Expanded geographic distribution and dietary strategies of the earliest Oldowan hominins and Paranthropus

Abstract

The oldest Oldowan tool sites, from around 2.6 million years ago, have previously been confined to Ethiopia's Afar Triangle. We describe sites at Nyayanga, Kenya, dated to 3.032 to 2.581 million years ago and expand this distribution by over 1300 kilometers. Furthermore, we found two hippopotamid butchery sites associated with mosaic vegetation and a C4 grazer-dominated fauna. Tool flaking proficiency was comparable with that of younger Oldowan assemblages, but pounding activities were more common. Tool use-wear and bone damage indicate plant and animal tissue processing. Paranthropus sp. teeth, the first from southwestern Kenya, possessed carbon isotopic values indicative of a diet rich in C4 foods. We argue that the earliest Oldowan was more widespread than previously known, used to process diverse foods including megafauna, and associated with Paranthropus from its onset.

Full text

HOMININ EVOLUTION
Expanded geographic distribution and dietary strategies
of the earliest Oldowan hominins and Paranthropus
Thomas W. Plummer
1,2,3,4
*, James S. Oliver
5
,EmmaM.Finestone
6,7
, Peter W. Ditchfield
8
,
Laura C. Bishop
9,10
, Scott A. Blumenthal
4,11,12
, Cristina Lemorini
13
,IsabellaCaricola
13,14
,SharaE.Bailey
3,15
,
Andy I. R. Herries
16,17
, Jennifer A. Parkinson
4,18
, Elizabeth Whitfield
9
, Fritz Hertel
19
,
Rahab N. Kinyanjui
4,20,21
,ThomasH.Vincent
9
, Youjuan Li
22,23
, Julien Louys
24
, Stephen R. Frost
11
,
David R. Braun
25,26
,JonathanS.Reeves
26
, Emily D. G. Early
4,27
,BlastoOnyango
20
, Raquel Lamela-Lopez
2,3
,
Frances L. Forrest
28,29
,HuaiyuHe
30
, Timothy P. Lane
9
,MarineFrouin
31
, Sébastien Nomade
32,33
,
Evan P. Wilson
2,3
, Simion K. Bartilol
34
, Nelson Kiprono Rotich
35
, Richard Potts
4,20
The oldest Oldowan tool sites, from around 2.6 million years ago, have previously been confined
to Ethiopia’s Afar Triangle. We describe sites at Nyayanga, Kenya, dated to 3.032 to 2.581 million
years ago and expand this distribution by over 1300 kilometers. Furthermore, we found two
hippopotamid butchery sites associated with mosaic vegetation and a C
4
grazer–dominated fauna.
Tool flaking proficiency was comparable with that of younger Oldowan assemblages, but pounding
activities were more common. Tool use-wear and bone damage indicate plant and animal tissue
processing. Paranthropus sp. teeth, the first from southwestern Kenya, possessed carbon isotopic
values indicative of a diet rich in C
4
foods. We argue that the earliest Oldowan was more widespread
than previously known, used to process diverse foods including megafauna, and associated with
Paranthropus from its onset.
The appearance of Oldowan tools around
2.6 million years ago (Ma) was a techno-
logical breakthrough that used system-
atically produced, sharp-edged flakes for
cutting and cobbles or cores for percus-
sion (1). Although the Oldowan is often attrib-
uted to the genus Homo, multiple hominin
taxa overlapped temporally and geographical-
ly with these early tools, and it is possible that
other genera, such as Paranthropus,made
and/or used them. Some have linked emer-
gent Oldowan technology to the first access to
or more efficient processing of nutrient-rich
animal carcasses [for example, (2,3)]. Others
have argued that plant food processing was
the primary goal of early Oldowan stone tool
usage, with increased carnivory (and butchery
with stone tools) being added to the behav-
ioral repertoire after 2 Ma (4,5). The evolu-
tionary benefits connected with the emergence
of Oldowan technology are unclear because
of the paucity of late Pliocene Oldowan sites,
hitherto known only from the Afar Triangle of
Ethiopia at Gona and Ledi-Geraru, localities
found approximately 50 km away from each
other (6,7). In this study, we report 3.032–
2.595 Ma deposits at Nyayanga, Kenya, that
expand the geographic range of the earliest
Oldowanbymorethan1300kmandtherange
of Paranthropus by approximately 230 km to
southwestern Kenya. Archeological findings
demonstrate that hominins used tools to butch-
er a variety of animals, including megafauna,
and process diverse plants at the Oldowan’s
inception.
