78,000-year-old record of Middle and Later Stone Age innovation in an East African tropical forest

78,000-year-old record of Middle and Later Stone Age innovation in an East African tropical forest Corrected: Publisher correction ARTICLE DOI: 10.1038/s41467-018-04057-3 OPEN 78,000-year-old record of Middle and Later Stone Age innovation in an East African tropical forest 1,2,3 4 5,6 7,8 9 4,10 Ceri Shipton , Patrick Roberts , Will Archer , Simon J. Armitage , Caesar Bita , James Blinkhorn , 11 2,12 13 8,14 4,15 Colin Courtney-Mustaphi , Alison Crowther , Richard Curtis , Francesco d’ Errico , Katerina Douka , 16 4,17 18 13 19 Patrick Faulkner , Huw S. Groucutt , Richard Helm , Andy I. R Herries , Severinus Jembe , 20,21 17 10 22 14,23 Nikos Kourampas , Julia Lee-Thorp , Rob Marchant , Julio Mercader , Africa Pitarch Marti , 24 25 19 26 27,28 Mary E. Prendergast , Ben Rowson , Amini Tengeza , Ruth Tibesasa , Tom S. White , 4,29 4 Michael D. Petraglia & Nicole Boivin The Middle to Later Stone Age transition in Africa has been debated as a significant shift in human technological, cultural, and cognitive evolution. However, the majority of research on this transition is currently focused on southern Africa due to a lack of long-term, stratified sites across much of the African continent. Here, we report a 78,000-year-long archeological record from Panga ya Saidi, a cave in the humid coastal forest of Kenya. Following a shift in toolkits ~67,000 years ago, novel symbolic and technological behaviors assemble in a non- unilinear manner. Against a backdrop of a persistent tropical forest-grassland ecotone, localized innovations better characterize the Late Pleistocene of this part of East Africa than alternative emphases on dramatic revolutions or migrations. 1 2 McDonald Institute for Archaeological Research, University of Cambridge, Downing Street Cambridge, Cambridge CB2 3DZ, UK. British Institute in Eastern Africa, Laikipia Road, Kileleshwa, Nairobi, Kenya. Centre of Excellence for Australian Biodiversity and Heritage, Australian National University, Canberra, ACT 2601, Australia. Department of Archaeology, Max Planck Institute for the Science of Human History, Kahlaische Strasse 10, Jena D-07745, Germany. 5 6 Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Pl. 6, Leipzig 04103, Germany. Department of Archaeology, University of Cape Town, Rondebosch 7701 Western Cape, South Africa. Department of Geography, Royal Holloway, University of London, Egham, Surrey TW20 OEX, UK. SSF Centre for Early Sapiens Behavior (SapienCe), University of Bergen, Øysteinsgate 3, Postboks 7805, Bergen 5020, 9 10 Norway. Malindi Museum, National Museums of Kenya, Malindi, Kenya. Department of Archaeology, Classics and Egyptology, University of Liverpool, 12–14 Abercromby Square, Liverpool L69 7WZ, UK. Department Environment, York Institute for Tropical Ecosystems, University of York, Heslington, York 12 13 YO10 5NG, UK. School of Social Sciences, The University of Queensland, St Lucia, QLD 4072, Australia. Department of Archaeology and History, The Australian Archaeomagnetism Laboratory, Palaeoscience Labs, La Trobe University, Melbourne Campus, Bundoora, VIC 3086, Australia. UMR 5199 PACEA, CNRS/Université de Bordeaux, Bâtiment B18, Allée Geoffroy Saint Hilaire, CS, 50023 - 33615 PESSAC CEDEX, France. Research Laboratory for Archaeology and the History of Art, Dyson Perrins Building, South Parks Road, Oxford OX1 3QY, UK. Faculty of Arts and Social Sciences, Department of Archaeology, The University of Sydney, Sydney, NSW, Australia. School of Archaeology, University of Oxford, 36 Beaumont Street, Oxford OX1 2PG, UK. 18 19 Canterbury Archaeological Trust, 92A Broad Street, Canterbury, Kent CT1 2LU, UK. Coastal Forests Conservation Unit, National Museums of Kenya, 20 21 Kilifi, Kenya. Centre for Open Learning, University of Edinburgh, Paterson’s Land, Edinburgh EH8 8AQ Scotland, UK. Biological and Environmental Sciences, University of Stirling, Stirling FK9 4LA Scotland, UK. Department of Anthropology and Archaeology, University of Calgary, 2500 University Drive, Calgary, AB T2N 1N4, Canada. Grup de Recerca Aplicada al Patrimoni Cultural (GRAPAC), Department of Animal Biology, Plant Biology and Ecology, Faculty of Biosciences, Autonomous University of Barcelona (UAB), Campus Bellaterra, Bellaterra 08193, Spain. Department of Sociology and Anthropology, Saint Louis University, Avenida del Valle 34, Madrid 28003, Spain. Invertebrate Biodiversity, National Museum Wales, Cathays Park, Cardiff CF10 3NP, UK. Department of Anthropology and Archaeology, University of Pretoria, cnr Lynnwood Road and Roper Street, Hatfield, South Africa. 27 28 University Museum of Zoology, Cambridge Downing Street, Cambridge CB2 3EJ, UK. Department of Life Sciences, The Natural History Museum, Cromwell Road, London SW7 5BD, UK. Human Origins Program, Smithsonian Institution, Washington, DC 20560, USA. Correspondence and requests for materials should be addressed to C.S. (email: ceri.shipton@anu.edu.au) or to P.R. (email: roberts@shh.mpg.de) or to N.B. (email: boivin@shh.mpg.de) NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 1 | | | 1234567890():,; A ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 he terms Middle Stone Age (MSA) and Later Stone Age forest and savannah (SM5). Excavations at PYS have revealed (LSA) have long been used to frame discussions of beha- exceptional preservation and stratigraphic integrity, and a record Tvioral and cultural change in Africa . Changes in lithic of human activity back to ~78 ka, including a rich technological production (such as making elongate blades and stone-tipped sequence that includes lithic forms elsewhere associated with the 2, 3 4, 5 arrows ), the appearance of symbolic material culture , and MSA and LSA. Alongside a rich range of paleoecological indi- subsistence diversification associated with the MSA and LSA cators, these features mean that the site offers a rare opportunity have all been identified as important thresholds in human cog- to study human behavioral changes in an evolutionarily critical, 3, 6 nitive and social evolution . Many researchers have highlighted but poorly-understood, region of Africa. the revolutionary nature of MSA and LSA human capacities, in some cases arguing that they reflect cognitive evolutionary Results developments or that they stimulated pan-African and global Stratigraphy and chronology. The 3 m deep excavated sequence 8, 9 migrations from 60,000 years ago (ka) onwards . On the other at PYS encompasses 19 layers (Supplementary Note 2) (Fig. 1b). hand, recent discoveries in southern Africa have suggested a more Three lithological boundaries divide the profile into four main 10–13 gradual development of these material culture traits . lithostratigraphic units that are discussed in detail in Supple- The debate as to the significance and tempo of behavioral mentary Note 1. A series of 20 stratigraphically ordered and changes during the MSA and LSA has largely focussed on the internally consistent radiocarbon and optically stimulated lumi- 12, 14 temperate and coastal environments of southern Africa . This nescence (OSL) age estimates, when included in a Bayesian is due to a general lack of well-dated, well-stratified records in model, show human occupation in every Marine Isotope Stage other key areas of the African continent, particularly across (MIS), from late MIS5 c. 78 ka into the Holocene (Supplementary the period 80–40 ka, though the Haua Fteah and Taforalt in Note 3). Geoarchaeological and micromorphological studies 15, 16 North Africa are notable exceptions . Long-term, dated indicate that the sequence consists of fine colluvia, spalling, and records from East Africa remain scarce. Several East African sites anthropogenic deposits with abundant organic and cultural have produced evidence for novel practices, reflected in the microremains (e.g., fauna, flora, lithic microdebitage – Supple- appearance of backed stone tools and beads over the last mentary Notes 4 and 5). 17–22 60–40,000 years . However, the chronologies and environ- Geomorphological and magnetic susceptibility proxies for mental contexts of these key behavioral transitions are not clear. human occupation intensity (Fig. 2; Supplementary Note 2) Sites >50,000 years old have only just begun to be identified indicate a pattern of intermittent pulses of human activity. In the beyond the East Africa Rift System and are limited to the Lake early part of the sequence, in Layers 19, 18, and the lower part of 24–26 Victoria region . Layer 17 (~78–73 ka), occupation intensity is low. This is The new archeological cave site of Panga ya Saidi (PYS) followed by a possible hiatus when both lithics and charcoal >c. described here offers an opportunity to address geographical and 125 μm drop off, and magnetic susceptibility and biogenic input ecological biases in our understanding of early human behavioral signals are dramatically reduced (Fig. 2). Human occupation and cultural change. The site is situated 15 km from the present- proxies subsequently show a general trend of increasing intensity day shoreline in the Zanzibar-Inhambane coastal forest mosaic from Layer 16 (beginning ~67 ka) to the Holocene (Fig. 2). that runs along the East African littoral, with the coastal shelf dropping below −125 m depth within 5 km of the modern coastline (Fig. 1a) (Supplementary Note 1). The cave is in the Cultural artifacts. The long artifactual sequence is comprised Dzitsoni limestone hills on an ecotone between lowland tropical of 17 ocher fragments, eight worked bone artifacts, 88 ostrich Layer ab Date in ka 0.5 1 1 3 7.5 7.5 14.5 PYS PYS 10 10 48.5 12 12 58.5 13 S Sh h 61.5 B Bn n Mombasa 16 71 – 67 S Sh h L Lt t Low coastal plain 76.5 Tropical moist forest Savannas Foot plateau Flooded savannas Ditzoni upland C Ch h Montane grasslands Coastal range 1 m Deserts High coastal plain 0 km 500 km 0 km 20 km OSL sample Radiocarbon Mangroves B Bn n Lakes Ash Burrow 0 mi 500 mi 0 mi 20 mi Fig. 1 Environmental setting and PYS stratigraphic section. a The location of PYS in the tropical moist forest of coastal East Africa, situated in the Ditzoni upland, southeastern Kenya. b The stratigraphic sequence of PYS showing the Layers and modeled ages, with example of a micromorphological thin section, illustrating the rich biogenic/anthropogenic contents in the sediments (Sh shell, Bn bone, Lt lithic, Ch charcoal). Note that Layer 12 was not continuous across the whole excavation and did not occur in this section. Age estimates are shown as the median of the highest posterior density age range for simplicity 2 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | I NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE a d f 40 mm 50 mm b c 50 mm 30 mm 45 mm 35 mm Fig. 2 Selected artifacts from PYS. a Levallois core from Layer 11. b Two backed lithic artifacts from Layer 11. c Backed lithic artifact from Layer 3. d Notched bone from Layer 8. e Notched bone from Layer 9. f Ocher crayon from Layer 10. g Ostrich eggshell bead from Layer 8. h Conus shell bead from Layer 16. i Gastropod shell bead from Layer 4 Stable isotopes Magnetic Mollusks Charcoal 18 13 susceptibility Biogenic content δ O(‰) δ C(‰) species >25um Lithic Lithic Lithic (counts per 0.5cm ) Sedimentology Depth Layer (VPDB) (VPDB) Phytoliths Bovids count nXLF nXFD% diameter Tetrapods density materials weight LM SA LM PSA LM PSA LM PSA Phytoliths LM PS absent below 16 layer 13 Sandier upwards 17 Redox pedo- features LM Sc 100% 0 100% 0 60 0.4 0.8 1.000.4 0 5 0 100 0 100% 0 Artefacts/1 60 100% 0 (g) 5 86 4 2 0 16 12 8 4 0 4 0 Clayey Silty Sand Phytoliths key Bovid key Biogenic key Tetrapods key Lithic materials key Grass Suni, duiker, dik dik Ash intraclast Bovids and suids Chert Bats and rodents Woody Bushbuck, kudu Bone Limestone Other tetrapods Palm Reedbuck, waterbuck Snail shell Quartz Hartebeest, topi, wildebeest Charcoal Buffalo Isotopic particles, Coprolith, quartz lithics, Unspecified large bovids Plant pseudomorph Fig. 3 The palaeoenvironmental and human occupation proxies from PYS. From left to right: Sedimentology (LM(SC) sandy clayey loam, LM(PS) pebbly sandy loam, LM(PSA) pebbly sandy ashy loam, LM(SA) sandy ashy loam); Depth; Layer divisions; Box and whisker plots of stable oxygen and carbon isotope values of mammalian teeth; Phytoliths, showing the proportion of grass, palm, and woody taxa; Percentage of different bovids in Minimum Number of Individuals (MNI); Terrestrial mollusk rarefied species count; Magnetic susceptibility (XLF and XFD%); Biogenic content of micromorphology thin sections; Microcharcoal abundance; Proportions of selected faunal groups as a percentage of total tetrapod MNI; Lithic density; Lithic material types; Lithic weight (mean debitage weight) eggshell beads, 27 marine shell beads, five exotic manuports, as made using variations of the Levallois method, and large retou- well as >30,000 knapped stone artifacts, including Levallois cores ched points (Supplementary Note 4), comfortably fitting with and backed artifacts—typical of MSA and LSA technologies, other contemporary MSA assemblages . Immediately after the 6, 27, 28 respectively (Fig. 3). The metrics and characteristics of hiatus, in Layer 16, there is a shift in rock type proportions from these technologies are discussed in greater detail in Supplemen- principally microcrystalline limestone to cryptocrystalline quartz tary Note 4. The possible hiatus or ephemeral occupation in the and chert, and a concomitant reduction in artifact size (Fig. 2, PYS sequence between 73 and 67 ka corresponds with a change in Supplementary Note 4). Size reduction is evident across all stone the stone artifact sequence. In the early part of the sequence artifacts, including within cryptocrystalline materials and retou- stone artifacts are characterized by large flakes (Fig. 2), often ched tools, demonstrating that this change does not simply reflect NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 3 | | | 2,5 2,0 1,5 1,0 1,0 0,5 m 0,5 m Depth (m) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 variation in procured raw material package size (Supplementary and δ O ranges of fauna after Layer 17 (Fig. 2) (Supplementary Note 4). Data 1)reflects the persistence of C and C vegetation, and a 3 4 Coupled with the switch to small cryptocrystalline stone tools variety of water sources affected to differing extents by after ~67 ka is an increased use of bipolar technology (Supple- evapotranspiration in the vicinity of PYS; likely in the form of mentary Fig. 4). A shift in technological emphasis toward bipolar a forest-grassland ecotonal setting. A consistent ecotonal situation strategies and a reduction in stone tool size is considered a key is supported by zooarchaeological data for the persistent presence 27, 30, 31 marker of LSA behavior elsewhere in Africa . A change of open and bush/forest adapted mammalian species, as well from Levallois to prismatic blade technology and the appearance as phytolith datasets that document the continued occurrence of backed crescentic tools have also been highlighted as indicators of woody, grass, and palm phytoliths in the immediate site 6, 32 of the LSA . However, in the PYS sequence, Levallois environment (Fig. 2). technology occurs alongside backed crescents in Layers 11 and 12 (~51–48 ka) and Layers 5–3 (~14–1 ka) (Supplementary Discussion Note 4) (Fig. 3). Prismatic blade production is rare in the PYS The paleoecological datasets from PYS agree with other East sequence, but bipolar and Levallois blades become common in African records that point to a period of low amplitude envir- the upper part of the sequence (Layers 8–3) (~25–1 ka) onmental change throughout much of MIS4-1, particularly from (Supplementary Note 4). Levallois and bipolar technology, blade records nearer the coast where the maritime influence buffers 33, 34 production, and backed crescentic tools occur recurrently and against temperature extremes . The consistent presence of intermittently, with no evidence for a unilinear accumulation of the forest-grassland ecotone throughout the last ~67,000 years traits or a uniform uptake of the latter three traits as a package. accompanies evidence for increasing occupation intensity at PYS, Despite changes in technology, stone artifacts remain consistently perhaps suggesting a growing human presence in the region small and were predominantly produced on cryptocrystalline linked to the use of mosaic habitats. Moreover, alongside the materials after ~67 ka. magnetic susceptibility data for increased occupation intensity Beads, ocher fragments, and worked bone have been associated from 60 ka, this may imply, as has been suggested elsewhere, that with behavioral complexity in the Late Pleistocene and occur with the MSA–LSA transition of East Africa is a long-term pattern of increasing regularity through the sequence at PYS (Supplemen- 33, 35 change related to growing population densities . tary Note 4). The earliest bead, a Conus sp. shell spire, occurs in Indeed, the PYS sequence does not document a radical change Layer 16 which dates from between ~67–63 ka (Fig. 3). At ~33 ka in technological or cultural behavior in East Africa ~60–50 ka that (Layer 9) the most common beads were Conus shell spires (n = might be suggestive of cognitive or technological “revolutions” or 13) (Fig. 3). Recurrent engagement with coastal resources for 7, 8 migrations . Instead, PYS documents a long-term assembly, and symbolic use is behavioral, rather than geographic, given minimal intermittent presence, of various innovative traits. In particular, changes in distance from the shoreline during the site’s there is no dramatic appearance of an LSA technological package occupation (Fig. 1a) (Supplementary Note 1). Ostrich eggshell and, instead, older MSA technological traits, such as Levallois beads at PYS reach their highest frequency ~25 ka (Layer 8, n = cores, exist alongside the development of backed artifacts 70), while fully-manufactured beads made of marine shells are the and blade production. The principal change in the sequence is dominant types during the Holocene (Layers 1–5) (Fig. 2). Layers the reduction in lithic size and the shift to cryptocrystalline 8-10 (~48–25 ka) produced two modified ocher fragments (Fig. 3), materials ~67 ka. as well as carved bone and tusk artifacts, including a decorated From MIS6-3 Homo sapiens began to adapt to a diversity of 13, 36 37, 38 39, 40 bone tube (Fig. 3) and a small bone point ; artifact types that coastal , tropical forest , and hyper-cold environ- have been argued to be characteristic of the LSA . In contrast ments across Africa and Eurasia. Humans appear to have adapted to many revolutionary or unilinear interpretations of technolo- locally to these environments, gradually developing new symbolic gical and cultural development, the PYS sequence reveals a forms, technological production strategies, and subsistence pattern of intermittent presence of different technological traits behaviors. It seems that the Middle and Late Pleistocene of Africa and symbolic artifacts that have been associated with the MSA is best characterized by diverse Homo sapiens populations, and LSA. adopting a range of survival strategies and new forms of social communication on an intermittent, ad hoc basis in different 41, 42 environmental and climatic contexts . It is this adaptive Paleoecology. Numerous proxies point to broad perseverance plasticity that truly defines the expansion and development of of tropical forest and grassland environments throughout the humans as a global species. sequence (Supplementary Note 5). This is indicated, for example, by the consistently high (> 25 species per sample) terrestrial Methods mollusk diversity, with most of the mollusk species requiring Excavation. Our excavation is in a large rockshelter in the first chamber of the humid shady conditions. collapsed-roof cave, near to the main entrance. Three seasons of excavation Sedimentology and magnetic susceptibility data indicates a between 2010 and 2013 have exposed a 3 m deep sequence, with the trench measuring 3.5 × 2 m at the top stepping into 2 × 1 m at the base. The trenches were shift to drier conditions following the early occupation in Layers excavated using the single context method according to the Museum of London 17–19 ~78–73 ka (Fig. 2) (see also Supplementary Note 5). This is protocols with the addition of total excavated sediment volume recorded for each supported, albeit with a lag, by the stable isotope data which context. Any animal burrows were excavated by hand and their contents discarded. shows higher stable carbon (δ C) in Layers 10–12 relative to In situ excavated deposits were dry-sieved on site through a 5 mm mesh. A program of bulk soil sampling for flotation (0.5 mm) and wet-sieving (1 mm) 13–16 and 17 (Supplementary Tables 12 and 13), indicative of an to recover archeobotanical, zooarcheological, and paleoenvironmental increase in the presence of C resources, most likely in the form of microremains was also implemented (see below). Where possible, a minimum grassland, in the diets of fauna being exploited at the site in the sample volume of 60 litres per context was maintained. Smaller contexts (<60 L) region c. 48 ka. Visually, δ O also tracks this trend but there is were sampled in their entirety. Nineteen stratigraphic layers were identified in the section, each corresponding to a particular excavation context. A continuous no statistically significant difference between Layer 17 and Layers column bulk sediment sample was taken for palaeoenvironmental analyses with 13–16 or Layers 10–12 (Supplementary Tables 14 and 15). 