Isotopic evidence for an early shift to C

4
resources by
Pliocene hominins in Chad
Julia Lee-Thorp
a,1
, Andossa Likius
b
, Hassane T. Mackaye
b
, Patrick Vignaud
c
, Matt Sponheimer
d
, and Michel Brunet
c,e
a
Research Laboratory for Archaeology, Oxford University, Oxford OX1 3QY, United Kingdom;
b
Dpartement de Palontologie, Universit de NDjamena,
NDjamena, Chad;
c
Institut de Paloprimatologie, Palontologie Humaine: Evolution et Paloenvironnements, Centre National de la Recherche Scientique
Unit Mixte de Recherche 7262, Universit de Poitiers, 86022 Poitiers Cedex, France;
d
Department of Anthropology, University of Colorado at Boulder,
Boulder, CO 80309; and
e
Chaire de Palontologie Humaine, Collge de France, 75232 Paris Cedex 05, France
Edited by Stanley H. Ambrose, University of Illinois at UrbanaChampaign, Urbana, IL, and accepted by the Editorial Board October 5, 2012 (received for
review March 11, 2012)
Foods derived from C
4
plants were important in the dietary ecol-
ogy of early Pleistocene hominins in southern and eastern Africa,
but the origins and geographic variability of this relationship re-
main unknown. Carbon isotope data show that Australopithecus
bahrelghazali individuals from Koro Toro in Chad are signi-
cantly enriched in
13
C, indicating a dependence on C
4
resources.
As these sites are over 3 million years in age, the results extend the
pattern of C
4
dependence seen in Paranthropus boisei in East Africa
by more than 1.5 million years. The Koro Toro hominin fossils were
found in argillaceous sandstone levels along with abundant graz-
ing and aquatic faunal elements that, in combination, indicate the
presence of open to wooded grasslands and stream channels asso-
ciated with a greatly enlarged Lake Chad. In such an environment,
the most abundant C
4
plant resources available to A. bahrelghazali
were grasses and sedges, neither of which is usually considered as
standard great ape fare. The results suggest an early and funda-
mental shift in hominin dietary ecology that facilitated the exploi-
tation of new habitats.
R
ecent stable carbon isotope and microwear studies of early
hominin diets have challenged widely held perceptions that
the development of their robust craniomandibular morphology,
large molars and premolars, and thick enamel, are adaptations
for processing hard foods associated with increasingly xeric
environments in the Pliocene and Pleistocene (1, 2). In partic-
ular, carbon isotope (
13
C/
12
C) ratios of fossil tooth enamel have
demonstrated previously unsuspected contributions to austra-
lopithecine diets. Most
13
C data obtained for Australopithecus
africanus and Paranthropus robustus in South Africa suggest sig-
nicant dietary contributions from
13
C-enriched biomass (37),
which can be seasonally variable (8, 9). [By convention,
13
C/
12
C
ratios are expressed as parts per 1,000 in the notation relative
to the VPDB standard, as
13
C = (R
sample
/R
standard
1)
1,000, where R =
13
C/
12
C.] In eastern Africa, data for Para-
nthropus boisei suggest a very strong C
4
emphasis by 1.8 Ma (10,
11). Ardipithecus ramidus from Aramis is a notable exception,
showing little engagement with C
4
resources despite their abun-
dance in the local environment (12). On present evidence, the
earliest known evidence for C
4
consumption is observed in three
of four A. africanus Makapansgat individuals from the Late
Pliocene at 2.8 Ma (4).
Australopithecus bahrelghazali is the only Pliocene hominin,
and indeed the only primate, to have been found at the Koro Toro
(KT) sites (KT12 and KT13) in the Chad Basin, well beyond the
abundant fossil occurrences in the Rift Valley to the East, and
the southern African sites to the South (Fig. 1) (1315). Bio-
chronological estimates for both sites indicate ages of 33.5 Ma
(13, 15), whereas cosmogenic
10
Be/
9
Be nuclide dating of the KT
12 sediments containing the A. bahrelghazali holotype specimen,
KT12/H1, provides an age of 3.58 0.27 Ma (16). These dates
suggest contemporaneity with Australopithecus afarensis in
eastern Africa.
The fragmentary remains of several A. bahrelghazali individuals
occur in a poorly consolidated, perilacustrine, clayey sandstone
together with terrestrial (Suidae, Loxodonta, Stegodon, Giraffa,
Parmularius, Antidorcas, Sivatherium, Cerathotherium, Hipparion,
Bovini, Kobus) and freshwater (Perciformes, Siluriformes, Trionyx,
cf. Tomistoma, Anatidae, Hexaprotodon, and Lutrinae indet.)
taxa (1315, 17). Relative abundances of bovids at KT13 show a
predominance of open-country grazers (17), including Reduncini
indicative of edaphic grasslands (18). The lack of Aepyceros,
Tragelaphini, and primates (other than A. bahrelghazali) suggests
a paucity of woodland or thicket (1315, 17), but the occurrence
(10% of terrestrial fauna) of Giraffa (19) indicates the con-
tinued presence of some tree cover in the Pliocene sites. Existing
faunal enamel
13
C data are consistent with these observations,
showing overall C
4
plant dominance (20). Together, the paleo-
environmental data indicate a northern Chad Basin landscape
with extensive grasslands and sparser trees, bordering an ex-
panded lake and associated stream channels.