Nyayanga (0° 23.909'S, 34° 27.115'E) is an
archeological and paleontological locality on
the western shoreline of the Homa Peninsula
(Fig. 1A) [(8), section 1]. The peninsula is lo-
cated on the southern margin of the Winam
Gulf of Lake Victoria, within the east-west–
oriented Nyanza Rift between the two main
branches of the East African Rift System (9).
It is dominated by the Homa Mountain car-
bonatite complex, which on its flanks bears
alluvial, fluvial, and lacustrine sediments that
range in age from 6 Ma through the Holocene
(10–13). Sediments at Nyayanga are exposed
in a 40,000 m
2
amphitheater and a gully that
can be traced for 500 m upslope (Fig. 1). Ex-
cavations and surface collection focused on
the top half of the oldest bed (Fig. 1, NY-1),
which yielded Oldowan artifacts, Paranthropus
sp. fossils, and faunal fossils in overbank de-
posits from a westward-flowing paleochannel
[(8), section 2].
The age of the Nyayanga Beds is constrained
by (U-Th)/He dating of apatite crystals, mag-
netostratigraphy, lithostratigraphic correlation
with the Rawi Fm (11,13) deposited north of
Homa Mountain, and biostratigraphy. The
(U-Th)/He dating of apatite crystals yielded
ages of 2.87 ± 0.79 Ma and 2.98 ± 0.50 Ma from
two tuffaceous silts in NY-1 (fig. S1 and table
S1) [(8), section 3]. Magnetostratigraphic sam-
pling was carried out on the excavation 3 slope
andintrenches9and11(Fig.1)[(8), section 4].
The Nyayanga sequence shows reversed polar-
ity in unit A, intermediate-normal to normal
polarity from basal to middle NY-1, normal
polarity from middle NY-1 through NY-2, in-
termediate normal polarity in the base of NY-3,
and reversed polarity at the top of NY-3 (Fig. 1,
fig. S2, and table S2). The (U-Th)/He apatite
crystal dates suggest that the normal interval
corresponds to the C2An.1n Subchron between
3.032and2.595Ma(14). This is similarly in-
dicated by the lithostratigraphic correlation of
the Nyayanga Beds with the Rawi Fm (11,13),
which was also deposited during the C2An.1n
Subchron. Biostratigraphy [(8), section 6]
is consistent with a late Pliocene age, includ-
ing more archaic examples of two pig species,
the suine Metridiochoerus andrewsi and the
tetraconodont Notochoerus cf. scotti, than the
nearby locality of Kanjera South from 2 Ma
(13), as well as an equid sample composed ex-
clusively of hipparionin (Eurygnathohippus
sp.) fossils. The latter indicates a date earlier
than the 2.3 Ma dispersal of Equus across
Africa (15). Acknowledging the wide 1suncer-
tainty, the combination of (U-Th)/He apatite
RESEARCH
Plummer et al., Science 379, 561–566 (2023) 10 February 2023 1of6
1
Department of Anthropology, Queens College, Flushing, NY, USA.
2
The CUNY Graduate Center, New York, NY, USA.
3
New York Consortium in Evolutionary Primatology, New York, NY, USA.
4
Human Origins Program, National Museum of Natural History, Smithsonian Institution, Washington, DC, USA.
