Diet and the evolution of the earliest
Mark F. Teaford*†and Peter S. Ungar‡
*Department of Cell Biology and Anatomy, Johns Hopkins University School of Medicine, 725 North Wolfe Street, Baltimore, MD 21205; and‡Department
of Anthropology, University of Arkansas, Old Man 330, Fayetteville, AR 72701
Edited by F. Clark Howell, University of California, Berkeley, CA, and approved October 5, 2000 (received for review August 4, 2000)
Over the past decade, discussions of the evolution of the earliest
human ancestors have focused on the locomotion of the austra-
lopithecines. Recent discoveries in a broad range of disciplines
have raised important questions about the influence of ecological
factors in early human evolution. Here we trace the cranial and
dental traits of the early australopithecines through time, to show
that between 4.4 million and 2.3 million years ago, the dietary
capabilities of the earliest hominids changed dramatically, leaving
them well suited for life in a variety of habitats and able to cope
with significant changes in resource availability associated with
long-term and short-term climatic fluctuations.
scenarios of human origins (1, 2). Surprisingly, less attention has
been focused on the role played by diet in the ecology and
evolution of the early hominids (as usually received). Recent
work in a broad range of disciplines, such as paleoenvironmental
studies (3, 4), behavioral ecology (5), primatology (6), and
isotope analyses (7), has rekindled interests in early hominid
diets. Moreover, important new fossils from the early Pliocene
raise major questions about the role of dietary changes in the
need to focus not just on how the earliest hominids moved
between food patches, but also on what they ate when they got
This paper presents a review of the fossil evidence for the diets
of the Pliocene hominids Ardipithecus ramidus, Australopithecus
anamensis, Australopithecus afarensis, and Australopithecus africa-
nus. These hominids offer evidence for the first half of human
evolution, from our split with prehistoric apes to the earliest
as a roughly linear sequence from Ardipithecus to A. africanus,
they give us a unique opportunity to examine changes in dietary
adaptations of our ancestors over nearly 2 million years. We also
trace what has been inferred concerning the diets of the Miocene
hominoids to put changes in Pliocene hominid diets into a broader
temporal perspective. From such a perspective, it becomes clear
that the dietary capabilities of the early hominids changed dramat-
ically in the time period between 4.4 million and 2.3 million years
ago. Most of the evidence has come from five sources: analyses of
tooth size, tooth shape, enamel structure, dental microwear, and
early australopithecines, to increased dietary flexibility in the face
of climatic variability. Moreover, changes in diet-related adapta-
tions from A. anamensis to A. afarensis to A. africanus suggest that
hard, abrasive foods became increasingly important through the
Pliocene, perhaps as critical items in the diet.
ince the discovery of Australopithecus afarensis, many re-
searchers have emphasized the importance of bipedality in
incisors compared with molars and speculated that this might be
associated with terrestrial seed eating, as seen in Theropithecus
today. Although this idea has been the subject of some controversy
(12), Jolly’s efforts have stimulated considerable research on the
variety of living and fossil primates. Hylander (13), for example,
examined the relationship of incisor row length (relative to body
size) in a range of living anthropoids and found that those species
with larger incisors tend to consume larger, tougher fruits, whereas
those with smaller front teeth tend to feed on smaller foods, or
those that require less extensive incisal preparation, such as leaves
have looked to incisor size in early hominids and other fossil
primates for clues concerning diet.
What can incisor size tell us of the diets of Miocene apes?
Unfortunately, not as much as one would like. Ideally, to consider
relative incisor sizes among taxa, we need estimates of species body
weights based on attributes independent of the dentition. Such
estimates are unavailable for most taxa. Furthermore, Miocene
apes as a whole evidently had small incisors compared with extant
hominoids, in much the same way that platyrrhines as a whole have
relatively smaller incisors than do catarrhines, regardless of diet
(14). Such phylogenetic effects make it difficult to find an extant
comparative baseline series with which to compare these basal taxa
of uncertain phyletic affinities.
On the other hand, incisor size might give us some clues to diet
and tooth use for the early australopithecines, and we have good,
many of these taxa. If we look at a regression of maxillary central
incisor breadth on body size for species representing a variety of
catarrhine genera, we see a separation of cercopithecines (with
relatively larger incisors) above the line and colobines below
(Fig. 1). Furthermore, more frugivorous chimpanzees and or-
angutans fall above the line, whereas gibbons and gorillas fall
close to the line, with relatively smaller incisors. Indeed, values
for the living frugivorous great apes fall above the 95% confi-
dence limits of expected incisor size for modern catarrhines. The
human values fall below the 95% confidence limits, indicating
that we have very small incisors relative to body size.
are remarkably similar, and they fall very close to the regression
line, much like the gorilla. These results are similar to those
reported by Kay (21) and Ungar and Grine (17) and suggest that
these hominids used their incisors in ingestion to a similar
degree, although they all probably used these teeth less than
either the chimpanzee or orangutan. These data can also give us
some idea of whether a taxon often eats foods that require incisal
than orangutans, and they depend on smaller fruits requiring
little incisal preparation (17, 22, 23). From this perspective, the
australopithecines probably put less emphasis on foods that
require substantial incisor use, such as those with thick husks and
those with flesh adherent to large, hard seeds. Body weight
estimates and incisor size data for Ardipithecus ramidus and
Australopithecus garhi should provide even more insights.