100 g samples taken at every 2 cm of depth. Deposits exposed in PYS Trench 4 From ~67–48 ka, paleoenvironmental proxies at PYS show that (2013 excavation) were logged on-site following standard sedimentological 43–45 the local environment underwent relatively little variation until procedures . Sediment color, texture, composition, structures, postdepositional the final occupation of the cave at ~0.5 ka (Fig. 2). The wide δ C disturbance and the nature, and geometry of layer boundaries were recorded for 4 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE each layer resolved by the excavators. Layers were grouped into higher-order, Fourteen C measurements were produced at the Oxford Radiocarbon multilayer lithostratigraphic units, as shown in Supplementary Fig. 1. Accelerator Unit (ORAU) (n = 10) and by Beta Analytic (n = 4). Most charcoal samples were prepared using the standard acid–base–acid (ABA) protocol . While for the younger material (<30 ka BP) this is usually sufficient, it has been shown Micromorphology. A set of 23 undisturbed micromorphology sediment samples that ABA does not efficiently decontaminate older charcoal samples when were collected from the excavated profile in clear polyurethane boxes. In view of compared with the more rigorous protocol: acid–base oxidation/stepped the large dimensions and stratigraphic complexity of the excavated trench, sam- combustion (ABOx-SC) . Paired ABA and ABOx-SC preparation was used on one pling was at a reconnaissance scale, concentrating on layer boundaries and dis- sample from context 413 C. For this particular sample, ABA and ABOx tinctive features. Sample boxes were labeled, photographed and plotted on the methodologies produced identical AMS ages. This is probably due to the profile drawing before removal from the profile (Supplementary Fig. 1). Out of this exceptional state of preservation of the charcoal in this part of the sequence as well sample set, 10 samples from the Pleistocene part of the profile were processed for as due to the fact that it is not very old sample. For material from lower levels micromorphological analysis at the Thin Section Micromorphology Laboratory, where only ABA dates from Beta Analytic exist these should be considered University of Stirling (sample code PYS; thin sections manufactured by George minimum ages only. This is demonstrated by the much older ages returned from MacLeod). Samples were air-dried and impregnated with polyester (polylite) resin ostrich eggshell samples from similar contexts. following standard procedures (http://www.thin.stir.ac.uk/). Ca. 30 μm thick, It is not possible to interpret radiocarbon ages reliably without calibration, due uncovered, large format thin sections (7.5 × 11 cm) were manufactured from the to variation in the concentration of radiocarbon in the atmosphere through time. hardened impregnated blocks. All terrestrial C measurements were calibrated using IntCal13 and SHCal13, the 55, 56 Thin sections were observed with a polarizing microscope at magnifications of most recent internationally agreed calibration curves available . As PYS lies ×12.5 to ×400, using plain polarized (PPL), cross-polarized (XPL), and oblique close to the equator, a 68.2/31.8 northern/southern hemisphere split was used in incident light (OIL). Relative abundance of sediment/soil components was the calibration curve, taking into account the position of the site relative to the 46, 47 estimated using standard semi-quantitative estimation charts . Key sediment inter-tropical convergence zone. The radiocarbon ages obtained from PYS are constituents larger than ca. 50 μm (mineral grains; pedoclasts; biogenic particles, summarized in Supplementary Table 1. etc.) were point-counted using a 0.5 × 0.5 cm grid overlay printed on clear acetate. Point-counted biogenic particles included bone fragments (from small cave OSL. Optically stimulated luminescence (OSL) samples were obtained by ham- vertebrates and larger vertebrate fauna—the latter possibly including human prey); mering metal tubes into section faces following cleaning. The samples were sealed small vertebrate coproliths; indeterminate isotropic particles; shell fragments; using adhesive tape. Following transport to the Royal Holloway University of charcoal (both woody and non-woody tissue); burnt plant pseudomorphs; ash and London Luminescence Laboratory, samples were processed under subdued orange ash intraclasts; quartz debitage (Supplementary Fig. 2). light. The outer 5 cm of sample (presumed as being exposed to sunlight) was removed and retained for background radioisotope concentration determination. Paleomagnetism. A sub-sample of the continuous column sample from PYS was Quartz was extracted from the part of each sample not exposed to sunlight transported to The Australian Archaeomagnetism Laboratory for preparation and following burial. As the bedrock at PYS is primarily composed of carbonate, analysis. Once in the laboratory samples were dried to standardize water content, samples were initially wet-sieved to isolate the 212–180 µm size fraction. This crushed with a non-magnetic mortar, and pestle to become homogenized and then removes large bedrock clasts from the sample before acid treatment, meaning that packed into standard 8cc palaeomagnetic plastic cubes. The approach to the the possibility of incorporating grains liberated by dissolution of the bedrock was analysis follows that of Herries and a number of analyses were run to establish minimized. the magnetic mineralogy of the samples. These included mass specific low-field The volume of 1 M HCl and H O were used to remove carbonates and organic 2 2 susceptibility (χlf), mass specific high-field susceptibility (χhf), mass specific fre- matter from the 212–180 µm fraction, respectively. The samples were then re- quency dependant susceptibility (χfd) and saturation isothermal remanent mag- sieved at 180 µm and quartz extracted from the >180 µm fraction using density netization acquisition curves and backfields. This was done to understand the separations at 2.62 and 2.70 g/cm followed by a HF acid etch (23 M HF for 60 min mineralogy driving magnetic susceptibility change through identifying ferrimag- followed by 10 M HCl rinse). The resulting, etched samples were sieved at 150 µm netic vs. anti-ferromagnetic minerals on the basis of their coercivity, establish the to remove partially dissolved grains. All samples were then stored in opaque concentration of magnetic minerals present within each sample, and examine grain containers prior to measurement. size and domain state trends in the sequence. All OSL measurements were carried out using a Risø TL/OSL-DA-15 57 58, 59 The magnetic susceptibility of each sample was measured using a Bartington automated dating system , fitted with a single-grain OSL attachment . Single- MS2 susceptibility-meter connected to an MS2B sensor. Samples were measured at grains were stimulated using a 10 mW Nd: YVO4 solid-state diode-pumped green 2 57 0.47 kHz (low, χlf) and 4.7 kHz (high, χhf) to attain both low and high frequency laser (532 nm) focused to yield a nominal power density of 50 W/cm . All susceptibility values. These measurements were then used to compute the infrared (IR) stimulation was carried out using an IR (870 nm) laser diode array frequency dependence of the magnetic susceptibility (χfd) and then expressed as a yielding a power density of 132 mW/cm . OSL passed through 7.5 mm of Hoya percentage (χfd%) using the formula stated by Dearing et al. . Isothermal U-340 filter and was subsequently detected using an Electron Tubes Ltd 9235QB15 remanent magnetizations (IRM) were induced up to 1 T (representing the photomultiplier tube. saturation IRM: SIRM) using a magnetic measurements pulse magnetizer Irradiation was carried out using a 40 mCi 90 Sr/90Y beta source providing ~6 (MMPM10) and measurements made on an AGICO JR6 magnetometer. Forward- Gy/min. This source is calibrated relative to the National Physical Laboratory, field measurements were taken at 20, 500, 600 mT, and 1 T and back-field Teddington 60Co γ-source (Hotspot 800) . Due to the spatial inhomogeneity of measurements at 20, 40, 100, 150, 200, 250, and 300 mT. Full IRM acquisition beta emitters across the active face of our 90 Sr/90Y beta source we calibrated the curves and backfields were also produced for each layer of the site. HIRM dose rate to each individual grain position on a single-grain disc using the 50 62 measurements were also taken using the method described by Liu et al. . Soft IRM method reported by Armitage et al. . For more detail on measurement and quality measurements showing the concentration of ferrimagnetic minerals were taken at criteria see Supplementary Note 3. 