5
Anthropology Section, Illinois State Museum Springfield, IL, USA.
6
The Cleveland
Museum of Natural History, Cleveland, OH, USA.
7
Department of Archeology, Max Planck Institute for the Science of Human History, Jena, Germany.
8
School of Archeology, University of
Oxford, Oxford, UK.
9
Research Centre in Evolutionary Anthropology and Paleoecology, School of Biological and Environmental Sciences, Liverpool John Moores University, Liverpool, UK.
10
The
Sino-British College, University of Shanghai for Science and Technology, Shanghai, China.
11
Department of Anthropology, University of Oregon, Eugene, OR, USA.
12
University of British Columbia,
Vancouver, BC, Canada.
13
LTFAPA Laboratory, Department of Classics, Sapienza University of Rome, Rome, Italy.
14
School of History, Classics and Archeology, Newcastle University, Newcastle
upon Tyne, UK.
15
Department of Anthropology, New York University, New York, NY, USA.
16
The Australian Archaeomagnetism Laboratory, Dept. Archaeology and History, La Trobe University,
Melbourne Campus, 3086 Bundoora, Victoria, Australia.
17
Paleo-Research Institute, University of Johannesburg, Johannesburg, South Africa.
18
Department of Anthropology, University of San
Diego, San Diego, CA, USA.
19
Department of Biology, California State University, Northridge, CA, USA.
20
National Museums of Kenya, Nairobi, Kenya.
21
Max Planck Institute for Geoanthropology,
Jena, Germany.
22
Department of Geoscience, University of Wisconsin-Madison, Madison, WI, USA.
23
Institute of Geology, China Earthquake Administration, Beijing, China.
24
Australian Research
Centre for Human Evolution, Griffith University, Brisbane, Australia.
25
Department of Anthropology, George Washington University, Washington, DC, USA.
26
Max Planck Institute for Evolutionary
Anthropology, Leipzig, Germany.
27
Arizona Museum of Natural History, Mesa, AZ, USA.
28
Department of Sociology and Anthropology, Fairfield University, Fairfield, CT, USA.
29
American Museum
of Natural History, New York, NY, USA.
30
Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China.
31
Department of Geosciences, Stony Brook University, Stony Brook,
NY, USA.
32
Laboratoire des Sciences du Climat et de l’Environnement, Université de Versailles Saint-Quentin and Paris Saclay, Gif Sur Yvette, France.
33
Institut Pierre Simon Laplace, Guyancourt,
France.
34
Institute of Nuclear Science and Technology, University of Nairobi, Nairobi, Kenya.
35
Institute of Nuclear Chemistry and Technology, Warsaw, Poland.
*Corresponding author. Email: thomas.plummer@qc.cuny.edu

Plummer et al., Science 379, 561–566 (2023) 10 February 2023 2of6
Fig. 1. Nyayanga gully, stratigraphy, magnetostratigraphic data, and apatite crystal dating results. (A) Topographic map of the gully system showing the
locations of geologic trenches and excavations. (B) Composite stratigraphic column of the Nyayanga Beds showing the stratigraphic placement of excavations 3 and
5 and the magnetostratigraphic profile. Reversed polarity is shown in white, intermediate normal polarity in gray, and normal polarity in black.
RESEARCH |RESEARCH ARTICLE

crystal dates, biostratigraphy, and the transi-
tional nature of the magnetostratigraphy of
lower NY-1 supports deposition early in the
temporal range of the C2An.1n Subchron.
We recovered 330 artifacts from the upper
half of NY-1; 135 were recovered in situ from
excavations 3 and 5, and 195 were recovered
from the surface [(8), section 7]. The overall
technological attributes of tools, such as core
and flake sizes and the number of flake scars
on cores, are similar to other Oldowan as-
semblages (Fig. 2A and table S4). Nyayanga
hominins efficiently removed flakes from cores
using unifacial, bifacial, and multifacial reduc-
tion that is also comparable with technology
at other Oldowan localities (7). The presence of
cortical flakes and hammerstones with batter-
ing damage is consistent with on-site flake
production through hard hammer percussion.