This paper was submitted directly (Track II) to the PNAS office.
†To whom reprint requests should be addressed. E-mail: email@example.com.
Article published online before print: Proc. Natl. Acad. Sci. USA, 10.1073?pnas.260368897.
Article and publication date are at www.pnas.org?cgi?doi?10.1073?pnas.260368897
December 5, 2000 ?
vol. 97 ?
One of the hallmarks of the australopithecines has always been
their large, relatively flat molars (24–29). There are certainly
differences in the amount of occlusal relief between gracile and
robust australopithecines (30) (see below). However, by compari-
son with other primates, the australopithecines’ molars are still flat
plot of mandibular postcanine tooth area (MD ? BL, the product
of maximal mesiodistal and buccolingual diameters), where most
taxa have teeth larger than those of the modern orangutan (Fig. 2).
The only exception is Ardipithecus, which is more chimp-sized
in the P4–M1region, but intermediate between chimpanzees and
orangutans in the M2–M3region. Again, interpretations of such
differences are hampered by the lack of body size estimates for
Ardipithecus, but if a body size estimate of 51 kg is used for A.
anamensis (the average of the two different estimates based on
the tibia) (18), McHenry’s ‘‘megadontia quotient’’ for this taxon
is essentially identical to that for A. afarensis (Fig. 3). In other
words, its molars are large for a hominoid, but smaller than those
of A. africanus or the ‘‘robust’’ australopithecines.
range of mandibular molar sizes (Fig. 2). Many have postcanine
tooth areas larger than that of Ardipithecus, and some (such as
Ouranopithecus) even have larger postcanine tooth areas than that
of A. anamensis, but as all body size estimates for them have been
computed from dental remains, a megadontia quotient cannot be
computed. The main message from a simple look at postcanine
tooth size is that the earliest hominids make a nice progression
leading into subsequent hominids, but they do not have larger
postcanine teeth than all of the middle to late Miocene hominoids.
This might just mean that there are a variety of body sizes
sampled in these taxa. However, as shown by the work of Lucas
and colleagues (39), variations in tooth size are a means of
adapting to changes in the external characteristics of foods, such
as their size, shape, and abrasiveness. Clearly, some of these food
characteristics were changing during the evolution of the earliest
hominids, as postcanine teeth became relatively larger and
larger. However, evidence from the middle to late Miocene
shows that tooth size, by itself, cannot pinpoint the initial change
to a hominid diet, at least not with the samples at hand.
ratio of the areas of M1 and M3 (Fig. 4). Lucas et al. (39) showed
that this ratio was inversely related to the percentage of leaves,
of M1 to M3 area consumed more fruit than did those with a low
M1 to M3 ratio. When this is computed for the earliest hominids,
plus a sample of Miocene apes, a clear separation is evident, with
the early hominids, including Ardipithecus, showing higher ratios
than the Miocene apes. So, does this indicate more fruit in the diet
of the earliest hominids? To begin to answer this question, we must
look at analyses of tooth shape.
Variations in tooth shape are a means of adapting to changes in
the internal characteristics of foods, such as their strength,
toughness, and deformability (39–43). Clearly, foods are com-
plicated structures; thus it is impossible to describe all of the
Dashed lines indicate 95% confidence limits of the least-squares regression
plot (data from refs. 15 and 17–20).
Relative maxillary first incisor sizes in catarrhines. MD, mesiodistal.
apes, early hominids, and extant apes (data from refs. 8, 18–20, and 31–37).
Summed mandibular postcanine tooth areas (P4–M3) in Miocene
from refs. 18, 20, 27, 31, and 38).
Megadontia quotients for early hominids and extant primates (data
mesiodistal and buccolingual diameters (data from refs. 8, 18–20, and 31–37).
Ratios of M1 to M3 areas, defined as the products of maximal
Teaford and UngarPNAS ?
December 5, 2000 ?
vol. 97 ?
no. 25 ?
internal characteristics that might have confronted the earliest
hominids’ teeth. However, another approach is to describe the
capabilities of those teeth.