Hence, A. bahrelghazali survived in an environment more open
than that inhabited by most other Pliocene taxa including its
contemporary A. afarensis in eastern Africa (18, 2123). Their
survival in such a habitat was possibly facilitated by dietary niche
expansion to include novel C
4
resources, the exploitation of
which is uncommon among extant great apes including savanna-
dwelling chimpanzees (24, 25), although more common among
many savanna baboons (26, 27). The appearance of C
4
resources
among early hominins would represent a signicant ecological
shift, but the timing and geographical variability across Africa
remains unclear. Here, we explore A. bahrelghazalis diet by com-
paring the carbon isotope composition of their tooth enamel
against that of the coexisting fauna, and against the published
data for early hominins elsewhere.
Results
A. bahrelghazali
13
C values range from 0.8 to 4.4. They
are similar to, but reach slightly lower values than, those of the
Alcelaphini, Reduncini, Equidae, Suidae, and Proboscidea from
the KT sites (Fig. 2, Table S1). The detailed laser scan data for
two of the hominids are given in Table S2. The results indicate a
predominance of C
4
dietary resources (5580% by linear in-
terpolation). Carbon isotope data alone cannot distinguish
whether carbon of C
4
origin was from plant or animal sources,
but in this case, the high proportions suggest that the primary C
4
dietary resources were plant staples. Consumption of animals
(e.g., termites, rodents, grazing herbivores) reliant on the abundant
Author contributions: J.L.-T., A.L., H.T.M., P.V., and M.B. designed research; J.L.-T., M.S.,
and M.B. performed research; A.L., H.T.M., and P.V. contributed sample materials; J.L.-T.
and M.S. analyzed data; and J.L.-T. and M.B. wrote the paper.
The authors declare no conict of interest.
This article is a PNAS Direct Submission. S.H.A. is a guest editor invited by the Editorial
Board.
1
To whom correspondence should be addressed. E-mail: julia.lee-thorp@rlaha.ox.ac.uk.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1204209109/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1204209109 PNAS Early Edition | 1 of 4
A
N
T
H
R
O
P
O
L
O
G
Y
C
4
vegetation cannot be excluded and they may have formed
components of the diet, as inferred for the South African aus-
tralopithecines (5, 7, 28). Very high proportions of animal food,
however, are not plausible for hominins given that even recent
humans such as the Kalahari San rely most heavily on plant foods
(80%) and less on game (29). Moreover, hominins lack the ap-
propriate dental morphology. Therefore, we focus on C
4
plants
in the discussion below.
It is readily apparent that the A. bahrelghazali
13
C values
differ from those of most hominins analyzed so far (Fig. 3). They
differ from
13
C values for Ar. ramidus, which largely reect
C
3
resource consumption (KruskalWallis test and Conover
Inman pairwise comparisons; P < 0.001). A. bahrelghazali is
more enriched in
13
C than the South African austral-
opithecines A. africanus and P. robustus (P < 0.01). In con-
trast, its
13
C values are not signicantly different from those
of P. boisei from the Olduvai and Lake Turkana regions in
East Africa, where mean
13
C (x = 1.3 0.9) suggests as
much as 85% of carbon from C
4
sources (10, 11).
Discussion
Although necessarily few because of the low numbers of
A. bahrelghazali tooth specimens, the
13
C data demonstrate
that a reliance on C
4
resources is not an aberration conned to
P. boisei, but that it had a much earlier genesis. The Chadian
australopithecines occupied a largely open landscape dominated
by C
4
plants, in which traditional hominoid woodland or for-
est-derived resources were sparse, so a signicant C
4
dietary
component might be considered unsurprising. However, both
tropical grasses and sedges, which form the major and minor C
4
components of the ora in the open, low to midlatitude regions
of Africa, respectively (3032), are unusual foods for hominoids.
C
4
, or Kranz, grasses are challenging for herbivores gener-
ally, and because hominins lack dental and digestive adapta-
tions of grazing taxa to cope with the bundle sheath cell walls
and abundant silica phytoliths (30), one must consider whether
plant parts other than leaves/blades may have been the major
focus. Jolly (33) famously suggested a gelada baboon analogy
for early hominin diet focused on exploitation of grass seeds
and rhizomes. However, the seeds are seasonally restricted,
small food packages that are costly to harvest, and most grass
rhizomes are not eshy storage organs (see gure 1j in ref. 34),
so the potential value of these resources for hominins has been
neglected (see 28). Nevertheless, East African savanna baboons
regularly exploit all of the above-ground, fresh parts of taxa
such as Cynodon and Sporobolus on a seasonal basis (26, 3538).