20 mT. S-ratios (IRM-300 mT/SIRM and IRM-100 mT/SIRM) were calculated to establish variation in the grain size of ferromagnetic minerals. Remanence of Bayesian methods. The absolute age determinations were used to construct an age coercivity values were also calculated to further establish the mineralogy, grain size, model using Bayesian software (OxCal 4.3 ) and the INTCAL13 curve. The and domain states of the magnetic minerals present within the samples. determinations were input as values in fraction modern (fM) plus or minor fM errors at 1σ (R_F14C in OxCal). In order to determine whether there are pro- blematic determinations that do not agree with the prior framework, an outlier Charcoal abundance. From the continuous column sample described in SM3, detection method was applied. When there is a lack of agreement with the prior subsamples of 1 cm were extracted for charcoal abundance analysis. The addition framework, significant outlier results allow us to quantify the degree of difference. of sodium hexametaphosphate to a beaker containing the sample was used to Values excessively higher than the prior outlier probabilities applied are auto- disaggregate the samples and aid in the separation of the organic material and the matically down-weighted in the models. A posterior outlier probability of 0.5 clay particles . The contents of the beaker was then passed through a 125 μm sieve means that the radiocarbon likelihood of the sample is only included in half of the and the trapped content transferred to a gridded Petri dish. Pieces of charcoal were runs of the model. The two Beta dates were assigned a 0.3 value to reflect the identified using comparative collections and the total charcoal count was deter- inadequate chemical protocol applied in the decontamination of these samples mined through visual inspection and manipulation with a metal probing needle prior to measurement. under a Zeiss Stemi 2000-C optical stereomicroscope (×10–40 magnifications). The Only age estimates older than 20,000 years ago were included in the model. The microscopic charcoal size identified represents charcoal that was locally produced inclusion of younger ages does not affect the older ages at all. Younger ages were during fires within the catchment area of the site . excluded because of the large span of the dates; the model runs with a broad resolution of 50–100 years for the older data, whereas 20 years would be more Radiocarbon dating. Items selected for radiocarbon dating were a human bone appropriate for the younger data. In each Bayesian model, a start and end and a charred sorghum seed (identified by Alison Crowther) from the upper part of boundary was added in order to bracket the archeological phases within the the sequence; as well as unidentified charcoal, and ostrich eggshell pieces including sequence. The posterior distributions of these boundaries facilitated determination a bead, from the middle part of the sequence. These dating samples were either of probability distribution functions (PDF) for the beginning and ending of these recovered during excavation or taken from the section at the end of excavation. phases of activity. Due to the presence of a depositional hiatus represented in the NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 5 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 later part of Layer 17, we used two boundaries between the end of 17 and start of one articular surface and/or key landmark, enabling identification to element and Layer 16. In addition, due to uncertainties related to the reliability of sample OSL-7 usually to taxon or at least taxonomic group (e.g., “bird,”“small carnivore”). (see text above) these age estimates were not included in the model. MaxID specimens were identified in all contexts in both trenches. Second, minimally identifiable (minID) bones include limb shafts and axial fragments that can generally be identified at least to carcass size; these were identified in nine high- Lithics. A total of 30,420 lithic artifacts were recovered through the excavation priority contexts from Trench 4, in order to obtain a sample for taphonomic seasons at PYS between 2010 and 2013. The analyses proceeded by classifying all analysis. Third, nonidentified (NID) bones were separated from minID specimens lithics in accordance with stratigraphic context, raw-material type, and technolo- and weighed in these selected contexts. On average across all contexts in Trench 4, gical class. All artifacts were subsequently weighed and counted according to these 10% of the assemblage was maximally identifiable. When minID specimens categories. Cores and retouched artifacts were further classified in accordance with identified in nine selected contexts were included the average identification rate reduction strategy and typology. For unretouched flakes, blades, Levallois flakes, rose to 37%. 64 65 and bipolar flakes were counted. Levallois , bipolar , laminar, and discoidal Taxonomic identifications were made on the basis of extensive reference strategies as well as resultant blanks were all documented to varying frequencies in collections housed at the National Museums of Kenya (Nairobi) Osteology Unit. a number of the layers. Calculations of the minimum number of individuals (MNI) were made using the resulting database following completion of this analysis and took into account specimen laterality, size, and where relevant, age estimates. It should be stressed Beads, osseous artifacts, and ocher. Over two hundred potential beads, bone tools, engraved bone and stone objects, and pigment lumps recovered during that limb shafts were only studied in selected contexts and therefore could not be excavation, were examined under a low power reflected light microscope in search used to calculate the minimum number of elements (MNE) or the resulting MNI for anthropogenic modifications. When necessary, sediment was carefully removed values for layers or phases, as is standard practice in contexts where density- under the microscope with a soft brush or a wet tooth pick. This resulted in the mediated attrition has occurred. It is therefore possible that MNI estimates will be retention of 159 pieces bearing compelling traces of manufacture and use, too low in some instances. unmodified or marginally modified shell fragments probably used as beads, and modified and unmodified lumps of iron-rich rock and sediments, possibly used to extract ocher powder. Stable carbon and oxygen isotope analysis of mammalian tooth enamel. Stable The retained artifacts were examined at magnifications between ×4 and ×40, isotope analysis of mammalian tissues has frequently been used to assess the diets and photographed with a motorized Leica Z6 APOA microscope equipped with a 71–73 and ecologies of East African fossil fauna . This work primarily relies on the Leica Application Suite (LAS) and Multifocus module, and Leica Map DCM 3D distinction between the C -or C -photosynthetic pathways at the base of East 3 4 computer software. The Multifocus module enables the acquisition of extended African foodwebs. In the context of tropical and sub-tropical forest ecologies, this depth of field images. Once the digital images had been complete for different distinction can be used to assess the degree of faunal reliance on C forest resources heights, algorithms in the software compile them into a single composite images as opposed to C plant resources available in open habitats, with C plants being 4 4 that significant extends the depth of field, and provides clarity in viewing the entire 13 74–76 enriched in C relative to C plants . object. This distinction is further enhanced by the “canopy effect” whereby vegetation The selected areas of one Conus shell were scanned using a Sensofar Sneox growing under a closed forest canopy is strongly depleted in C (with scanning confocal microscope with a ×20 objective. The resulting files were correspondingly lower measured δ C) due to low light and the presence of large analyzed with Mountains 7.2 software. For further details on the analytical 77, 78 amounts of respired CO . This results in the tissues of animals consuming protocols for beads, osseuous artifacts, and ocher see Supplementary Note 4 and forest vegetation, as well as forest herbivores, having lower δ C values than the references therein. A full description of all osseous artifacts, beads, and used animals pending some, or all, of their time consuming open-habitat foodstuffs ocher will be reported elsewhere, but Fig. 3 shows examples of each of these artifact e.g. . types from PYS excavation, and Supplementary Figure 19 gives an overview of their Stable oxygen isotope measurements from mammalian enamel can yield further distribution through the sequence. 80, 81 paleoecological information about water and food . Given a constant source of water, plant water δ O will primarily reflect the impacts of relative humidity on Phytolith analysis. Seventeen sediment samples from the continuous stratigraphic leaf water evapotranspiration, with decreasing humidity resulting in increased δ O 82–84 column described in SM3 were processed for phytolith analysis. Of these, the lower values . In a tropical or sub-tropical setting, increased forest cover, and the six samples were barren. We followed extraction protocols employed on Middle resulting shade and increased humidity, will lead to decreased evapotranspiration 66 67 18 79 Stone Age sites from adjacent countries and modern topsoils . The procedure and therefore decreased δ O, especially on the forest floor . As faunal tooth 18 18 included sieving, drying, deflocculating, acid/base treatment, and sequential density enamel δ O primarily reflects water and food-water δ O, herbivores feeding and separation by manipulating the specific gravity of sodium polytungstate. Aliquot drinking in forests can be expected to have lower enamel δ O than those feeding mounting was with “Entellan New,” which allowed for microscopic inspection in open, irradiated areas. This is complicated by physiological and behavioral (×40) and 3D rotation before drying. The average count per slide was 238 phy- variables . Nevertheless, animals consuming plants and water in a shaded, forested toliths. The inferential baseline was grounded on East African phytoliths from setting will broadly reflect the corresponding lower levels of evapotranspiration in 68, 69 plants and soils . Preservation was adequate for morphometric analysis and their enamel δ O. type identification. Stable carbon and oxygen isotope analysis of faunal tooth enamel excavated from the various archeological Phases at PYS was undertaken in order to directly assess the diets and ecologies of animals being exploited by humans living at the Mollusk analysis. Macro marine molluskan remains were identified and counted site at different points in time. This should, in turn, provide information regarding by Patrick Faulkner using comparative modern reference collections. Since no fluctuations in the degree of forest cover surrounding PYS in the past. Faunal evidence for subsistence on these was found prior to Layer 6, they will be reported enamel samples were taken from all available PYS Layers. A broad selection of in detail elsewhere. PYS is an open-roofed cave with substantial input from the species was sampled for each Layer based on availability. Where possible, up to five external environment. The paleoenvironmental samples recovered from PYS members of each species/genus were sampled per layer grouping (Supplementary contained terrestrial mollusk fauna representing the area in and immediately Data 1). Faunal samples were identified to species and/or genus level using the outside the cave over a long time period . There is no evidence that any of these substantial reference collection available at the National Museums of Kenya, species have been transported to the site through natural or anthropogenic Nairobi. The full list of faunal samples and tooth identifications analyzed in this processes. study are shown in Supplementary Data 1. Excluding large marine shells that form part of the archeological record, the Air-abrasion was used to remove any adhering detrital material from the teeth aquatic component of the assemblage