For example, tough foods, those that are difficult to fracture, are
generally sheared between the leading edges of sharp crests. In
contrast, hard brittle foods, those that are easy to fracture but
reciprocally concave, highly crested teeth have the capability of
efficiently processing tough items such as insect exoskeletons and
leaves, whereas rounder and flatter cusped teeth are best suited for
more folivorous species have the highest shearing quotients, fol-
lowed by those that prefer brittle, soft fruits; finally, hard-object
feeders have the lowest shearing quotients (21, 44).
Shearing crest studies have been conducted on early Miocene
African apes and middle to late Miocene European apes. These
studies suggest a considerable range of diets in these forms. For
example, Rangwapithecus and Oreopithecus have relatively long
shearing crests, suggesting folivory; Ouranopithecus has extremely
other Miocene taxa studied, such as Proconsul and Dryopithecus,
have the intermediate length crests of a frugivore (14, 45).
As for the early hominids, A. africanus had more occlusal relief
than did Paranthropus robustus, suggesting a dietary difference
between these species (30). Additional preliminary shearing quo-
tient studies support this idea while reaffirming that the australo-
pithecines, as a group, had relatively flat, blunt molar teeth and
lacked the long shearing crests seen in some extant hominoids (28).
By itself, this indicates that the earliest hominids would have had
difficulty breaking down tough, pliant foods, such as soft seed coats
and the veins and stems of leaves—although they probably were
capable of processing buds, flowers, and shoots.
Interestingly, as suggested by Lucas and Peters (46), another
tough pliant food they would have had difficulty processing is
meat. In other words, the early hominids were not dentally
preadapted to eat meat—they simply did not have the sharp,
reciprocally concave shearing blades necessary to retain and cut
such foods. In contrast, given their flat, blunt teeth, they were
admirably equipped to process hard brittle objects. What about
soft fruits? It really depends on the toughness of those fruits. If
they were tough, then they would also need to be precisely
retained and sliced between the teeth. Again, early hominids
would be very inefficient at it. If they were not tough, then the
hominids could certainly process soft fruits.
In sum, Miocene apes show a range of adaptations, including
folivory, soft-fruit eating, and hard-object feeding. This range
exceeds that of living hominoids and especially the early homi-
nids. Although studies of shearing crest length have been
conducted on only some of the early hominids, all evidence
indicates that the australopithecines had relatively flat molar
teeth compared with many living and fossil apes. These teeth
were well suited for breaking down hard, brittle foods, including
some fruits and nuts, and soft, weak foods, such as flowers and
buds; but again, they were not well suited for breaking down
tough pliant foods such as stems, soft seed pods, and meat.
Another area of interest regarding dental functional anatomy is the
study of enamel thickness. There are certainly methodological
differences between studies (47–52), but the consensus still seems
to be that the australopithecines had relatively thick enamel com-
pared with living primates, and that many of the Miocene apes also
had thick enamel (24, 28, 48–49, 51, 53–54). Interestingly, this
perspective may be changing as we get glimpses of more and more
new taxa. For instance, Conroy et al. (55) have noted that Otavipi-
thecus may have had thin enamel, and White et al. (8) have made
the same observation for Ardipithecus. Granted, in neither case do
still, the figures that have been quoted (less than 1 mm for
Otavipithecus and 1.1–1.2 mm for Ardipithecus) are far less than
those quoted for the australopithecines.
So what might be the functional significance of enamel
thickness? The most frequently cited correlations are between
the consumption of hard food items, or abrasive food items, and
thick molar enamel (58–59). There are many potential compli-
cating factors (51, 56, 59–60); thus it is perhaps not surprising
that the correlation between enamel thickness and diet is not a
perfect one (57). Moreover, thick enamel by itself does not
necessarily provide protection against hard objects, which com-
monly cause fracture of enamel (61). The best protection against
this is prism or crystallite decussation or interweaving. Maas (62,
63), Rensberger (64, 65), and others (42, 59) have shown that
prism and crystallite orientations can give clues to intricate
details of dental function, and that decussation can be an
effective crack-stopping mechanism in many animals. Only
anecdotal references to this phenomenon in Miocene apes and
early hominids have been made thus far, largely because more
detailed work generally requires the sectioning and etching of
teeth. Still, after some discussion and debate (48–49, 53), a
of prism decussation. Thus, the thick enamel of the early
hominids may have been a means of resisting breakage during
the consumption of hard objects and an adaptation that pro-
longed the life of the tooth, given an abrasive diet.
Numerous workers have recognized that microscopic wear on
the incisors and molars of primates reflects tooth use and diet.