The more nutritious and lower ber grass meristems may be
exploited year-round (36, 37). Furthermore, some tropical
grasses including Sporobolus rangei have eshy tubers or corms
that form an important component of Amboseli baboons diets
(26, 35, 37).
Several studies have suggested that exploitation of energy-rich
geophytes or underground storage organs (USOs) (bulbs, corms,
tubers, and rhizomes) represented an important adaptive shift in
hominin evolution, especially as related to the transition to early
Homo (e.g., 3941). In general, most USOs are C
3
. The main
exceptions are the C
4
members of the Cyperaceae (sedges),
many of which produce tubers and more occasionally, corms,
in addition to rhizomes and stolons. Species in the C
4
genus
Cyperus form prominent and highly productive components of
the vegetation fringing African tropical and subtropical wetlands
including the Chad basin (42). The nutritive value of the rhi-
zomes of the most abundant C
4
taxa fringing Lake Chad today,
Cyperus papyrus, Cyperus laevigatus, and Cyperus articulatus (42)
remains unclear given the prevalence of indigestible structural
carbohydrates and/or secondary compounds (43). Cyperus species
also occur in both swamp, and drier sandy or loamy settings
where there is seasonal moisture (43); many of these taxa of
grassier habit bear eshy tubers or corms (e.g., Cyperus
rotundus, Cyperus esculentus, Cyperus blysmoides, Cyperus obtusi-
orus) or swollen bases (e.g., Cyperus bulbosus, Cyperus amauropus,
Cyperus merkeri) that are eaten by both baboons and humans
Fig. 1. Map of Africa showing the Chad Basin and the location of the KT12
and KT13 sites, and the hominin sites in South Africa and eastern Africa from
which the
13
C data shown in Fig. 3 were obtained.
Fig. 2. Plot of
13
C values for A. bahrelghazali and Alcelaphini, Equidae,
Suidae, Proboscidea, Reduncini, and Bovini from the KT sites, and data for
Paracamelus sp., Giraffa sp., and Diceros bicornis from the nearby, but older,
site of Kossom Bougoudi (labeled KB browsers). Data for KT12 and KT13
from this study are combined with published data (20) for KT sites, and for
KB (data provided in Table S1). The KB browsers are shown here to indicate
values for C
3
feeders in the region, because although Giraffa fossils are
present in the KT sites (20), no suitable tooth enamel was available for

13
C analysis.
2 of 4 | www.pnas.org/cgi/doi/10.1073/pnas.1204209109 Lee-Thorp et al.
today (3638, 43). Ethnohistoric sources suggest that Cyperus
usitatus, Cyperus fulgens, and C. esculentus were important plant
resources for Late Stone Age foragers in the semiarid South
African Karoo (44). The above-ground parts of many sedges are
also eaten. For instance, humans use the culms of C. papyrus as
casual or famine foods (43) and baboons the owers and bulbs
of Cyperus grandibulbosus (38).
Given the nature of the Chad basin environment, it would
seem that abundant edible C
4
plant resources were available to
A. bahrelghazali, including the below- and above-ground parts of
several grass and sedge taxa. Moreover, many of the invertebrate
and vertebrate fauna are likely to be enriched in
13
C in a C
4
-rich
biome. The remaining (C
3
) portion of A. bahrelghazalis diet may
have included C
3
macrophytes, as suggested in ref. 41, or the
more traditional hominoid foods such as fruits, pods, owers,
and insects. The semiarid terrestrial ora that today stretches away
from the current Lake Chad and colonizes the dunes includes
Hyphaene thebaica (doum palm), Acacia nilotica, and Ficus
ingens (42), all of which produce widely consumed fruits and
pods (e.g., refs. 3538). The presence of Giraffa in the KT sites
suggests that similar species might have been present in the
Pliocene environment.
Whether the A. bahrelghazali individuals relied on C
4
grasses
or sedges, or both, these resources are seldom consumed (if at
all) by most primates. Savanna baboons are a notable exception
as they spend signicant amounts of time foraging for C
4
plants
(26, 3538), although the sparse existing
13
C data for these
baboons in East Africa show a lower reliance on such resources
than one might expect based on foraging time. Bone collagen
and enamel values for modern Papio anubis individuals (n = 3,
respectively) suggest C
4
contributions of <20% in East Turkana
and Naivasha (45) and 50% in Laikipia (46). Many of the main
C
4
plant resources (i.e., grasses and sedges) are high in ber
and/or complex starches, although some parts, such as grass
meristems, offer higher nutrient rewards. For instance, the tubers
of cultivated C. esculentus varieties are high in structural car-
bohydrates and lignin (47). In comparison with baboons, values
for the hominins A. bahrelghazali and P. boisei suggest higher C
4
contributions. A focus on the C
4
plant parts as suggested above
may be consistent with observations that occlusal microwear in
A. anamensis, A. afarensis, and P. boisei (1, 48) lack indications
for the processing of hard items.