Artifacts were manufactured from a diverse
array of raw materials, including rhyolite, quartz-
ite, and quartz. The Nyayanga assemblage is
distinct in containing a high frequency of cores
(20.6%, n= 68) (Fig. 2B) and a large percent-
age of artifacts preserving evidence of percus-
sive activities (7.0%, n= 23) (Fig. 2C).
A total of 1776 bones were recovered in situ
from NY-1 in excavation 3 (n= 1580) and ex-
cavation 5 (n=196).Themostcommontaxain
excavations 3 and 5 are hippopotamids [57.1
and 61.9% of the number of identified speci-
mens (NISP), respectively] followed by bovids
(19.2 and 22.2% of the NISP, respectively) (fig. S4
Plummer et al., Science 379, 561–566 (2023) 10 February 2023 3of6
Fig. 2. Oldowan artifact technological analysis. (A) Photos of a dorsal flake
(Exc3-1475), ventral flake (Exc3-1413), and core (NY17-128) from Nyayanga next to
a principal component analysis based on major technological attributes of Early
Stone Age artifact assemblages and a capuchin-derived assemblage (table S4)
[(8), section 9]. Assemblages are plotted according to principal component 1
(xaxis) and principal component 2 (yaxis). The Nyayanga assemblage (NYA) falls
within the shaded ellipse that represents the 95% confidence interval for Oldowan
sites. A scree plot with eigenvalue percentage of variance for principal components 1
through 11, and a loadings plot showing the contribution of each variable are shown
to the right. (B) The frequency of cores in other Oldowan assemblages compared with
those of Nyayanga (table S4) and photos of two Nyayanga cores (NY17-54, top;
NY17-55, bottom). (C) The frequency of artifacts with percussion damage in other
Oldowan assemblages compared with those of Nyayanga (table S4) with photos of a
Nyayanga pounded piece (Exc3-485).
RESEARCH |RESEARCH ARTICLE

and tables S5 and S6) [(8), section 8]. The high
in situ frequencies of animals preferring near-
water habitats [such as hippopotamids, turtles,
crocodilians, and cane rats (Thryonomyidae)]
reflect a riparian depositional context. Bone
surface preservation was highly variable, but
more than 85% of the sample in both exca-
vations showed no or minimal weathering,
which is consistent with the results of rapid
burial by fluvial sediments (fig. S7).
Hippopotamid butchery is documented in
both excavation 3 and excavation 5. A mini-
mum of two hippopotamid individuals were
recovered from excavation 3 (fig. S5). The more
complete individual is composed of 241 bone
fragments from across the skeleton, including
a large axial bone concentration likely mark-
ing its death site. Stone tools (n= 42) were
closely associated with the skeleton, includ-
ing several tools recovered in direct physical
contact with hippopotamid bones. Despite the
varied bone preservation, one hippopotamid rib
fragment exhibits a deep cutmark with clearly
preserved internal striations (Fig. 3B), and three
stone flakes (detached pieces) exhibit use-wear
indicative of butchery.
In excavation 5, 39 hippopotamid bones, like-
ly from a single individual, were found spatially
associated with 14 artifacts (fig. S6). One cluster
of bones consisted of girdle elements (scapula,
innominate), appendicular elements (proximal
half of tibia, calcaneum), a flake, and a split
cobble with percussion damage. The anterior
tuberosity of the tibia has a series of four short,
parallel cutmarks (Fig. 3A). A second cluster of
bones, located 2 m away, consists of a broken
humerus, a flake, a rib fragment, and a man-
uport. The nonanatomical placement of these
bones, some with hominin damage, and asso-
ciated artifacts (one with use-wear indicative
of butchery) in a fine silt suggests that the
bones may have been moved by hominins
while butchering the carcass.