For example, those primates that often use their front teeth in
ingestion have high densities of microwear striations on their
incisors. Furthermore, folivores have a high incidences of long
narrow scratches on their molars, whereas frugivores have more
pits on those surfaces. Among frugivores, hard-object feeders
have even higher pit incidences than soft-fruit eaters. These and
other relationships between microwear and feeding behaviors in
living primates have been used to infer diet in fossil forms.
Miocene apes have a remarkable range of microwear patterning,
greatly exceeding that of living hominoids. For example, rela-
tively high scratch densities suggest that Micropithecus, Rang-
wapithecus, and especially Oreopithecus (66) included more
leaves in their diets. In contrast, high pit percentages suggest that
Griphopithecus and Ouranopithecus (66) were hard-object spe-
cialists. Finally, intermediate microwear patterns suggest that
most other species studied, such as Gigantopithecus, Dendro-
pithecus, Proconsul, Dryopithecus, and, perhaps, Sivapithecus
(66–68), had diets dominated by soft fruits. These data give us
a glimpse of the extraordinary variation from which the last
common ancestor of apes and hominids evidently arose.
Unfortunately, little is known about the microwear of early
australopithecines. No microwear research has yet been pub-
lished for either Ardipithecus ramidus or A. anamensis, although
there has been some done on A. afarensis and A. africanus. The
work done on A. afarensis has been largely qualitative and
focused on the anterior teeth, and it suggests that these hominids
were beginning to exploit savanna resources (69). Furthermore,
Ryan and Johanson (70) argued that A. afarensis had a mosaic
of gorilla-like fine wear striae and baboon-like pits and mi-
croflakes, indicating the use of incisors to strip gritty plant parts
such as seeds, roots, and rhizomes. These authors also suggested
that there was a functional shift in the P3complex from ape-like
slicing and cutting to hominid puncture-crushing.
Work done on A. africanus has been more quantitative but has
focused on comparing this taxon to Paranthropus robustus rather
than to extant hominoids. Grine (71) found that A. africanus
molars have lower incidences of pitting than seen for Paranthro-
www.pnas.orgTeaford and Ungar
more homogeneity in orientation. Grine argued that compared
with the ‘‘robust’’ forms, A. africanus ate more soft fruits and
leaves. Comparisons with work from Teaford (72) places A.
africanus between Cebus olivaceus on one hand and Pan troglo-
dytes on the other. Work on A. africanus incisors has shown that
this taxon has higher microwear feature densities on all surfaces
examined than does Paranthropus (17). This suggests that A.
africanus processed a greater variety of foods with its front teeth,
including larger, more abrasive ones, than were encountered by
Paranthropus. Comparisons with an extant baseline series exam-
ined by Ungar (73) puts Australopithecus between Pongo pyg-
of anterior tooth use in ingestion.
In sum, then, the microwear suggests that, by the end of the
Miocene, hominoids had a wide range of diets. In contrast, A.
afarensis probably focused on soft fruit but also began to
incorporate into its diet abrasive, terrestrial resources that
required incisal stripping. A. africanus may still have focused on
soft fruit, particularly that which required a moderate amount of
incisal preparation. Clearly, considerably more work is needed
on these and other early hominids to put together a reasonable
picture of diet based on microwear evidence.
Finally, there are other lines of evidence that we can examine to
look for evidence of diet. Mandibular fragments are among the
most common bony remains found at hominid fossil sites, and
the architecture of this bone has been adapted to withstand
stresses and strains associated with oral food processing. Thus its
morphology probably reflects some aspects of diet. Analyses of
australopithecine mandibular biomechanics have focused on
corpus size and shape.
Comparisons with extant hominoids have shown that A.
afarensis and A. africanus have relatively thick mandibular
corpora (74, 75). The same pattern was also found for Paran-
thropus boisei and P. robustus. Fig. 5 shows mandibular robust-
icity index values for extant great apes, some Miocene apes, and
early australopithecines. The early hominids show relatively
thicker mandibular corpora than extant great apes and Miocene
catarrhines, suggesting a morphological shift in the former.
Both functional and nonfunctional interpretations have been
offered to explain this phenomenon. For example, it may simply
or a reduced canine. This is not a likely explanation, however, as
australopithecines still have relatively broad mandibles when
considered relative to molar size, and there appears to be no
relationship between mandibular robusticity and relative canine
size among the australopithecines (75).
Despite some inherent difficulties, it seems more likely that
the unique shape of the australopithecine mandibular corpus
relates to the functional demands of mastication. Thickened
mandibles can act to resist extreme stresses associated with
transverse bending (that is, ‘‘wishboning’’) and torsion. Because
wishboning stresses decline toward the back of the corpus,
torsion is likely a more important explanation. Corpus torsion
can result from bite force and muscle activity during mastication.