Our data show that by 3.5 Ma A. bahrelghazali was fully
engaged in exploiting C
4
biomass. The results imply that aus-
tralopithecines had become broad generalists foraging opportu-
nistically for locally abundant resources that included signicant
quantities of savanna resources, unlike chimps. Alternatively, they
might imply an emerging specialism on grasses and/or sedges. In
the former case, we should expect strong variability related to the
environment differences, whereas in the latter we might expect
most early australopithecine diets to reect a strong C
4
focus in
all habitats where both grasses and sedges are predominantly C
4
.
This could be tested by isotopic analyses of temporally and mor-
phologically similar A. afarensis, which generally inhabited more
wooded environments in eastern Africa (18, 21, 22) and might
be expected to consume more C
3
resources there if they were
dietary generalists. Whatever the case, we suggest that a ca-
pacity to exploit C
4
resources provided the means to survive in
environments completely unlike the earlier ancestral forests or
denser woodlands.
Methods
Because there are very few A. bahrelghazali specimens and it was imperative
to minimize the impact of sampling, we followed a dual approach using
laser ablation (LA) mass spectrometry (following methods outlined in ref. 49)
for
13
C determination of two of the hominin specimens, and standard acid
hydrolysis preparation for the third hominin and all of the fauna. LA analysis
permitted isotopic determination from a minute enamel sliver previously
dislodged from KT12/H1a during morphometric measurements. Although
this method is minimally invasive, there are drawbacks, including lower
precision (49), and lack of comparability for
18
O/
16
O ratios because the laser
releases all oxygen species in enamel. Therefore, we rst compared
13
C
results for the two methods, using two fossil teeth in our collection (LT4
and LT6), and thereafter, three KT faunal specimens. The results showed rea-
sonable agreement (see Table S3 for the KT data). LT4 and LT6 were sub-
sequently used as internal standards to monitor the response of the laser
and to calibrate the data. We considered that the lower isotopic homoge-
neity in enamel was outweighed by the requirement for a standard of
similar mineralogy to the samples, given the sensitivity of LA to this factor
(SI Methods).
All teeth were cleaned as outlined in SI Methods. Enamel powders destined
for standard 100% H
3
PO
4
hydrolysis mass spectrometry were pretreated rst
(SI Methods) before analysis in a GasBench II coupled to a Thermo Finnigan
Delta V. Multiple measurements of NBS 19 and laboratory calcium carbonate
standards, using H
3
PO
4
hydrolysis, indicated precision better than 0.1 for
13
C/
12
C, and 0.2 for
18
O/
16
O, whereas multiple (n = 10) determinations of
the fossil enamel fragments (LT4 and LT6) gave means for
13
C of 7.73
0.15 and 1.20 0.06, respectively.
For LA mass spectrometry, the sample tooth or fragment was placed
alongside the two enamel standards in a laser chamber, and purged with
helium, usually overnight. CO
2
was generated by ring a New Wave MIR 10
CO
2
laser in several bursts, whereafter the gas was cryogenically puried and
introduced to the continuous-ow gas chromatographyisotope ratio mass
spectrometry system (details are given in SI Text, including data and compar-
isons in Tables S2 and S3). All data were blank corrected. The two internal
enamel standards (LT4 and LT6), gave long-term
13
C means of 8.03 0.8
(n = 45) and 0.49 1.01 (n = 53), respectively. The data for unknowns were
calibrated using a regression calculated from the two standards, for each run.
All measurements were performed in the University of Bradford (Bradford,
UK) Stable Light Isotope Laboratory.
ACKNOWLEDGMENTS. We thank Andrew Gledhill for assistance with mass
spectrometry, Zachary Sharp for design and manufacture of the laser cell,
Ben Passey for advice on laser ablation, and S. Riffaut, Amy Jeffrey, and
Tom Frankland for drawing the gures. We thank Stanley Ulijaszek, the
Editor, Stanley Ambrose, and two anonymous referees for their helpful
comments. Research was funded by the Universities of Bradford, Oxford,
and Poitiers [Sciences Fondamentales et Appliques- Centre National de
la Recherche Scientique (CNRS), Institut de Paloprimatologie, Palonto-
logie Humaine: Evolution et Paloenvironnements, Unit Mixte de
Fig. 3. Box-and-whiskers plot of the
13
C data for all probable hominins
(excluding Homo) published to date. The high C
3
and C
4
contributions to the
diets of Ar. ramidus and P. boisei, respectively, mark out distinctive diets
toward both ends of the C
3
to C
4
spectrum. A. africanus data stand out for
their high variability, which is atypical among fossil and extant mammals.
Data sources are ref. 13 for Ar. ramidus, the South African australopith data as
summarized in ref. 8, and the P. boisei data in refs. 11 and 12.
Lee-Thorp et al. PNAS Early Edition | 3 of 4
A
N
T
H
R
O
P
O
L
O
G
Y
Recherche 7262] and Universit et Centre National dAppui la Recherche,
NDjamena, Chad, Ministre des Affaires Etrangres (Commission des Fouilles
Paris and Service de Coopration et dAction Culturelle Ambassade de
France NDjamena), the CNRS/Institut cologie et Environnement and
ANR-09-BLAN-0238 project, and the National Science Foundation (Reveal-
ing Hominids Origins Initiative).