is extremely small and is less significant than or tooth fragments to be studied. Gentle abrasion with a diamond-tipped drill was that seen in faunas from the modern coastal forests on Zanzibar. As a result, performed along the full length of the buccal surface of the tooth or tooth fragment transport by fluvial activity or regular flooding of the area can be ruled out. The in order to maximize the period of formation represented by the resulting isotopic diverse community of snails could not have been sustained in a closed cave analysis. The resulting enamel powder was pretreated using standard, published environment but would have required continual external input in the form of leaf protocols in order to remove any organic or secondary carbonate contaminates. litter and the snails themselves or their shells. The non-marine mollusk record This consisted of a wash in 1.5% sodium hypochlorite for 60 min, followed by three from PYS therefore offers a long record of the local environment through time. rinses in purified H O. A volume of 0.1 M acetic acid was then added for 10 min 86, 87 prior to another three rinses in purified H O . Tetrapod analysis. The zooarchaeological analysis of PYS osseous remains took Gases were evolved from the treated samples using 100% Phosphoric Acid. t 13 18 place in 2012 (Trench 3) and 2014 (Trench 4) and produced a database of 5256 δ C and δ O of the resulting gases was measured using a Thermo Gas Bench 2 identified specimens (Number of Identified Specimens, NISP). A total of 6.4 kg of connected to a Thermo Delta V Advantage Mass Spectrometer at the Division of bone was excavated from Trench 3 and 14.7 kg from Trench 4. The Trench 4 study Archeological, Geographic and Environmental Sciences Bradford University. δ C included both taxonomic identification and taphonomic analysis, but microfauna and δ O measurements from samples were compared against international were not analyzed in Phase 1 due to time constraints. In Trench 3, microfauna were standards (NBS 19 and CO-9) registered by the International Atomic Energy identified to a very general level (e.g., “Muridae”). Agency (five of each for a run of 60). Replicate analysis of in-house OES and Initial sorting of faunal remains created up to three categories: first, maximally MERCK standards (six of each for a run of 60) suggests that machine measurement 13 18 identifiable (maxID) specimens include teeth and those bones that preserve at least error is c. ±0.1‰ for δ C and ±0.2‰ for δ O. 6 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE Analysis of variance (ANOVA) tests were performed on faunal enamel δ C 23. Groucutt, H. S. et al. Rethinking the dispersal of Homo sapiens out of Africa. and δ O to determine whether these isotopic parameters differed between groups Evolut. Anthropol.: Issues, News, Rev. 24, 149–164 (2015). of stratigraphic layers. The stratigraphic layers were grouped based on meaningful 24. Faith, J. T. et al. Paleoenvironmental context of the Middle Stone Age chronological and stratigraphic divisions as follows: 1–3, 4–6, 7–9, 10–12, 13–16, record from Karungu, Lake Victoria Basin, Kenya, and its implications and 17. Where variance was found to be significant following ANOVA testing, a for human and faunal dispersals in East Africa. J. Hum. Evol. 83,28–45 post-hoc Tukey pair-wise comparison was performed in order to determine which (2015). layer groupings were significantly different from each other in terms of δ C and 25. Tryon, C. A. et al. The Pleistocene prehistory of the Lake Victoria basin. Quat. 18 88 δ O. All statistical analyses were conducted using the free program R software . Int. 404, 100–114 (2016). 26. Blegen, N., Faith, J. T., Mant-Melville, A., Peppe, D. J. & Tryon, C. A. The Data availability. The authors declare that all data supporting the findings of this Middle Stone Age after 50,000 years ago: new evidence from the Late study are available upon request from the authors. The artifacts and faunal remains Pleistocene sediments of the eastern Lake Victoria basin, Western Kenya. from the Panga ya Saidi excavation are curated in the National Museum of Kenya, PaleoAnthropology 139, 139–169 (2017). Nairobi, under the site code PYS and the suffixes 10, 11, and 13 (denoting the year 27. Villa, P. et al. Border Cave and the beginning of the Later Stone Age in South of excavation). Africa. Proc. Natl Acad. Sci. USA 109, 13208–13213 (2012). 28. d’Errico, F. et al. Early evidence of San material culture represented by organic artifacts from Border Cave, South Africa. Proc. Natl Acad. Sci. USA 109, Received: 23 June 2017 Accepted: 29 March 2018 13214–13219 (2012). 29. Tryon, C. A. & Faith, J. T. Variability in the Middle Stone Age of eastern Africa. Curr. Anthropol. 54, S234–S254 (2013). 30. Eren, M. I., Diez-Martin, F. & Dominguez-Rodrigo, M. An empirical test of References the relative frequency of bipolar reduction in Beds VI, V, and III at Mumba 1. Goodwin, A. & vanRiet Lowe, C. The Stone Age cultures of South Africa. Ann. Rockshelter, Tanzania: implications for the East African Middle to Late Stone South Afr. Mus. 27,1–289 (1929). Age transition. J. Archaeol. Sci. 40, 248–256 (2013). 2. Lombard, M. Quartz-tipped arrows older than 60 ka: further use-trace 31. Pargeter, J. Lithic miniaturization in Late Pleistocene southern Africa. evidence from Sibudu, KwaZulu-Natal, South Africa. J. Archaeol. Sci. 38, J. Archaeol. Sci.: Rep. 10, 221–236 (2016). 32. Tryon, C. A. et al. Middle and Later Stone Age chronology of Kisese II 1918–1930 (2011). 3. Henshilwood, C. S. Late Pleistocene techno-traditions in southernAfrica: a rockshelter (UNESCO World Heritage Kondoa Rock-Art Sites), Tanzania. review of the Still Bay and Howiesons Poort, c. 75–59 ka. J. World Prehistory PLOS One 13, e0192029 (2018). 25, 205–237 (2012). 33. Blome, M. W., Cohen, A. S., Tryon, C. A., Brooks, A. S. & Russell, J. The 4. Wadley, L. Those marvellous millennia: the Middle Stone Age of southern environmental context for the origins of modern human diversity: a synthesis Africa. Azania.: Archaeol. Res. Afr. 50, 155–226 (2015). of regional variability in African climate 150,000–30,000 years ago. J. Hum. 5. Jacobs, Z. et al. Ages for the Middle Stone Age of southern Africa: implications Evol. 62, 563–592 (2012). 34. Moernaut, J. et al. The seismic-stratigraphic record of lake-level fluctuations in for human behavior and dispersal. Science 322, 733–735 (2008). 6. Ambrose, S. H. Small things remembered: origins of early microlithic industries Lake Challa: hydrological stability and change in equatorial East Africa over in sub‐Saharan Africa. Archeol. Pap. Am. Anthropol. Assoc. 12,9–29 (2002). the last 140kyr. Earth Planet. Sci. Lett. 290, 214–223 (2010). 7. Klein, R. G. Archeology and the evolution of human behavior. Evolut. 35. Tryon, C. A. & Faith, J. T. A demographic perspective on the Middle to Later Anthropol. Issues News Rev. 9,17–36 (2000). Stone Age transition from Nasera rockshelter, Tanzania. Philos. Trans. R. Soc. 8. Mellars, P. Why did modern human populations disperse from Africa ca. B 371, 20150238 (2016). 60,000 years ago? A new model. Proc. Natl Acad. Sci. USA 103, 9381–9386 36. O’Connor, S., Louys, J., Kealy, S. & Samper Carro, S. C. Hominin dispersal and settlement east of Huxley’s Line: the role of sea level changes, island size, (2006). 9. Shea, J. J. & Sisk, M. L. Complex projectile technology and Homo sapiens and subsistence behavior. Curr. Anthropol. 58, S567–S582 (2017). dispersal into western Eurasia. PaleoAnthropology 2010, 100–122 (2010). 37. Mercader, J. Forest people: the role of African rainforests in human evolution 10. McBrearty, S. & Brooks, A. S. The revolution that wasn’t: a new interpretation and dispersal. Evol. Anth. 11, 117–124 (2002). 38. Roberts, P. et al. Direct evidence for human reliance on rainforest resources in of the origin of modern human behavior. J. Hum. Evol. 39, 453–563 (2000). 11. Conard, N. J. & Will, M. Examining the causes and consequences of short- late Pleistocene Sri Lanka. Science 347, 1246–1249 (2015). term behavioral change during the Middle Stone Age at Sibudu, South Africa. 39. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western Siberia. Nature 514, 445–449 (2014). PLOS One 10, e0130001 (2015). 12. Henshilwood, C. S. et al. A 100,000-year-old ochre-processing workshop at 40. Fiedel, S. J. & Kuzmin, Y. V. Radiocarbon date frequency as an index of Blombos Cave, South Africa. Science 334, 219–222 (2011). intensity of Paleolithic occupation of Siberia: did humans react predictably to 13. Marean, C. W. et al. Early human use of marine resources and pigment in climate oscillations? Radiocarbon 49, 741–756 (2007). 41. Lombard, M. & Parsons, I. What happened to the human mind after the South Africa during the Middle Pleistocene. Nature 449, 905–908 (2007). 14. Marean, C. W. Pinnacle Point Cave 13B (Western Cape Province, South Howiesons Poort? Antiquity 85, 1433–1443 (2011). Africa) in context: the Cape floral kingdom, shellfish, and modern human 42. d’Errico, F. et al. Identifying early modern human ecological niche expansions and associated cultural dynamics in the South African Middle Stone Age. Proc. origins. J. Hum. Evol. 59, 425–443 (2010). 15. Bouzouggar, A. et al. 82,000-year-old shell beads from North Africa and Natl Acad. Sci. USA 114, 7869–7876 (2017). implications for the origins of modern human behavior. Proc. Natl Acad. Sci. 43. Goldberg, P. & Macphail, R. Practical and Theoretical Geoarchaeology USA 104, 9964–9969 (2007). (Blackwell Publishing Ltd, Hoboken, NJ, 2006). 44. Tucker, M. E. Sedimentary Rocks in the Field: A Practical Guide(John Wiley & 16. Douka, K. et al. The chronostratigraphy of the Haua Fteah cave (Cyrenaica, northeast Libya). J. Hum. Evol. 66,39–63 (2014). Sons, Chichester, 2011). 17. Gliganic, L. A., Jacobs, Z., Roberts, R. G., Domínguez-Rodrigo, M. & Mabulla, 45. Ayala, G., Canti, M., Heathcote, J., Sidell, J. & Usai, M. Geoarchaeology: Using Earth Sciences to Understand the Archaeological Record (English Heritage, A. Z. New ages for Middle and Later Stone Age deposits at Mumba rockshelter, Tanzania: optically stimulated luminescence dating of quartz and Swindon, 2007). feldspar grains. J. Hum. Evol. 62, 533–547 (2012). 46. Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G.. & Tursina, T. Handbook for 18. Ambrose, S. H. Chronology of the Later Stone Age and food production in Soil Thin Section Description (Waine Research, Albrighton, 1985). 47. Stoops, G. Guidelines for Analysis and Description of Soil and Regolith Thin East Africa. J. Archaeol. Sci. 25, 377–392 (1998). 