Therefore, it may be that australopithecine mandibular mor-
phology reflects elevated stresses associated with unusual me-
chanical demands. Daegling and Grine (75) suggest that austra-
lopithecines may have eaten fibrous, coarse foods that required
repetitive loading. While this fails to explain why colobines do
not have thick corpora, it does suggest a fundamental difference
between australopithecines and living great apes that may reflect
a shift in diet in the early hominids.
Studies of corpus shape in A. anamensis and Ardipithecus
ramidus will likely provide further clues regarding differences in
mandibular architecture between great apes and later australo-
pithecines. Corpus robusticity indices for A. anamensis below M1
average 53.5 (M. Leakey, personal communication). These
values fall at the upper range for extant hominoids (Pan ?
39.2–57.8; Gorilla ? 43.5–59.7; Pongo ? 35.7–52.0) and at the
lower end of the range for later fossil hominids (A. afarensis ?
48.4–68.9, A. africanus ? 54.8–79.0) (Fig. 5) (data from Daeg-
ling and Grine and Lockwood et al.) (75, 85).
In sum, the architecture of the mandibular corpus suggests
that the ‘‘gracile’’ australopithecines differed from living apes in
their abilities to dissipate masticatory stresses. Taken with other
lines of evidence, this certainly suggests a difference in diet
between living apes and A. anamensis, and between A. anamensis
and later hominids, with A. anamensis intermediate between the
African ape and later australopithecine conditions.
The australopithecines exhibited a complex of morphological fea-
tures related to diet that are unique compared with living homi-
noids or Miocene apes. These early hominids all had small- to
moderate-sized incisors; large, flat molars with little shear poten-
tial; a ratio of first to third molar area that was low compared with
those of extant apes, but generally higher than those of Miocene
apes; thick tooth enamel; and thick mandibular corpora. This suite
of traits is distinctive of australopithecines and suggests a dietary
shift at or near the stem of hominid evolution. Their thick-
enameled, flattened molars would have had great difficulty prop-
agating cracks through tough foods, suggesting that the australo-
pithecines were not well suited for eating tough fruits, leaves, or
meat. The dental microwear data agree with this conclusion, as the
australopithecine patterns documented to date are most similar to
those of modern-day seed predators and soft fruit eaters. Further-
more, given their comparatively small incisors, these hominids
extensive incisal preparation. Instead, the australopithecines would
have easily been able to break down hard, brittle foods. Their large
flat molars would have served well for crushing, and their thick
enamel would have withstood abrasion and fracture. Their man-
dibular corpora would probably have conferred an advantage for
resisting failure, given high occlusal loads. In essence, for much of
their history, the australopithecines had an adaptive package that
not particularly tough. The early hominids could also have eaten
both abrasive and nonabrasive foods. This ability to eat both hard
the early hominids particularly well suited for life in a variety of
habitats, ranging from gallery forest to open savanna.
Leakey, personal communication).
Mandibular corpus shape (data from refs. 75, 76, and 85 and M.
Teaford and Ungar PNAS ?
December 5, 2000 ?
vol. 97 ?
no. 25 ?
Does this mean we can talk of a characteristic ‘‘australo-
the australopithecines shared many features in common, they
also differed from one another, suggesting a change in diet
through time. Such morphological changes occurred as a mosaic,
much as that seen for locomotor anatomy.
Much of the evidence for Ardipithecus ramidus is not yet
available, but despite its thin molar enamel and absolutely
smaller teeth than those of later hominids, it shows molar size
proportions that may hint at dietary changes to come. A.
anamensis shows the first indications of thicker molar enamel in
a hominid, and its molar teeth were equivalent in size to those
of A. afarensis. Still, its mandibular corpus is intermediate in
robusticity between those of living great apes and later austra-
lopithecines. This combination of features suggests that A.
anamensis might have been the first hominid to be able to
effectively withstand the functional demands of hard and per-
haps abrasive objects in its diet, whether or not such items were
frequently eaten or were only an important occasional food
source. A. afarensis was similar to A. anamensis in relative tooth
sizes and probably enamel thickness, yet it did show a large
increase in mandibular robusticity. This increase may be due to
changes in peak force magnitude or degree of repetitive loading
in mastication. Either way, hard and perhaps abrasive foods may
have become even more important components of the diet of A.
afarensis. A. africanus shows yet another increase in postcanine
tooth size, which by itself would suggest an increase in the sizes
and abrasiveness of foods. However, its molar microwear does
not show the degree of pitting one might expect from a classic
hard-object feeder. Thus, even A. africanus has evidently not
begun to specialize in hard objects, but rather has emphasized
dietary breadth. In contrast, subsequent ‘‘robust’’ australo-
pithecines do show hard-object microwear and craniodental
specializations, suggesting a substantial departure in feeding
adaptive strategies early in the Pleistocene.