1. Ungar PS, Scott RS, Grine FE, Teaford MF (2010) Molar microwear textures and the
diets of Australopithecus anamensis and Australopithecus afarensis. Philos Trans R Soc
Lond B Biol Sci 365(1556):33453354.
2. Ungar PS, Sponheimer M (2011) The diets of early hominins. Science 334(6053):
190193.
3. Lee-Thorp JA, van der Merwe NJ, Brain CK (1994) Diet of Australopithecus robustus at
Swartkrans deduced from stable carbon isotope ratios. J Hum Evol 27:361372.
4. Sponheimer M, Lee-Thorp JA (1999) Reconstructing the diet of the early hominid
Australopithecus africanus using
13
C/
12
C analysis. Science 283(5400):368370.
5. Lee-Thorp JA, Thackeray JF, van der Merwe NJ (2000) The hunters and the hunted
revisited. J Hum Evol 39(6):565576.
6. van der Merwe NJ, Thackeray JF, Lee-Thorp JA, Luyt J (2003) The carbon isotope
ecology and diet of Australopithecus africanus at Sterkfontein, South Africa. J Hum
Evol 44(5):581597.
7. Sponheimer M, et al. (2005) Hominins, sedges, and termites: New carbon isotope data
from the Sterkfontein valley and Kruger National Park. J Hum Evol 48(3):301312.
8. Sponheimer M, et al. (2006) Isotopic evidence for dietary variability in the early
hominin Paranthropus robustus. Science 314(5801):980982.
9. Lee-Thorp JA, Sponheimer M, Passey BH, de Ruiter DJ, Cerling TE (2010) Stable iso-
topes in fossil hominin tooth enamel suggest a fundamental dietary shift in the Pli-
ocene. Philos Trans R Soc Lond B Biol Sci 365(1556):33893396.
10. van der Merwe NJ, Masao FT, Bamford MK (2008) Isotopic evidence for contrasting
diets of early hominins Homo habilis and Australopithecus boisei of Tanzania. S Afr J
Sci 104:153155.
11. Cerling TE, et al. (2011) Diet of Paranthropus boisei in the early Pleistocene of East
Africa. Proc Natl Acad Sci USA 108(23):93379341.
12. White TD, et al. (2009) Macrovertebrate paleontology and the Pliocene habitat of
Ardipithecus ramidus. Science 326(5949):8793.
13. Brunet M, et al. (1995) The rst australopithecine 2,500 kilometres west of the Rift
Valley (Chad). Nature 378(6554):273275.
14. Brunet M, et al. (1996) Australopithecus bahrelghazali, une novelle espece dHomi-
nide ancient de la region de Koro Toro. C R Acad Sci Paris 322(IIa):907913.
15. Brunet M, et al. (1997) Tchad: Un nouveau site Hominids Pliocne. C R Acad Sci
Paris 324(IIa):341345.
16. Lebatard AE, et al. (2008) Cosmogenic nuclide dating of Sahelanthropus tchadensis
and Australopithecus bahrelghazali: Mio-Pliocene hominids from Chad. Proc Natl
Acad Sci USA 105(9):32263231.
17. Geraads D, Brunet M, Mackaye HT, Vignaud P (2001) Pliocene Bovidae (Mammalia)
from the Koro Toro Australopithecine sites, Chad. J Vertebr Paleontol 21:335346.
18. Reed KE (1997) Early hominid evolution and ecological change through the African
Plio-Pleistocene. J Hum Evol 32(23):289322.
19. Fara E, Likius A, Mackaye HT, Vignaud P, Brunet M (2005) Pliocene large-mammal
assemblages from Northern Chad: Sampling and ecological structure. Natur-
wissenschaften 92(11):537541.
20. Zazzo A, et al. (2000) Herbivore paleodiet and paleoenvironmental changes at Chad
during the Pliocene using stable isotope ratios of tooth enamel carbonate. Paleobi-
ology 26:294309.
21. Andrews PJ, Humphrey L (1999) African Miocene environments and the transition to
early hominines. African Biogeography, Climate Change and Human Evolution, eds
Bromage T, Schrenk F (Oxford Univ Press, Oxford), pp 282315.
22. Reed KE (2008) Paleoecological patterns at the Hadar hominin site, Afar Regional
State, Ethiopia. J Hum Evol 54(6):743768.
23. Su DF, Harrison T (2007) The paleoecology of the Upper Laetolil Beds at Laetoli;
a reconsideration of the large mammal evidence. Hominin Environments in the East
African Pliocene: An Assessment of the Faunal Evidence, eds Bobe R, Alemseged Z,
Behrensmeyer AK (Springer, Dordrecht, The Netherlands), pp 279313.
24. Schoeninger MJ, Moore J, Sept JM (1999) Subsistence strategies of two savanna
chimpanzee populations: The stable isotope evidence. Am J Primatol 49(4):297314.