Tool-damaged bones of nonhippopotamid
taxa were also found in the excavations and
Plummer et al., Science 379, 561–566 (2023) 10 February 2023 4of6
Fig. 3. Stone tool–damaged fossilized bones from Bed NY-1. (A) Hippopotamid
tibia (Exc5-170, proximal end oriented to left) displaying a series of identically
oriented cut marks with striae on anterior tibial crest. (B) Cut mark on a
hippopotamid rib (Exc3-1685) displaying striae and a concretion filling the
middle of the mark. (C) Parallel cut marks extending along the spine of size
three bovid scapula NY17-1. (D) A series of parallel cut marks (top) as well as
percussion load points and flake scars created during marrow processing
(bottom) visible on a size three bovid long bone shaft fragment (NY15-61).
RESEARCH |RESEARCH ARTICLE

eroding out of NY-1 in the amphitheater. A
size three bovid scapular spine fragment with
cut marks was found eroding from an in-situ
context at about the same level as the exca-
vation 5 hippopotamid (Fig. 3C). Other bones
from NY-1 with cut marks or percussion dam-
age show that hominins were consuming
both meat and marrow (Fig. 3D), a finding
supported by use-wear analysis. Overall fre-
quencies of hominin damage from the exca-
vations are low: 0.9 and 1.9% at excavations
3 and 5, respectively (table S7). In part, this
reflects poor surface preservation of many of
the fossils as well as fragmentation of ribs.
Use-wear observed on 30 quartz, quartzite,
granite, carbonatite, and rhyolite tools from
NY-1 confirm hominin processing of faunal
remains and plant tissue [(8), section 9]. Use-
wear found on six pounded pieces (16)and
17 flaked pieces (cores) show macro- and mi-
crotraces related to pounding activities (fig.
S8). Percussive stone tools were heavily used,
showing deep pits and developed polishes
and striations at low andhighmagnification,
which, based on modern experiments, require
at least several hours of use to emerge. On the
basis of experimental analogs (figs. S9 and S10
and tables S8 and S10), the quartzite and
rhyolite Oldowan percussive tools at Nyayanga
were used to process soft (such as soft tubers,
vegetables, or fruits) and hard (such as fibrous
tubers or woody parts) plant tissues (fig. S11
and table S11). Macro- and microtraces related
to cutting and scraping on six detached pieces
(flakes) and one flaked piece show that similar
materials were being cut and pounded (fig. S12
and tables S9 and S12). Five quartz detached
pieces from excavation 3 show traces indica-
tive of underground storage organ, wood, and
animal processing. A rhyolite flaked piece from
excavation 5 and a NY-1 surface collected de-
tached piece also have use-wear related to
butchery (fig. S12 and table S12).
Stable carbon isotopic analysis of pedogenic
carbonates, dietary reconstruction by using
tooth enamel isotopes, and bovid taxonomic
frequencies indicate that hominin activities
took place in a wooded grassland to grassy
woodland, bushland, or shrubland along a
stream channel within a mesic savanna biome
characterized by an abundance of C
4
grasses
and herbaceous plants (figs. S13 to S17 and
tables S13 and S14) [(8), sections 10 to 12].
Similar C
4
grazer–dominated ecosystems are
documented at the Ethiopian sites of Ledi-
Geraru (~2.8 Ma) (17) and Mille-Logya (~2.8 to
~2.4 Ma) (18), indicating that early represen-
tatives of both Paranthropus and Homo were
found in substantially open ecosystems. The
riparian setting, nearby freshwater spring, and
ecotone with open habitats provided Nyayanga
hominins with a diverse array of plant and
animal foods, shelter, and potable water.
Two hominin individuals from Bed NY-1
are assigned to Paranthropus sp. (Fig. 4) [(8),
section 13]. KNM-NG 77315 is a relatively com-
plete left upper molar, probably M
2
(second
upper molar), from surface collection, with a
crown area above the range of that of P. boisei
and P. robustus samples (tables S17 and S18).