In sum, diet was probably an important factor in the origin and
early evolution of our family. The earliest australopithecines
show a unique suite of diet-related features unlike those of
Miocene apes or living hominoids. Such features suggest that the
earliest hominids may have begun to experiment with harder,
This does not mean that all of the australopithecines were
specialized hard-object feeders. It merely means that, through
time, they acquired the ability to feed on hard objects. Many
modern primates need to consume critical ‘‘fall-back foods’’ at
certain times of the year (6), and it may well be that the earliest
australopithecines resorted to the consumption of hard objects
only in such situations, whereas the robust australopithecines
relied on them far more regularly.
Another important aspect of early hominid trophic adaptations
is evident from data presented here—the dietary shift from apes to
early hominids did not involve an increase in the consumption of
tough foods, and so the australopithecines were not preadapted for
suggesting that the australopithecines did in fact consume signifi-
cant amounts of meat (7) and (ii) nutritional work suggesting that
meat may have provided critical nutrients for both young and old
hominids (77–79). There would seem to be three different ways to
reconcile these perspectives. First, the present study has reviewed
only craniodental features related to diet. If the australopithecines
used other means for ingesting and processing meat (e.g., tools),
craniodental evidence suggests (80, 81). Second, the heavy C3
signature found in A. africanus (7) may reflect the consumption of
underground storage organs of C3 plants rather than meat (82).
Third, the functional analyses of the teeth assume that all meat has
the same degree of toughness. This may not be the case. Studies of
the physical properties of food have thus far focused on plant
remains, with only brief mention of the toughness of materials like
skin (40, 46). Variations in toughness between animal tissues might
matrix. Furthermore, the physical effects of decomposition might
render meat less tough and more readily processed by hominids. If
this is so, it could be further evidence in support of scavenging as
part of the early hominid way of life.
patterns of climatic change for some time (3, 4). The focus of much
of the recent work has been on the origin of the genus Homo. Can
the dietary shifts in the earliest hominids also be tied to such
changes? Whereas there is some evidence of large-scale climatic
changes around the Mediterranean (83) and unusual faunal turn-
over in parts of western Asia (84), there are no large-scale changes
is the slow and inexorable cooling and drying of the Miocene, but
perhaps the crucial result of this was an increase in microhabitat
variability. Certainly, there are limits to our paleoecological evi-
dence from this period, but as Potts (4) has noted, ‘‘in general, the
oldest hominids were associated with a diverse range of habitats.’’
These included lake and river margins, woodland, bushland, and
savanna. Potts (4) has emphasized that locomotor versatility was a
In such a land of variable opportunities, the generalized cranio-
dental toolkit of the earliest hominids may have had a distinct
advantage, as it allowed our forbears the flexibility to cope with
short-term and long-term climatic variations and the resultant
changes in resource availability.
We are grateful to the Governments of Ethiopia, Kenya, and Tanzania and
especially to the National Museums of Ethiopia, Kenya, and Tanzania for
permission to study early hominid specimens in their care. This work was
supported by National Science Foundation Grants SBR 9804882 and
1. Lovejoy, C. O. (1975) in Primate Functional Morphology and Evolution, ed.
Tuttle, R. L. (Mouton, The Hague), pp. 291–326.
2. Susman, R. L., Stern, J. T. & Jungers, W. L. (1984) Folia Primatol. 43, 113–156.
3. Vrba, E. S. (1995) in Paleoclimate and Evolution, with Emphasis on Human
Origins, eds. Vrba, E. S., Denton, G. H., Partridge, T. C. & Burckle, L. H. (Yale
Univ. Press, New Haven, CT), pp. 24–45.
4. Potts, R. (1998) Yearbook Phys. Anthropol. 41, 93–136.
5. O’Connell, J. F., Hawkes, K. & Blurton Jones, N. G. (1999) J. Hum. Evol. 36,
7. Sponheimer, M. & Lee-Thorp, J. A. (1999) Science 283, 368–370.
8. White, T. D., Suwa, G. & Asfaw, B. (1994) Nature (London) 371, 306–312.
9. Asfaw, B., White, T., Lovejoy, O., Latimer, B., Simpson, S. & Suwa, G. (1999) Science
10. Ward, C., Leakey, M. & Walker, A. (1999) Evol. Anthropol. 7, 197–205.
11. Jolly, C. J. (1970) Man 5, 1–26.
12. Dunbar, R. I. M. (1976) J. Hum. Evol. 5, 161–167.
13. Hylander, W. L. (1975) Science 189, 1095–1098.
14. Kay, R. F. & Ungar, P. S. (1997) in Function, Phylogeny and Fossils: Miocene
Hominoids and Great Ape and Human Origins, eds. Begun, D. R., Ward, C. &
Rose, M. (Plenum, New York), pp. 131–151.