25. Sponheimer M, et al. (2006) Do savanna chimpanzees consume C
4
resources? J Hum
Evol 51(2):128133.
26. Alberts SC, et al. (2005) Seasonality in Primates: Studies of Living and Extinct Human
and Non-Human Primates, eds Brockman DK, van Schaik CP (Cambridge Univ Press,
Cambridge, UK), pp 157195.
27. Codron D, Lee-Thorp JA, Sponheimer M, de Ruiter D, Codron J (2008) What insights
can baboon feeding ecology provide for early hominin niche differentiation? Int J
Primatol 29:757772.
28. Peters CR, Vogel JC (2005) Africas wild C
4
plant foods and possible early hominid
diets. J Hum Evol 48(3):219236.
29. Tanaka J (1980) The San, hunter-gatherers of the Kalahari: A study in ecological
anthropology (Univ of Tokyo Press, Tokyo).
30. Sage RF (2004) The evolution of C
4
photosynthesis. New Phytol 161:341370.
31. Hesla ABI, Tieszen LL, Imbamba SK (1982) A systematic survey of C
3
and C
4
photo-
synthesis in the Cyperaceae of Kenya, East Africa. Photosynthetica 16:196205.
32. Stock WD, Chuba DK, Verboom GA (2004) Distribution of South African C
3
and C
4
species of Cyperaceae in relation to climate and phylogeny. Austral Ecol 29:313319.
33. Jolly CJ (1970) The seed-eaters: A new model of hominid differentiation based on
a baboon analogy. Man (Lond) 5:526.
34. Dominy NJ, Vogel ER, Yeakel JD, Constantino Lucas PW (2008) Mechanical properties
of plant underground storage organs and implications for dietary models of early
hominins. Evol Biol 35:159175.
35. Post DG (1982) Feeding behaviour of Yellow baboons (Papio cynocephalus) in the
Amboseli National Park, Kenya. Int J Primatol 3(4):403430.
36. Barton RA, Whiten A, Byrne RW, English M (1993) Chemical composition of baboon
plant foods: Implications for the interpretation of intra- and interspecic differences
in diet. Folia Primatol (Basel) 61(1):120.
37. Altmann SA, Post DG, Klein DF (1987) Nutrients and toxins of plants in Amboseli,
Kenya. Afr J Ecol 25:279293.
38. Swedell L, Hailemeskel G, Schreier A (2008) Composition and seasonality of diet in wild
Hamadryas baboons: Preliminary ndings from Filoha. Folia Primatol (Basel) 79(6):476490.
39. Conklin-Brittain NL, Wrangham RW, Smith CC (2002) A two-stage model of increased
dietary quality in early hominid evolution: The role of ber. Human Diet: Its Origin
and Evolution, eds Ungar PS, Teaford MF (Bergin & Garvey, Westport, CT), pp 6176.
40. Laden G, Wrangham R (2005) The rise of the hominids as an adaptive shift in fallback
foods: Plant underground storage organs (USOs) and australopith origins. J Hum Evol
49(4):482498.
41. Wrangham RW, Cheney D, Seyfarth R, Sarmiento E (2009) Shallow-water habitats as
sources of fallback foods for hominins. Am J Phys Anthropol 140(4):630642.
42. Lemoalle J (2005) African wetlands in a semi arid climate: The Chad Basin as an ex-
ample. The Worlds Largest Wetlands: Ecology and Conservation, eds Fraser LH, Keddy PA
(Cambridge Univ Press, Cambridge, UK), pp 316346.
43. Simpson DA, Inglis CA (2001) Cyperaceae of economic, ethnobotanical and horticul-
tural importance: A checklist. Kew Bull 56:257360.
44. Youngblood D (2004) Identication and quantication of edible plant foods in the
Upper (Nama) Karoo, South Africa. Econ Bot 58(Supplement):S43S65.
45. Ambrose SH, DeNiro M (1986) The isotopic ecology of East African mammals. Oeco-
logia 69:395406.
46. Levin NE, Simpson SW, Quade J, Cerling TE, Frost SR (2008) Herbivore enamel
carbon isotope composition and the environmental context of Ardipithecus at
Gona, Ethiopia. The Geology of Early Humans in the Horn of Africa, eds Quade J,
Wynn JG (The Geological Society of America, Boulder, CO), Special Paper 446, pp
215234.
47. Pascual B, Maroto JV, Lpez-Galarza S, Sanbautista A, Alagarda J (2000) Chufa
(Cyperus esculentus L. var. sativus Boeck.): An unconventional crop. Studies related to
applications and cultivation. Econ Bot 54(4):439448.
48. Ungar PS, Grine FE, Teaford MF (2008) Dental microwear and diet of the Plio-Pleistocene
hominin Paranthropus boisei. PLoS One 3(4):e2044.
49. Passey BH, Cerling TE (2006) In situ stable isotope analysis (
13
C,
18
O) of very small
teeth using laser ablation GC/IRMS. Chem Geol 235:238249.