KNM-NG 77316 is a nearly complete lingual
portion of a left lower molar, probably M
1
(first
lower molar), found in situ in excavation 3,
spatially associated with Oldowan artifacts
and a butchered hippopotamid. The Nyayanga
Paranthropus teeth have an average d
13
C
enamel
value of –0.7 ± 0.4‰(Fig. 4C), which demon-
strates a heavy reliance on C
4
foods. Thus, the
emergence of C
4
specialist diets coincided with
the appearance of at least one major aspect of
robust masticatory morphology (large post-
canine teeth) relatively early in the evolution
of Paranthropus [as opposed to (19)].
Paranthropus molar KNM-NG 77316 from
the excavation 3 hippopotamid butchery site is
a clear association of a hominin fossil with arti-
facts, raising the possibility that Paranthropus
made and/or co-opted stone tools. Although
its skull anatomy was not preserved, Nyayanga
Paranthropus was megadont and had flat mo-
lars with poor shearing capability. However,
its specialized gnathic morphology may not have
precluded tool use. Extraoral cutting and pound-
ing with stone tools could have provided access
to carcasses and within bone nutrients, and
made plant and animal tissue easier to chew and
digest (20), potentially allowing Paranthropus
to expand its diet. Although not found at
Nyayanga, Homo was also present in eastern
Africa at about the time of Nyayanga deposition
(21), so the Nyayanga artifacts cannot be defin-
itively attributed to a specific hominin genus.
Deposits at Nyayanga dated between 3.032
and 2.581 Ma show that at its earliest onset the
Oldowan was geographically more widely dis-
persed than previously known, a finding con-
sistent with a recently described Oldowan site
from around 2.4 Ma in North Africa (22).
Nyayanga artifacts wereusedtocut,scrape,
and pound large mammal and plant tissue,
demonstrating that at their emergence Oldo-
wan tools were used in a variety of actions to
access a broad array of food types. By 2 Ma,
Oldowan sites are found from northern to
southern Africa in both grassy and wooded
habitats (23), suggesting that one of the key
attributes of the technology was the flexibility
to process foods with different physical prop-
erties in a diversity of habitats.
The behaviors preserved at Nyayanga are at
least 600,000 years older than prior evidence
of megafaunal carcass and plant processing
and substantially predate the increase in abso-
lute brain size documented in the genus Homo
Plummer et al., Science 379, 561–566 (2023) 10 February 2023 5of6
Fig. 4. Paranthropus finds from Nyayanga. (A)Paranthropus sp. left upper molar KNM-NG 77315 found
on the surface of NY-1. (B)Paranthropus sp. left lower molar KNM-NG 77316 found in situ in NY-1 in
excavation 3. (C) Tooth enamel d
13
C of Nyayanga hominins and previously published d
13
C data from eastern
African hominins (19,25–34). Pliocene hominin teeth from Woranso-Mille are identified as “hominini indet
(W-M).”Mid-Pliocene hominin teeth from Lomekwi and Lothagam in the Turkana Basin previously attributed
to K. platyops (28) are identified here as “hominini indet (L-L)”(35). Hominin teeth from other areas of
the Omo-Turkana Basin that cannot be confidently attributed to a genus are identified as “hominini indet.”
RESEARCH |RESEARCH ARTICLE

after 2 Ma (24).ThelatePlioceneexpanded
geography of the earliest Oldowan, and new
evidence of its use in diverse tasks amplifies
our understanding of the adaptive advantage
of early stone technology in hominin diet and
foraging ecology.
REFERENCES AND NOTES
1. N. Toth, K. Schick, Azania 53,3–39 (2018).
2. M. J. Rogers, S. Semaw, in Sourcebook of Paleolithic Transitions
(Springer, 2009), pp. 155–171.