19. Brandt, S. A. et al. Early MIS 3 occupation of Mochena Borago Rockshelter, Sections (Soil Science Society of America Inc., Madison, WI, 2003). Southwest Ethiopian Highlands: implications for Late Pleistocene archaeology, 48. Herries, A. I. Archaeomagnetic evidence for climate change at Sibudu Cave. paleoenvironments and modern human dispersals. Quat. Int. 274,38–54 South. Afr. Humanit. 18, 131–147 (2006). (2012). 49. Dearing, J. A. et al. Frequency-dependent susceptibility measurements of 20. Pleurdeau, D. et al. Cultural change or continuity in the late MSA/Early LSA environmental materials. Geophys. J. Int. 124, 228–240 (1996). of southeastern Ethiopia? The site of Goda Buticha, Dire Dawa area. Quat. Int. 50. Liu, Q. et al. Environmental magnetism: principles and applications. Reviews of Geophysics 50, RG4002 (2012). 343, 117–135 (2014). 21. Miller, J. M. & Willoughby, P. R. Radiometrically dated ostrich eggshell beads 51. Bamber, R. Sodium hexametaphosphate as an aid in benthic sample sorting. from the Middle and Later Stone Age of Magubike Rockshelter, southern Mar. Environ. Res. 7, 251–255 (1982). Tanzania. J. Hum. Evol. 74, 118–122 (2014). 52. Lynch, J. A., Clark, J. S. & Stocks, B. J. Charcoal production, dispersal, and 22. Diez-Martín, F. et al. The MSA/LSA technological transition in East Africa. deposition from the Fort Providence experimental fire: interpreting fire New data from Mumba Rocksheiter Bed V and their implications in the origin regimes from charcoal records in boreal forests. Can. J. For. Res. 34, of modern human behaviour. J. Afr. Archaeol. 7, 147–173 (2009). 1642–1656 (2004). NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 7 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 53. Brock, F., Higham, T., Ditchfield, P. & Bronk Ramsey, C. Current 83. Yakir, D., Berry, J., Giles, L. & Osmond, C. B. Isotopic heterogeneity of water pretreatment methods for AMS radiocarbon dating at the Oxford in transpiring leaves: identification of the component that controls the δ Oof Radiocarbon Accelerator Unit (ORAU). Radiocarbon 52, 103–112 (2010). atmospheric O and CO . Plant Cell Environ. 17,73–80 (1994). 2 2 54. Bird, M. et al. Radiocarbon dating of “old” charcoal using a wet oxidation, 84. Sheshshayee, M. et al. Oxygen isotope enrichment (Δ O) as a measure of stepped-combustion procedure. Radiocarbon 41, 127–140 (1999). time-averaged transpiration rate. J. Exp. Bot. 56, 3033–3039 (2005). 55. Reimer, P. J. et al. IntCal13 and Marine13 radiocarbon age calibration curves 85. Levin, N. E., Cerling, T. E., Passey, B. H., Harris, J. M. & Ehleringer, J. R. A 0–50,000 years cal BP. Radiocarbon 55, 1869–1887 (2013). stable isotope aridity index for terrestrial environments. Proc. Natl Acad. Sci. 56. Hogg, A. G. et al. SHCal13 Southern Hemisphere calibration, 0–50,000 years USA 103, 11201–11205 (2006). cal BP. Radiocarbon 55,1–15 (2013). 86. Sponheimer, M. et al. Hominins, sedges, and termites: new carbon isotope 57. Bøtter-Jensen, L., Andersen, C., Duller, G. & Murray, A. Developments in data from the Sterkfontein valley and Kruger National Park. J. Hum. Evol. 48, radiation, stimulation and observation facilities in luminescence 301–312 (2005). measurements. Radiat. Meas. 37, 535–541 (2003). 87. Lee-Thorp, J. et al. Isotopic evidence for an early shift to C4 resources by 58. Duller, G., Bøtter-Jensen, L. & Murray, A. Combining infrared-and green- Pliocene hominins in Chad. Proc. Natl Acad. Sci. USA 109,20369–20372 (2012). laser stimulation sources in single-grain luminescence measurements of 88. R Core Team. 2013. R: A language and environment for statistical computing. feldspar and quartz. Radiat. Meas. 37, 543–550 (2003). R Foundation for Statistical Computing, Vienna, Austria. 59. Duller, G., Bøtter-Jensen, L., Murray, A. & Truscott, A. Single grain laser luminescence (SGLL) measurements using a novel automated reader. Nucl. Acknowledgements Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater. At. 155, 506–514 Funding was provided by the SEALINKS project under a European Research Council (1999). (ERC) grant (no. 206148) awarded to N.B. Permission to conduct the research was 60. Armitage, S. & Bailey, R. The measured dependence of laboratory beta dose granted by the Office of the President of the Republic of Kenya through affiliation with rates on sample grain size. Radiat. Meas. 39, 123–127 (2005). the National Museums of Kenya (NMK). We are grateful for the support of the NMK 61. Ballarini, M., Wintle, A. & Wallinga, J. Spatial variation of dose rate from beta and the British Institute in Eastern Africa. P.R. was funded by NERC and the Boise Fund sources as measured using single grains. Anc. TL 24,1–8 (2006). (University of Oxford). S.J.A. and F.D. acknowledge support from the Research Council 62. Armitage, S. J. et al. The southern route “out of Africa”: evidence for an early of Norway, through its Centres of Excellence funding scheme, SFF Centre for Early expansion of modern humans into Arabia. Science 331, 453–456 (2011). Sapiens Behaviour (SapienCE) (no. 262618). FD and AP were funded by the ERC grant, 63. OxCal version 4.3.2 (2017). TRACSYMBOLS (no. 249587), and the Agence Nationale de la Recherche (ANR-10- 64. Boëda, E. Le concept Levallois, variabilité des méthodes (CNRS éditions, Paris, LABX-52), LaScArBx Cluster of Excellence. A.P.M. holds a Beatriu de Pinós postdoctoral 1994). fellowship (2014 BP-A 00122) from the Agency for Management of University and 65. de la Peña, P. A qualitative guide to recognize bipolar knapping for flint and Research Grants, Government of Catalonia. A.C. and H.S.G. were funded by the British quartz. Lithic Technol. 40, 316–331 (2015). Academy. Additional support has been provided by the McDonald Institute for 66. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Phytoliths Archeological Research (University of Cambridge) and the Max Planck Society. For from Middle Stone Age habitats in the Mozambican Rift (105–29 ka). J. Hum. assistance in the field and with artifact analyses, we wish to thank Jackson Mupe, Yahya Evol. 64, 328–336 (2013). Lenga, Emmanuel Mupe, Mohammed Lenga, Louise Green, John Mpangarusyia, Tim 67. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Soil McEnright, James Munene, Doreen Mutoro, Joyce Wathira, and Emma Le Vraux. phytoliths from miombo woodlands in Mozambique. Quat. Res. 75, 138–150 Annette Guenzel and Michelle O’Reilly are thanked for assistance with image (2011). production. 68. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Phytoliths in woody plants from the Miombo woodlands of Mozambique. Ann. Bot. (Lond.). 104,91–113 (2009). Author contributions 69. Mercader, J. et al. Poaceae phytoliths from the Niassa Rift, Mozambique. N.B., C.S., R.H., and M.D.P. designed the research; C.S., P.R., M.E.P., J.B., R.H., A.T., N. J. Archaeol. Sci. 37, 1953–1967 (2010). K., A.C., S.J., W.A., R.T., S.A., K.D., F.D., A.P.M., B.R., T.W., J.M., C.C.M., R.C., and C.B. 70. Hunt, C. O. & Hill, E. A. in Molluscs in Archaeology: Methods, Approaches and collected the data; C.S., K.D., N.K., J.B., W.A., P.R., F.D., S.A., A.H., M.E.P., J.M., R.M., T. Applications Vol. 3 (ed. Michael J. Allen) 100–110 (Oxbow Books, Oxford, S.W., B.R., A.P.M., C.C.M., J.L.T., R.C., and P.F. analyzed the data; C.S., P.R., N.B., M.D. 2017). P., T.S.W., J.B., N.K., K.D., A.H., R.M., S.A., F.D., J.M., B.R., M.E.P., and H.S.G. wrote the 71. Levin, N. E., Simpson, S. W., Quade, J., Cerling, T. E. & Frost, S. R. Herbivore paper. enamel carbon isotopic composition and the environmental context of Ardipithecus at Gona, Ethiopia. Geol. Soc. Am. Spec. Pap. 446, 215–234 (2008). 72. Cerling, T. E. et al. Diet of Paranthropus boisei in the early Pleistocene of Additional information EastAfrica. Proc. Natl Acad. Sci. USA 108, 9337–9341 (2011). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 018-04057-3. 73. Uno, K. T. et al. Late Miocene to Pliocene carbon isotope record of differential diet change among East African herbivores. Proc. Natl Acad. Sci. USA 108, 6509–6514 (2011). Competing interests: The authors declare no competing interests. 74. Lee-Thorp, J. A., Van Der Merwe, N. J. & Brain, C. Isotopic evidence for dietary differences between two extinct baboon species from Swartkrans. Reprints and permission information is available online at http://npg.nature.com/ J. Hum. Evol. 18, 183–189 (1989). reprintsandpermissions/ 75. Kingston, J. D. & Harrison, T. Isotopic dietary reconstructions of Pliocene herbivores at Laetoli: Implications for early hominin paleoecology. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in Palaeogeogr. Palaeoclimatol. Palaeoecol. 243, 272–306 (2007). published maps and institutional affiliations. 76. Lee-Thorp, J. & Sponheimer, M. in Handbook of Paleoanthropology (eds Winifried Henke & Ian Tattersall) 441–464 (Springer, Berlin, 2015). 77. Farquhar, G. D., Ehleringer, J. R. & Hubick, K. T. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant. Biol. 40, 503–537 Open Access This article is licensed under a Creative Commons (1989). Attribution 4.0 International License, which permits use, sharing, 78. van der Merwe, N. J. & Medina, E. The canopy effect, carbon isotope ratios adaptation, distribution and reproduction in any medium or format, as long as you give and foodwebs in Amazonia. J. Archaeol. Sci. 18, 249–259 (1991). appropriate credit to the original author(s) and the source, provide a link to the Creative 79. Cerling, T. E., Hart, J. A. & Hart, T. B. Stable isotope ecology in the Ituri Commons license, and indicate if changes were made. The images or other third party Forest. Oecologia 138,5–12 (2004). material in this article are included in the article’s Creative Commons license, unless 80. Koch, P. L., Michener, R. & Lajtha, K. Isotopic study of the biology of modern indicated otherwise in a credit line to the material. If material is not included in the and fossil vertebrates. Stable Isot. Ecol. Environ. Sci. 2,99–154 (2007). article’s Creative Commons license and your intended use is not permitted by statutory 81. Krigbaum, J., Berger, M. H., Daegling, D. J. & McGraw, W. S. Stable isotope regulation or exceeds the permitted use, you will need to obtain permission directly from canopy effects for sympatric monkeys at Taï Forest, Côte d’Ivoire. Biol. Lett. 9, the copyright holder. To view a copy of this license, visit http://creativecommons.org/ 20130466 (2013). licenses/by/4.0/. 82. Flanagan, L. B., Comstock, J. P. & Ehleringer, J. R. Comparison of modeled and observed environmental influences on the stable oxygen and hydrogen isotope composition of leaf water in Phaseolus vulgaris L. Plant. Physiol. 96, © The Author(s) 2018 588–596 (1991). 8 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals
Loading next page...
