Conceived and designed the experiments: TWP PWD LCB JDK JVF DRB FH RBP RBP. Performed the experiments: TWP PWD LCB JDK JVF DRB FH. Analyzed the data: TWP PWD LCB JDK. Contributed reagents/materials/analysis tools: RBP RBP. Wrote the paper: TWP.
The authors have declared that no competing interests exist.
Major biological and cultural innovations in late Pliocene hominin evolution are frequently linked to the spread or fluctuating presence of C4 grass in African ecosystems. Whereas the deep sea record of global climatic change provides indirect evidence for an increase in C4 vegetation with a shift towards a cooler, drier and more variable global climatic regime beginning approximately 3 million years ago (Ma), evidence for grassland-dominated ecosystems in continental Africa and hominin activities within such ecosystems have been lacking.
We report stable isotopic analyses of pedogenic carbonates and ungulate enamel, as well as faunal data from ∼2.0 Ma archeological occurrences at Kanjera South, Kenya. These document repeated hominin activities within a grassland-dominated ecosystem.
These data demonstrate what hitherto had been speculated based on indirect evidence: that grassland-dominated ecosystems did in fact exist during the Plio-Pleistocene, and that early
The hominin fossil and archeological records of Africa exhibit substantial anatomical and behavioral change during the Plio-Pleistocene (∼1.5–3.0 Ma), including the evolution of
The composite stratigraphic log shows the basal three beds of the Southern Member (KS-1 to KS-3) and the base of KS-4. Spatially associated artifacts and fossils are found as diffuse scatters and also in more vertically discrete concentrations from the top of KS-1 through KS-3, with KS-2 providing the bulk of the archeological sample.
The late Pliocene Oldowan occurrences at Kanjera South are found on the northern margins of the Homa Mountain Carbonatite Complex, Homa Peninsula, southwestern Kenya (
KS-1 deposition began as a flow of pyroclastic material, possibly as a lahar, from the Homa Mountain complex in the south towards the depocenter in the Nyanza Rift graben. Lower KS-1 shows little internal stratification and no pedogenic development. In contrast, the well-bedded, better sorted and pedogenically modified upper parts of KS-1 represent reworking of the deposits by ephemeral streams running across the fan of the original pyroclastic flows. KS-2 represents a continuation of this environmental setting, with deposition by anastomosing channels flowing with intermittent, diffuse, generally low energy flow regimes and better-developed pedogenesis than KS-1.
KS-3 sees the transition to a wetter depositional environment, as evidenced by soft sediment deformation and the presence of a small channel, though stable land surfaces with pedogenesis continued to be found. KS-4 represents a continuation of this moister trend, with clays being deposited either during the transgression of a lake out of the depocenter to the north or during the formation of a wetland system. The homogeneity of the KS-4 clays favors the former interpretation, and the paleosol layers interbedded in KS-4 indicate intervals of lake regression sufficiently long for pedogenesis to take place.
A combination of biostratigraphy (co-occurrence of the equid
Except for artifacts and fauna found in thin, discontinuous conglomerate lenses, hominin activity was the primary agent of accumulation of the majority of archeological materials at this site
Habitats rich in plants using the C3 photosynthetic pathway, such as woodland and dry forests, are well-documented between 10 and 2 Ma in East Africa (
Shaded intervals represent pedogenic nodules forming in C3 dominated and C4 dominated environments. Paleosol carbonate δ13C values of -2 or greater are approximately equivalent to floral communities with 75% or more C4 plants
Isotopic analysis of enamel indicates that these taxa uniformly had a large amount of grass in their diets, reflecting the dominance of grass in the vegetation community (
A. Stable carbon isotopic composition of fossil mammal tooth enamel from KS-2 in Excavation 1. The KS-2 fauna is supplemented by several taxa unique to KS-1 or KS-3, or found on the surface of the KS-1 to KS-3 sequence, to provide a more complete sense of the diet of the mammalian community during the deposition of the archeological levels. The shading reflects the relative importance of C3 browse versus C4 grass in the diet, with δ13C values greater than -1 reflecting a diet with more than 75% C4 vegetation. Isotopic dietary classification follows others