3. I. Cáceres, N. Kandi, M. Sahnouni, Z. Harichane, J. van der Made,
Proceedings of the II Meeting of African Prehistory: Burgos
15–16 April, 2015 (Consorcio CENIEH, Burgos, Spain, 2017),
pp. 173–196.
4. M. Domínguez-Rodrigo, B. Martínez-Navarro, J. Hum. Evol. 62,
315–327 (2012).
5. E. Hovers, in Developments in Quaternary Sciences (Elsevier,
2012), vol. 16, pp. 51–68.
6. S. Semaw et al., Nature 385, 333–336 (1997).
7. D. R. Braun et al., Proc. Natl. Acad. Sci. U.S.A. 116, 11712–11717
(2019).
8. Materials and methods are available as supplementary
materials.
9. M. J. Le Bas, Carbonatite-Nephelinite Volcanism: an African
Case History (Wiley London, 1977).
10. A. K. Behrensmeyer et al., J. Hum. Evol. 29, 247–274
(1995).
11. P. Ditchfield, J. Hicks, T. Plummer, L. C. Bishop, R. Potts,
J. Hum. Evol. 36, 123–150 (1999).
12. T. Plummer, L. C. Bishop, P. Ditchfield, J. Hicks, J. Hum. Evol.
36, 151–170 (1999).
13. L. C. Bishop et al., in African Pa laeoecology and Huma n
Evolu tion , S. C. Reynolds, R. Bobe, Eds. (Cambridge Univ.
Press, 2022), pp . 360–375.
14. J. G. Ogg, in Geologic Time Scale 2020 (Elsevier, 2020),
pp. 159–192.
15. R. L. Bernor, M. J. Armour-Chelu, H. Gilbert, T. M. Kaiser,
E. Schulz, in Cenozoic mammals of Africa, L. Werdelin,
B. Sanders, Eds. (Univ. California Press, 2010),
pp. 685–722.
16. G. L. Isaac, J. W. K. Harris, E. M. Kroll, in Koobi Fora Research
Project, Volume 5: Plio-Pleistocene Archaeology (Clarendon
Press, 1997).
17. E. N. DiMaggio et al., Science 347, 1355–1359 (2015).
18. Z. Alemseged et al., Nat. Commun. 11, 2480 (2020).
19. J. G. Wynn et al., Proc. Natl. Acad. Sci. U.S.A. 117, 21978–21984
(2020).
20. K. D. Zink, D. E. Lieberman, Nature 531, 500–503
(2016).
21. B. Villmoare et al., Science 347, 1352–1355 (2015).
22. M. Sahnouni et al., Science 362, 1297–1301 (2018).
23. T. Plummer, Am. J. Phys. Anthropol. 125, 118–164 (2004).
24. S. C. Antón, R. Potts, L. C. Aiello, Science 345, (2014).
25. N. J. Van der Merwe, F. T. Masao, M. K. Bamford, S. Af. J. Sci.
104, 153–155 (2008).
26. T. D. White et al., Science 326,87–93 (2009).
27. T. E. Cerling et al., Proc. Natl. Acad. Sci. U.S.A. 108, 9337–9341
(2011).
28. T. E. Cerling et al., Proc. Natl. Acad. Sci. U.S.A. 110,
10501–10506 (2013).
29. M. Sponheimer et al., Proc. Natl. Acad. Sci. U.S.A. 110,
10513–10518 (2013).
30. N. E. Levin, Y. Haile-Selassie, S. R. Frost, B. Z. Saylor,
Proc. Natl. Acad. Sci. U.S.A. 112, 12304–12309 (2015).
31. J. R. Robinson, J. Rowan, C. J. Campisano, J. G. Wynn,
K. E. Reed, Nat. Ecol. Evol. 1, 159 (2017).
32. T. Lüdecke et al., Proc. Natl. Acad. Sci. U.S.A. 115,
13330–13335 (2018).
33. J. E. Martin, T. Tacail, J. Braga, T. E. Cerling, V. Balter,
Nat. Commun. 11, 3587 (2020).