 
/lp/springer_journal/78-000-year-old-record-of-middle-and-later-stone-age-innovation-in-an-ho4ttfJduP
Publisher
Nature Publishing Group UK
Copyright
Copyright © 2018 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
D.O.I.
10.1038/s41467-018-04057-3
Publisher site
See Article on Publisher Site

Abstract

Corrected: Publisher correction ARTICLE DOI: 10.1038/s41467-018-04057-3 OPEN 78,000-year-old record of Middle and Later Stone Age innovation in an East African tropical forest 1,2,3 4 5,6 7,8 9 4,10 Ceri Shipton , Patrick Roberts , Will Archer , Simon J. Armitage , Caesar Bita , James Blinkhorn , 11 2,12 13 8,14 4,15 Colin Courtney-Mustaphi , Alison Crowther , Richard Curtis , Francesco d’ Errico , Katerina Douka , 16 4,17 18 13 19 Patrick Faulkner , Huw S. Groucutt , Richard Helm , Andy I. R Herries , Severinus Jembe , 20,21 17 10 22 14,23 Nikos Kourampas , Julia Lee-Thorp , Rob Marchant , Julio Mercader , Africa Pitarch Marti , 24 25 19 26 27,28 Mary E. Prendergast , Ben Rowson , Amini Tengeza , Ruth Tibesasa , Tom S. White , 4,29 4 Michael D. Petraglia & Nicole Boivin The Middle to Later Stone Age transition in Africa has been debated as a significant shift in human technological, cultural, and cognitive evolution. However, the majority of research on this transition is currently focused on southern Africa due to a lack of long-term, stratified sites across much of the African continent. Here, we report a 78,000-year-long archeological record from Panga ya Saidi, a cave in the humid coastal forest of Kenya. Following a shift in toolkits ~67,000 years ago, novel symbolic and technological behaviors assemble in a non- unilinear manner. Against a backdrop of a persistent tropical forest-grassland ecotone, localized innovations better characterize the Late Pleistocene of this part of East Africa than alternative emphases on dramatic revolutions or migrations. 1 2 McDonald Institute for Archaeological Research, University of Cambridge, Downing Street Cambridge, Cambridge CB2 3DZ, UK. British Institute in Eastern Africa, Laikipia Road, Kileleshwa, Nairobi, Kenya. Centre of Excellence for Australian Biodiversity and Heritage, Australian National University, Canberra, ACT 2601, Australia. Department of Archaeology, Max Planck Institute for the Science of Human History, Kahlaische Strasse 10, Jena D-07745, Germany. 5 6 Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Pl. 6, Leipzig 04103, Germany. Department of Archaeology, University of Cape Town, Rondebosch 7701 Western Cape, South Africa. Department of Geography, Royal Holloway, University of London, Egham, Surrey TW20 OEX, UK. SSF Centre for Early Sapiens Behavior (SapienCe), University of Bergen, Øysteinsgate 3, Postboks 7805, Bergen 5020, 9 10 Norway. Malindi Museum, National Museums of Kenya, Malindi, Kenya. Department of Archaeology, Classics and Egyptology, University of Liverpool, 12–14 Abercromby Square, Liverpool L69 7WZ, UK. Department Environment, York Institute for Tropical Ecosystems, University of York, Heslington, York 12 13 YO10 5NG, UK. School of Social Sciences, The University of Queensland, St Lucia, QLD 4072, Australia. Department of Archaeology and History, The Australian Archaeomagnetism Laboratory, Palaeoscience Labs, La Trobe University, Melbourne Campus, Bundoora, VIC 3086, Australia. UMR 5199 PACEA, CNRS/Université de Bordeaux, Bâtiment B18, Allée Geoffroy Saint Hilaire, CS, 50023 - 33615 PESSAC CEDEX, France. Research Laboratory for Archaeology and the History of Art, Dyson Perrins Building, South Parks Road, Oxford OX1 3QY, UK. Faculty of Arts and Social Sciences, Department of Archaeology, The University of Sydney, Sydney, NSW, Australia. School of Archaeology, University of Oxford, 36 Beaumont Street, Oxford OX1 2PG, UK. 18 19 Canterbury Archaeological Trust, 92A Broad Street, Canterbury, Kent CT1 2LU, UK. Coastal Forests Conservation Unit, National Museums of Kenya, 20 21 Kilifi, Kenya. Centre for Open Learning, University of Edinburgh, Paterson’s Land, Edinburgh EH8 8AQ Scotland, UK. Biological and Environmental Sciences, University of Stirling, Stirling FK9 4LA Scotland, UK. Department of Anthropology and Archaeology, University of Calgary, 2500 University Drive, Calgary, AB T2N 1N4, Canada. Grup de Recerca Aplicada al Patrimoni Cultural (GRAPAC), Department of Animal Biology, Plant Biology and Ecology, Faculty of Biosciences, Autonomous University of Barcelona (UAB), Campus Bellaterra, Bellaterra 08193, Spain. Department of Sociology and Anthropology, Saint Louis University, Avenida del Valle 34, Madrid 28003, Spain. Invertebrate Biodiversity, National Museum Wales, Cathays Park, Cardiff CF10 3NP, UK. Department of Anthropology and Archaeology, University of Pretoria, cnr Lynnwood Road and Roper Street, Hatfield, South Africa. 27 28 University Museum of Zoology, Cambridge Downing Street, Cambridge CB2 3EJ, UK. Department of Life Sciences, The Natural History Museum, Cromwell Road, London SW7 5BD, UK. Human Origins Program, Smithsonian Institution, Washington, DC 20560, USA. Correspondence and requests for materials should be addressed to C.S. (email: ceri.shipton@anu.edu.au) or to P.R. (email: roberts@shh.mpg.de) or to N.B. (email: boivin@shh.mpg.de) NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 1 | | | 1234567890():,; A ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 he terms Middle Stone Age (MSA) and Later Stone Age forest and savannah (SM5). Excavations at PYS have revealed (LSA) have long been used to frame discussions of beha- exceptional preservation and stratigraphic integrity, and a record Tvioral and cultural change in Africa . Changes in lithic of human activity back to ~78 ka, including a rich technological production (such as making elongate blades and stone-tipped sequence that includes lithic forms elsewhere associated with the 2, 3 4, 5 arrows ), the appearance of symbolic material culture , and MSA and LSA. Alongside a rich range of paleoecological indi- subsistence diversification associated with the MSA and LSA cators, these features mean that the site offers a rare opportunity have all been identified as important thresholds in human cog- to study human behavioral changes in an evolutionarily critical, 3, 6 nitive and social evolution . Many researchers have highlighted but poorly-understood, region of Africa. the revolutionary nature of MSA and LSA human capacities, in some cases arguing that they reflect cognitive evolutionary Results developments or that they stimulated pan-African and global Stratigraphy and chronology. The 3 m deep excavated sequence 8, 9 migrations from 60,000 years ago (ka) onwards . On the other at PYS encompasses 19 layers (Supplementary Note 2) (Fig. 1b). hand, recent discoveries in southern Africa have suggested a more Three lithological boundaries divide the profile into four main 10–13 gradual development of these material culture traits . lithostratigraphic units that are discussed in detail in Supple- The debate as to the significance and tempo of behavioral mentary Note 1. A series of 20 stratigraphically ordered and changes during the MSA and LSA has largely focussed on the internally consistent radiocarbon and optically stimulated lumi- 12, 14 temperate and coastal environments of southern Africa . This nescence (OSL) age estimates, when included in a Bayesian is due to a general lack of well-dated, well-stratified records in model, show human occupation in every Marine Isotope Stage other key areas of the African continent, particularly across (MIS), from late MIS5 c. 78 ka into the Holocene (Supplementary the period 80–40 ka, though the Haua Fteah and Taforalt in Note 3). Geoarchaeological and micromorphological studies 15, 16 North Africa are notable exceptions . Long-term, dated indicate that the sequence consists of fine colluvia, spalling, and records from East Africa remain scarce. Several East African sites anthropogenic deposits with abundant organic and cultural have produced evidence for novel practices, reflected in the microremains (e.g., fauna, flora, lithic microdebitage – Supple- appearance of backed stone tools and beads over the last mentary Notes 4 and 5). 