15. Jungers, W. L. (1988) in Evolutionary History of the ‘‘Robust’’ Australopithecines,
ed. Grine, F. E. (de Gruyter, New York), pp. 115–125.
16. McHenry, H. M. (1992) Evol. Anthropol. 1, 15–20.
17. Ungar, P. S. & Grine, F. E. (1991) J. Hum. Evol. 20, 313–340.
19. Wood, B. A. (1991) Hominid Cranial Remains, Koobi Fora Research Project
(Clarendon, Oxford), Vol. 4.
20. Coffing, K., Feibel, C., Leakey, M. & Walker, A. (1994) Am. J. Phys. Anthropol.
www.pnas.orgTeaford and Ungar
21. Kay, R. F. (1984) in Adaptations for Foraging in Nonhuman Primates: Contri-
butions to an Organismal Biology of Prosimians, Monkeys and Apes, eds.
Rodman, P. S. & Cant, J. G. H. (Columbia Univ. Press, New York), pp. 21–53.
22. Ungar, P. S. (1994) Am. J. Phys. Anthropol. 95, 197–219.
23. Ungar, P. S. (1996) Am. J. Primatol. 38, 145–156.
24. Robinson, J. T. (1956) Mem. Transvaal Mus. 9, 1–179.
25. Wolpoff, M. H. (1973) Am. J. Phys. Anthropol. 39, 375–394.
26. Wood, B. A. & Abbott, S. A. (1983) J. Anat. 136, 197–219.
27. McHenry, H. M. (1984) Am. J. Phys. Anthropol. 64, 297–306.
28. Kay, R. F. (1985) Annu. Rev. Anthropol. 14, 315–341.
29. Suwa, G., Wood, B. A. & White, T. D. (1994) Am. J. Phys. Anthropol. 93,
30. Grine, F. E. (1981) S. Afr. J. Sci. 77, 203–230.
31. Mahler, P. E. (1973) Ph.D. thesis (Univ. of Michigan, Ann Arbor).
32. Alpagut, B., Andrews, P. & Martin, L. (1990) J. Hum. Evol. 19, 397–422.
33. Alpagut, B., Andrews, P., Fortelius, M., Kappelman, J., Temizsoy, I., C ¸elebi,
H. & Lindsay, W. (1996) Nature (London) 382, 349–351.
34. Andrews, P. (1978) Bull. Br. Mus. (Nat. Hist.) 30, 85–224.
35. Begun, D. R. & Gu ¨lec ¸, E. (1998) Am. J. Phys. Anthropol. 105, 279–314.
36. de Bonis, L. & Melentis, J. (1984) Cour. Forschungsinst. Senckenberg 69, 13–23.
37. Leakey, M. G., Feibel, C. S., McDougall, I., Ward, C. & Walker, A. (1998)
Nature (London) 393, 62–66.
38. McHenry, H. M. (1988) in Evolutionary History of the ‘‘Robust’’ Australo-
pithecines, ed. Grine, F. E. (de Gruyter, New York), pp. 133–147.
39. Lucas, P. W., Corlett, R. T. & Luke, D. A. (1986) Z. Morphol. Anthropol. 76,
40. Lucas, P. W. & Teaford, M. F. (1994) in Colobine Monkeys: Their Ecology, Behaviour
and Evolution, eds. Davies, A. G. & Oates, J. F. (Cambridge Univ. Press, Cambridge,
U.K.), pp. 173–203.
41. Spears, I. R. & Crompton, R. H. (1996) J. Hum. Evol. 31, 517–535.
42. Strait, S. G. (1997) Evol. Anthropol. 5, 199–211.
43. Yamashita, N. (1998) Am. J. Phys. Anthropol. 106, 169–188.
44. Meldrum, D. J. & Kay, R. F. (1997) Am. J. Phys. Anthropol. 102, 407–428.
45. Ungar, P. S. & Kay, R. F. (1995) Proc. Natl. Acad. Sci. USA 92, 5479–5481.
46. Lucas, P. W. & Peters, C. R. (2000) in Development, Function and Evolution of
Teeth, eds. Teaford, M. F., Smith, M. M. & Ferguson, M. W. J. (Cambridge
Univ. Press, Cambridge, U.K.), pp. 282–289.
47. Martin, L. B. (1985) Nature (London) 314, 260–263.
48. Beynon, A. D. & Wood, B. A. (1986) Am. J. Phys. Anthropol. 70, 177–193.
49. Grine, F. E. & Martin, L. B. (1988) in Evolutionary History of the ‘‘Robust’’
Australopithecines, ed. Grine, F. E. (de Gruyter, New York), pp. 3–42.