34. S. Semaw et al., Sci. Adv. 6, eaaw4694 (2020).
35. M. M. Skinner, M. G. Leakey, L. N. Leakey, F. K. Manthi,
F. Spoor, J. Hum. Evol. 145, 102820 (2020).
ACKNOW LEDGME NTS
The authors thank the National Museums of Kenya and M. Kibunjia,
F. K. Manthi, J. Kibii, and E. Ndiema for support and acknowledge
Kenya government permission granted by the Ministry of Sports,
Culture, and the Arts and by NACOSTI permit P/14/7709/701.
We thank P. Onyango for bringing the Nyayanga exposures to our
attention. N. Todd assisted with elephant tooth identification.
Funding: Funding from the L.S.B. Leakey Foundation (award
35805), the National Science Foundation (award 1327047), the
Wenner-Gren Foundation (award 9428), and the Professional
Staff Congress City University of New York Research Award
Program (award 60589-00 48) to T.P., and funding from the
William H. Donner Foundation and the Peter Buck Fund for
Human Origins Research to R.P., is gratefully acknowledged.
Author contributions: T.W.P., L.C.B., E.M.F., J.S.O., B.O., and
R.P. contributed to excavation strategy and research design.
Magnetostratigraphy sampling was conducted by P.W.D. and
analysis of samples and magnetostratigraphic interpretation was
undertaken by A.I.R.H. Sampling for (U-Th)/He apatite crystal
dates was undertaken by H.H. and Y.L. carried out (U-Th)/He
analysis. T.P.L., S.N., and M.F. provided additional geochronological
analysis and contributed to the geochronological framework.
Lithostratigraphy and sedimentology was conducted by P.W.D.,
E.W., and T.H.V. Paleosol carbonates were sampled and
analyzed for stable isotopic composition by P.W.D. and S.A.B.:
P.W.D. interpreted paleosol carbonate results, and S.A.B. sampled
and carried out stable isotopic analysis of tooth enamel. B.O.,
T.W.P., J.S.O., E.M.F., L.C.B., R.L.-L., F.L.F., J.L., F.H., and J.A.P.
contributed to field direction and data collection during field work.
S.A.B., T.W.P., P.W.D., R.N.K., E.W., L.C.B., T.H.V., S.R.F., and
R.P. contributed to the paleoenvironmental interpretation. S.E.B.
analyzed and described the hominin teeth, and S.R.F. analyzed the
monkey fossils. Other taxa were studied by F.H., J.L., E.D.G.E.,
L.C.B., T.W.P., and J.S.O. Bone identification and surface damage
were conducted by J.A.P., J.S.O., and T.W.P. Lithic technology
was studied by E.M.F. Additional insights on the lithic technology
were provided by D.R.B., J.S.R., C.L., I.C., S.K.B., N.K.R., and E.P.W.;
C.L. and I.C. conducted use-wear and residue analyses. T.W.P.
wrote the article with contributions from J.S.O., E.M.F., P.W.D.,
L.C.B., C.L., I.C., S.E.B., A.I.R.H., E.W., F.H., and Y.L. All authors
contributed editorial comments to the manuscript. Competing
Interests: The authors declare no competing interests. Data and
materials availability: All data supporting the findings of this
study are available within the paper and its supplementary
materials. All archeological and paleontological collections and field
records are archived in the Department of Earth Sciences in the
National Museums of Kenya in Nairobi. License information:
Copyright © 2023 the authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No
claim to original US government works. https://www.science.org/
about/science-licenses-journal-article-reuse
SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.abo7452
Materials and Methods
Figs. S1 to S17
Tables S1 to S21
References (36–135)
Submitted 22 February 2022; accepted 4 January 2023
10.1126/science.abo7452
Plummer et al., Science 379, 561–566 (2023) 10 February 2023 6of6
RESEARCH |RESEARCH ARTICLE