4 of 4 | www.pnas.org/cgi/doi/10.1073/pnas.1204209109 Lee-Thorp et al.
Supporting Information
Lee-Thorp et al. 10.1073/pnas.1204209109
SI Methods
In all cases, the tooth or fragment was cleaned with a soft brush to
remove loose sediment if present, swabbed with acetone, and then
air-dried.
For standard acid hydrolysis mass spectrometry, enamel powder
obtained by gentle abrasion with a diamond-tipped drill along
the length of a broken or nonocclusal surface was pretreated
to remove possible organic and carbonate contaminants. This
consisted of a series of washes in 1.5% (vol/vol) sodium hypo-
chlorite solution and then 0.1 M acetic acid, for 30 and 10 min,
respectively, interspersed with rinsing in puried H
2
O and
centrifuging (1). CO
2
was produced from 1.5 mg of enamel
powder by 100% H
3
PO
4
acid hydrolysis at 72 C in a GasBench
II, and introduced via a continuous-ow gas chromatograph
(GC) for measurement in a Thermo Finnigan Delta V. Precision
estimated by replicate measurements of NBS 19, and the Car-
raraZ interlaboratory standard, was 0.1 for both
13
C/
12
C and
18
O/
16
O. Replicate (n = 10) measurements of each of the fossil
enamel (bioapatite) standards, LT4 and LT6, indicated slightly
lower precision for LT4 (see below).
For laser ablation (LA) determinations, the sample tooth was
placed alongside two enamel fragment standards (LT4 and LT6)
in a 23-mm-diameter laser chamber with a ZnSe window, and
purged with helium (He) until outgassing ceased. Small amounts
(1030 nmol) of CO
2
were generated by four to ve laser shots
red using a New Wave MIR 10 CO
2
laser (10.6 m) operating
at 100 W and 5% power for 40-ms pulse duration (equivalent to
0.13 mJ), in a He atmosphere. The resultant gas was cryogeni-
cally puried before introduction of CO
2
to the continuous-ow
GCisotope ratio mass spectrometer. Each sample was mea-
sured in a short transect of four to ve scans (each consisting of
four to ve adjacent shots), spaced 0.5 mm apart, so that several

13
C values were generated for each specimen (Tables S2 and
S3). Measurements were blank corrected. Systematic isotope
fractionation and fractionation associated with laser ablation
production of CO
2
was monitored by coanalysis of the two in-
ternal tooth enamel standards in each run. LA mass spectrom-
etry determinations are less precise than acid hydrolysis mass
spectrometry (2), and additionally measurements along an
enamel transect are more variable because enamel increments
themselves are isotopically variable, compared with powdered
bulk enamel. This applies also to the two enamel fragments used
as standards. However, it was necessary to use enamel fragments
because of the sensitivity of LA to the sample material. Com-
parison of the
13
C results obtained by both methods for the
two enamel standards, and three KT faunal specimens, showed
agreement within these constraints (see Table S3 for the KT data).
Multiple LA analyses of
13
C LT4 (X =8.03 0.80; n = 45) and
LT6 (X = 0.49 1.01; n = 53) over 6 mo gave values similar
to the acid hydrolysis data (X =7.73 0.15 and 1.20 0.06;
n = 10). Laser-carbonate apparent fractionation (*
LASER-carb
) was
estimated using a regression comparison calculated from these two
enamel standards in each run. All measurements were carried out
in the Bradford University Stable Light Isotope Laboratory.
1. Sponheimer M, et al. (2005) Hominins, sedges, and termites: New carbon isotope data
from the Sterkfontein valley and Kruger National Park. J Hum Evol 48(3):301312.
2. Passey BH, Cerling TE (2006) In situ stable isotope analysis (
13
C,
18
O) of very small
teeth using laser ablation GC/IRMS. Chem Geol 235:238249.