50. Beynon, A. D., Dean, M. C. & Reid, D. J. (1991) Am. J. Phys. Anthropol. 86,
51. Macho, G. A. & Thackeray, J. F. (1992) Am. J. Phys. Anthropol. 89, 133–143.
52. Spoor, C. F., Zonneveld, F. W. & Macho, G. A. (1993) Am. J. Phys. Anthropol.
53. Gantt, D. G. (1986) in Comparative Primate Biology. Volume 1. Systematics,
Evolution, and Anatomy. Swindler, D. R. & Erwin, J. (Liss, New York), Vol.
1, pp. 453–475.
54. Andrews, P. & Martin, L. (1991) Philos. Trans. R. Soc. London B 334, 199–209.
55. Conroy, G. C., Pickford, M., Senut, B., Van Couvering, J. & Mein, P. (1992)
Nature (London) 356, 144–148.
56. Martin, L. B. (1983) Ph.D. thesis (Univ. of London, London).
57. Maas, M. C. & Dumont, E. R. (1999) Evol. Anthropol. 8, 133–152.
58. Kay, R. F. (1981) Am. J. Phys. Anthropol. 55, 141–151.
59. Dumont, E. R. (1995) J. Mammal. 76, 1127–1136.
60. Macho, G. A. & Berner, M. E. (1993) Am. J. Phys. Anthropol. 92, 189–200.
61. Teaford, M. F., Maas, M. C. & Simons, E. L. (1996) Am. J. Phys. Anthropol. 101,
62. Maas, M. C. (1993) Am. J. Phys. Anthropol. 92, 217–233.
63. Maas, M. C. (1994) Am. J. Phys. Anthropol. 95, 221–242.
64. Rensberger, J. M. (1997) in Tooth Enamel Microstructure, eds. Koenigswald, W.
v. & Sander, M. (Balkema, Rotterdam), pp. 237–257.
66. Ungar, P. S. (1996) J. Hum. Evol. 31, 335–366.
67. Teaford, M. F. & Walker, A. C. (1984) Am. J. Phys. Anthropol. 64, 191–200.
68. Daegling, D. J. & Grine, F. E. (1994) S. Afr. J. Sci. 90, 527–532.
69. Puech, P.-F. & Albertini, H. (1984) Am. J. Phys. Anthropol. 65, 87–91.
70. Ryan, A. S. & Johanson, D. C. (1989) J. Hum. Evol. 18, 235–268.
71. Grine, F. E. (1986) J. Hum. Evol. 15, 783–822.
72. Teaford, M. F. (1988) Scanning Microsc. 2, 1149–1166.
73. Ungar, P. S. (1998) Evol. Anthropol. 6, 205–217.
74. Hylander, W. L. (1988) in Evolutionary History of the ‘‘Robust’’ Australo-
pithecines, ed. Grine, F. E. (de Gruyter, New York), pp. 55–58.
75. Daegling, D. J. & Grine, F. E. (1991) Am. J. Phys. Anthropol. 86, 321–339.
76. Smith, R. J. (1980) Ph.D. dissertation (Yale Univ., New Haven, CT).
77. Milton, K. & Demment, M. (1988) Am. J. Primatol. 46, 45–52.
78. Milton, K. (1999) Evol. Anthropol. 8, 11–21.
79. Speth, J. D. (1989) J. Hum. Evol. 18, 329–343.
80. Blumenschine, R. J. & Cavallo, J. A. (1992) Sci. Am. 267, 90–96.
81. de Heinzelin, J., Clark, J. D., White, T., Hart, W., Renne, P., WoldeBabriel, G.,
Beyene, Y. & Vrba, E. (1999) Science 284, 625–629.
82. Lee-Thorp, J. (2001) in The Evolution of Human Diet, eds. Ungar, P. S. &
Teaford, M. F. (Greenwood, New Haven, CT), in press.
83. Bernor, R. L. (1983) in New Interpretations of Ape and Human Ancestry, eds.
Ciochon, R. L. & Corruccini, R. S. (Plenum, New York), pp. 21–64.
84. Barry, J. C. (1995) in Paleoclimate and Evolution, with Emphasis on Human
Origins, eds. Vrba, E. S., Denton, G. H., Partridge, T. C. & Burckle, L. H. (Yale
Univ. Press, New Haven, CT), pp. 115–134.
85. Lockwood, C. A., Kimbel, W. H. & Johanson, D. C. (2000) J. Hum. Evol. 39,
Teaford and UngarPNAS ?
December 5, 2000 ?
vol. 97 ?
no. 25 ?