These data provide the earliest isotopic evidence of an open habitat and a grassland-dominated ecosystem in East Africa. The presence of artifacts and archeological fauna both low and high in the KS-2 sequence and in the underlying KS-1 and overlying KS-3 indicates that hominins repeatedly visited this grass-rich area on the landscape for hundreds or even thousands of years. These data also substantively expand the known range of variation in Oldowan hominin habitat usage. Paleosol carbonate studies from the type locality of the Oldowan Industrial Complex, Olduvai Gorge, Tanzania, suggest that the Bed I and lower Bed II (∼1.7–1.87 Ma) basin margin was frequently well-wooded
These findings indicate that by ∼2.0 Ma Oldowan hominins had access to and used a broad spectrum of East African habitats, from open grassland to riparian forest. Stone tool manufacture and archeological site formation at this time is most likely attributable to the genus
The breadth of habitat use inferred for early
Excavation 1 was carried out within a grid of 169 1m×1m squares, excavated in 5 cm spits following site stratigraphy. Fossil and artifact-bearing horizons were dug with awls and dental picks. A Topcon total station was used for the precise determination of specimen N, E, and Z coordinates and in contour mapping. Object dip and orientation was measured with a Brunton compass. Sediments were dry sieved through 1 mm mesh. Sedimentary, taphonomic, and zooarcheological analyses indicate that the site assemblages formed predominantly through hominin activity
Pedogenic carbonates used in this analysis exhibited microstructure consistent with in situ formation without subsequent recrystallization. The difference in the δ 13C of occluded organic matter and pedogenic carbonate fits theoretical predictions for diagenetically unaltered materials. Diagenetic carbonate cements from the Kanjera Formation have negative δ 13C values (
Paleosol carbonate samples were washed in double distilled water and dried. The outer layers of the carbonate nodules were removed using a dental burr and discarded. The inner part of each carbonate nodule was crushed in an agate mortar and each sample was split into two aliquots, the first treated with 2% NaOHCl solution at 60°C for 24 hours to remove any organic contamination. The second aliquot was treated with 1M HCl until no reaction was observed and the remaining organic matter was washed to neutrality and freeze-dried. Organic samples were analyzed using flash combustion CF-IRMS. Samples were combusted in a Carlo Erba 1108 sample converter and the evolved gas was analyzed in a Europa Geo 20/20 gas source mass spectrometer at the University of Oxford.
Tooth enamel samples were carefully cleaned using an aluminium oxide air abrasive system to remove any adhering sediment and cementum. The outer surface of the enamel was abraded further, removing the outermost portion that was most likely to be diagenetically altered. Samples were then extracted from the cleaned enamel using a 0.5 mm diamond dental burr. Samples were ground and homogenized using an agate mortar. Powdered enamel samples were treated with 2% NaOHCl solution at 60°C for 24 hours to remove any organic contamination. Samples were then washed with double distilled water and treated with 0.1M CH3COOH at 25°C for 6 hours under vacuum to remove any secondary carbonate contamination. Samples were rinsed to neutrality and dried. All enamel samples for isotopic analysis were reacted with 100% phosphoric acid at 90°C in a common acid bath system. The evolved CO2 was pre-concentrated using a cold finger system and was analyzed at the University of Oxford using a VG Prism gas source isotope ratio mass spectrometer running in dual inlet mode.
Isotopic data from diagenetic sparry, pendant and poikilotopic calcite cements from KS-1 and KS-2, and from samples of carbonatite from the Homa Mountain carbonatite complex.
(0.03 MB DOC)
Paleosol Carbonate Isotopic Data.
(0.04 MB DOC)
Vertebrate taxon list from KS-2, Excavation 1. Isotopic dietary classification of Kanjera mammalian fossils follows others (23) using the isotopic data presented in
(0.04 MB DOC)
Stable isotopic composition of fossil eggshell and tooth enamel from Excavation 1.
(0.13 MB DOC)
We thank the Office of the President, Republic of Kenya, and the National Museums of Kenya for permission and support in conducting the field and laboratory studies described here. This research was conducted under the co-operative agreement between the National Museums of Kenya and the Smithsonian Institution. Logistical support was provided by the Human Origins Program of the Smithsonian Institution.