Lee-Thorp et al. www.pnas.org/cgi/content/short/1204209109 1 of 3
Table S1. Summary of the
13
C and
18
O data obtained for KT
faunal enamel from sites KT12 and KT13 used in Fig. 2
Reference no. Site Species
13
C
18
O
KT12/H1a KT12 A. bahrelghazali 4.4
KT12 P3/H2 KT12 A. bahrelghazali 0.8
KT13-96-H1 KT13 A. bahrelghazali 2.5 32.8
TC9000* KT13 Alcelaphini 0.1 36.6
TC6100* KT4 Alcelaphini 0.6 32.9
TC5600* KT12 Alcelaphini 1.8 35.8
TC10400* KT12 Alcelaphini 1.8 35.6
TC6400* KT4 Alcelaphini 1.9 34.4
TC8900* KT13 Alcelaphini 2.2 35.3
TC400* KT13 Hipparion cf. hasumense 0.2 30.2
TC200* KT12 Hipparion cf. hasumense 1.7 38.7
KT13-96-447 KT13 Equidae 1.6 29.1
KT13-96-515 KT13 Equidae 0.7 30.7
KT12-95-016 KT12 Equidae 0.3 30.9
KT13-96-379 KT13 Equidae 0.5 33.9
KT13-96-516 KT13 Equidae 1.3 31.6
KT12-95-007 KT12 Equidae 1.4 29.9
TC7500* KT12 Kolpochoerus afarensis 1.2 30.1
TC7300* KT12 Notochoerus euilus 0.9 33.8
TC7600* KT13 Notochoerus euilus 1.3 33.7
KT12-98 (1) KT12 Suidae 0.6 30.0
KT12-98 (2) KT12 Suidae 1.3 29.4
KT12-96-048 KT12 Suidae 1.5 26.9
KT13-96-149 KT13 Suidae 1.3 27.8
KT12-98 No # S KT12 Suidae 1.3 29.4
KT12-96-046 KT12 Suidae 0.3 30.5
KT12-96-050 KT12 Suidae 0.6 30.5
KT12-98 No # KT12 Suidae 2.2 30.0
TC7700* KT13 Loxodonta sp. 0.1 29.7
TC7800* KT8 Loxodonta sp. 0.6 29.5
TC7900* KT9 Loxodonta sp. 0.5 31
TC12600* KT Stegodon kaisensis 1.7 29.1
TC10600* KT13 Reduncini 4.5 32.9
TC10500* KT13 Reduncini 0.4 33.9
TC10550* KT13 Reduncini 1.8 31.1
TC5400* KT12 Bovini 5.4 36.9
TC8800* KT12 Bovini 1.3 35.3
TC10700* KT12 Bovini 1.1 33.5
TC12100* KB3 Paracamelus sp. 11.8 35.7
TC12300* KB3 Paracamelus sp. 10.5 33.8
TC12200* KB3 Paracamelus sp. 10.4 33.9
TC12500* KB3 Diceros bicornis 10.7 31.5
TC10300* KB26 Giraffa sp. 10.2 34.4
TC10200* KB3 Giraffa sp. 9.8 31.6
Data are from the present study, whereas those marked an asterisk (*) are
from ref. 1. Some values are reported for KT sites other than KT12 and 13, and
six specimens at the base of the table are from the older site of Kossom Bou-
goudi (1), shown here because no suitable browser tooth enamel from the KT
sites was available for isotopic measurement.
13
C values are reported relative
to VPDB, and
18
O values relative to VSMOW.
18
O is reported only for samples
determined by acid hydrolysis, and where samples were determined by both LA
and acid hydrolysis, only the latter
13
C values are reported in this table.
1. Zazzo A, et al. (2000) Herbivore paleodiet and paleoenvironmental changes at Chad during the Pliocene using stable isotope ratios of tooth enamel carbonate. Paleobiology 26:
294309.
Lee-Thorp et al. www.pnas.org/cgi/content/short/1204209109 2 of 3
Table S2.
13
C values obtained from the laser ablation transects
(labeled as consecutive scan number) for two A. bahrelghazali
specimens, KT12 P3/H2 and KT12/H1a
Reference no. Tooth Scan no.
13
C
KT12 P3/H2 P
3
1 1.8
2 0.6
3 0.3
4 1.1
Mean SD 0.80 0.87
KT12/H1a I
2
1 5.3
2 3.7
3 3.4
4 5.3
5 5.0
Mean SD 4.36 0.93
The sample of KT12/H1a measured here was an extremely tiny, previously
dislodged, enamel sliver. KT12 P3/H2 and KT12/H1 were reported in refs. 13.
The results for KT13-96-H1 are reported in the composite data (Table S1).

13
C values are reported relative to VPDB.
Table S3. Comparison of
13
C values obtained by LA compared
with those obtained by standard GasBench acid hydrolysis on
bulk enamel powder for the teeth of two KT suids and one equid
Reference no. Species
GasBench
LA method

13
C Mean
13
C SD No. scans
KT13-96-379 Equidae 0.5 0.65 0.09 4
KT12-98 Suidae 0.6 0.85 0.23 4
Individual 1
KT12-98 Suidae 1.3 0.31 1.42 5
Individual 2
Some variability along laser scans is expected but any possible seasonal
variability cannot be determined given the small surface area of the tooth
crown analyzed.
18
O values are not reported here because the two methods
measure different bioapatite componentsin particular, laser ablation measures
all oxygen-containing ions within the apatite (PO
4
, CO
3
, OH).
13
C values are
reported relative to VPDB.
1. Brunet M, et al. (1995) The rst australopithecine 2,500 kilometres west of the Rift Valley (Chad). Nature 378(6554):273275.
2. Brunet M, Beauvilian A, Coppens Y, Heintz E, Moutaye AHE, et al. (1996) Australopithecus bahrelghazali, une novelle espece dHominide ancient de la region de Koro Toro. C R Acad Sci
Paris 322(IIa):907913.
3. Brunet M, Beauvilian A, Geraads D, Guy F, Kasser M, et al. (1997) Tchad: Un nouveau site Hominids Pliocne. C R Acad Sci Paris 324(IIa):341345.
Lee-Thorp et al. www.pnas.org/cgi/content/short/1204209109 3 of 3