17–22 60–40,000 years . However, the chronologies and environ- Geomorphological and magnetic susceptibility proxies for mental contexts of these key behavioral transitions are not clear. human occupation intensity (Fig. 2; Supplementary Note 2) Sites >50,000 years old have only just begun to be identified indicate a pattern of intermittent pulses of human activity. In the beyond the East Africa Rift System and are limited to the Lake early part of the sequence, in Layers 19, 18, and the lower part of 24–26 Victoria region . Layer 17 (~78–73 ka), occupation intensity is low. This is The new archeological cave site of Panga ya Saidi (PYS) followed by a possible hiatus when both lithics and charcoal >c. described here offers an opportunity to address geographical and 125 μm drop off, and magnetic susceptibility and biogenic input ecological biases in our understanding of early human behavioral signals are dramatically reduced (Fig. 2). Human occupation and cultural change. The site is situated 15 km from the present- proxies subsequently show a general trend of increasing intensity day shoreline in the Zanzibar-Inhambane coastal forest mosaic from Layer 16 (beginning ~67 ka) to the Holocene (Fig. 2). that runs along the East African littoral, with the coastal shelf dropping below −125 m depth within 5 km of the modern coastline (Fig. 1a) (Supplementary Note 1). The cave is in the Cultural artifacts. The long artifactual sequence is comprised Dzitsoni limestone hills on an ecotone between lowland tropical of 17 ocher fragments, eight worked bone artifacts, 88 ostrich Layer ab Date in ka 0.5 1 1 3 7.5 7.5 14.5 PYS PYS 10 10 48.5 12 12 58.5 13 S Sh h 61.5 B Bn n Mombasa 16 71 – 67 S Sh h L Lt t Low coastal plain 76.5 Tropical moist forest Savannas Foot plateau Flooded savannas Ditzoni upland C Ch h Montane grasslands Coastal range 1 m Deserts High coastal plain 0 km 500 km 0 km 20 km OSL sample Radiocarbon Mangroves B Bn n Lakes Ash Burrow 0 mi 500 mi 0 mi 20 mi Fig. 1 Environmental setting and PYS stratigraphic section. a The location of PYS in the tropical moist forest of coastal East Africa, situated in the Ditzoni upland, southeastern Kenya. b The stratigraphic sequence of PYS showing the Layers and modeled ages, with example of a micromorphological thin section, illustrating the rich biogenic/anthropogenic contents in the sediments (Sh shell, Bn bone, Lt lithic, Ch charcoal). Note that Layer 12 was not continuous across the whole excavation and did not occur in this section. Age estimates are shown as the median of the highest posterior density age range for simplicity 2 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | I NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE a d f 40 mm 50 mm b c 50 mm 30 mm 45 mm 35 mm Fig. 2 Selected artifacts from PYS. a Levallois core from Layer 11. b Two backed lithic artifacts from Layer 11. c Backed lithic artifact from Layer 3. d Notched bone from Layer 8. e Notched bone from Layer 9. f Ocher crayon from Layer 10. g Ostrich eggshell bead from Layer 8. h Conus shell bead from Layer 16. i Gastropod shell bead from Layer 4 Stable isotopes Magnetic Mollusks Charcoal 18 13 susceptibility Biogenic content δ O(‰) δ C(‰) species >25um Lithic Lithic Lithic (counts per 0.5cm ) Sedimentology Depth Layer (VPDB) (VPDB) Phytoliths Bovids count nXLF nXFD% diameter Tetrapods density materials weight LM SA LM PSA LM PSA LM PSA Phytoliths LM PS absent below 16 layer 13 Sandier upwards 17 Redox pedo- features LM Sc 100% 0 100% 0 60 0.4 0.8 1.000.4 0 5 0 100 0 100% 0 Artefacts/1 60 100% 0 (g) 5 86 4 2 0 16 12 8 4 0 4 0 Clayey Silty Sand Phytoliths key Bovid key Biogenic key Tetrapods key Lithic materials key Grass Suni, duiker, dik dik Ash intraclast Bovids and suids Chert Bats and rodents Woody Bushbuck, kudu Bone Limestone Other tetrapods Palm Reedbuck, waterbuck Snail shell Quartz Hartebeest, topi, wildebeest Charcoal Buffalo Isotopic particles, Coprolith, quartz lithics, Unspecified large bovids Plant pseudomorph Fig. 3 The palaeoenvironmental and human occupation proxies from PYS. From left to right: Sedimentology (LM(SC) sandy clayey loam, LM(PS) pebbly sandy loam, LM(PSA) pebbly sandy ashy loam, LM(SA) sandy ashy loam); Depth; Layer divisions; Box and whisker plots of stable oxygen and carbon isotope values of mammalian teeth; Phytoliths, showing the proportion of grass, palm, and woody taxa; Percentage of different bovids in Minimum Number of Individuals (MNI); Terrestrial mollusk rarefied species count; Magnetic susceptibility (XLF and XFD%); Biogenic content of micromorphology thin sections; Microcharcoal abundance; Proportions of selected faunal groups as a percentage of total tetrapod MNI; Lithic density; Lithic material types; Lithic weight (mean debitage weight) eggshell beads, 27 marine shell beads, five exotic manuports, as made using variations of the Levallois method, and large retou- well as >30,000 knapped stone artifacts, including Levallois cores ched points (Supplementary Note 4), comfortably fitting with and backed artifacts—typical of MSA and LSA technologies, other contemporary MSA assemblages . Immediately after the 6, 27, 28 respectively (Fig. 3). The metrics and characteristics of hiatus, in Layer 16, there is a shift in rock type proportions from these technologies are discussed in greater detail in Supplemen- principally microcrystalline limestone to cryptocrystalline quartz tary Note 4. The possible hiatus or ephemeral occupation in the and chert, and a concomitant reduction in artifact size (Fig. 2, PYS sequence between 73 and 67 ka corresponds with a change in Supplementary Note 4). Size reduction is evident across all stone the stone artifact sequence. In the early part of the sequence artifacts, including within cryptocrystalline materials and retou- stone artifacts are characterized by large flakes (Fig. 2), often ched tools, demonstrating that this change does not simply reflect NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 3 | | | 2,5 2,0 1,5 1,0 1,0 0,5 m 0,5 m Depth (m) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 variation in procured raw material package size (Supplementary and δ O ranges of fauna after Layer 17 (Fig. 2) (Supplementary Note 4). Data 1)reflects the persistence of C and C vegetation, and a 3 4 Coupled with the switch to small cryptocrystalline stone tools variety of water sources affected to differing extents by after ~67 ka is an increased use of bipolar technology (Supple- evapotranspiration in the vicinity of PYS; likely in the form of mentary Fig. 4). A shift in technological emphasis toward bipolar a forest-grassland ecotonal setting. A consistent ecotonal situation strategies and a reduction in stone tool size is considered a key is supported by zooarchaeological data for the persistent presence 27, 30, 31 marker of LSA behavior elsewhere in Africa . A change of open and bush/forest adapted mammalian species, as well from Levallois to prismatic blade technology and the appearance as phytolith datasets that document the continued occurrence of backed crescentic tools have also been highlighted as indicators of woody, grass, and palm phytoliths in the immediate site 6, 32 of the LSA . However, in the PYS sequence, Levallois environment (Fig. 2). technology occurs alongside backed crescents in Layers 11 and 12 (~51–48 ka) and Layers 5–3 (~14–1 ka) (Supplementary Discussion Note 4) (Fig. 3). Prismatic blade production is rare in the PYS The paleoecological datasets from PYS agree with other East sequence, but bipolar and Levallois blades become common in African records that point to a period of low amplitude envir- the upper part of the sequence (Layers 8–3) (~25–1 ka) onmental change throughout much of MIS4-1, particularly from (Supplementary Note 4). Levallois and bipolar technology, blade records nearer the coast where the maritime influence buffers 33, 34 production, and backed crescentic tools occur recurrently and against temperature extremes . The consistent presence of intermittently, with no evidence for a unilinear accumulation of the forest-grassland ecotone throughout the last ~67,000 years traits or a uniform uptake of the latter three traits as a package. accompanies evidence for increasing occupation intensity at PYS, Despite changes in technology, stone artifacts remain consistently perhaps suggesting a growing human presence in the region small and were predominantly produced on cryptocrystalline linked to the use of mosaic habitats. Moreover, alongside the materials after ~67 ka. magnetic susceptibility data for increased occupation intensity Beads, ocher fragments, and worked bone have been associated from 60 ka, this may imply, as has been suggested elsewhere, that with behavioral complexity in the Late Pleistocene and occur with the MSA–LSA transition of East Africa is a long-term pattern of increasing regularity through the sequence at PYS (Supplemen- 33, 35 change related to growing population densities . tary Note 4). The earliest bead, a Conus sp. shell spire, occurs in Indeed, the PYS sequence does not document a radical change Layer 16 which dates from between ~67–63 ka (Fig. 3). At ~33 ka in technological or cultural behavior in East Africa ~60–50 ka that (Layer 9) the most common beads were Conus shell spires (n = might be suggestive of cognitive or technological “revolutions” or 13) (Fig. 3). Recurrent engagement with coastal resources for 7, 8 migrations . Instead, PYS documents a long-term assembly, and symbolic use is behavioral, rather than geographic, given minimal intermittent presence, of various innovative traits. In particular, changes in distance from the shoreline during the site’s there is no dramatic appearance of an LSA technological package occupation (Fig. 1a) (Supplementary Note 1). Ostrich eggshell and, instead, older MSA technological traits, such as Levallois beads at PYS reach their highest frequency ~25 ka (Layer 8, n = cores, exist alongside the development of backed artifacts 70), while fully-manufactured beads made of marine shells are the and blade production. The principal change in the sequence is dominant types during the Holocene (Layers 1–5) (Fig. 2). Layers the reduction in lithic size and the shift to cryptocrystalline 8-10 (~48–25 ka) produced two modified ocher fragments (Fig. 3), materials ~67 ka. as well as carved bone and tusk artifacts, including a decorated From MIS6-3 Homo sapiens began to adapt to a diversity of 13, 36 37, 38 39, 40 bone tube (Fig. 3) and a small bone point ; artifact types that coastal , tropical forest , and hyper-cold environ- have been argued to be characteristic of the LSA . In contrast ments across Africa and Eurasia. Humans appear to have adapted to many revolutionary or unilinear interpretations of technolo- locally to these environments, gradually developing new symbolic gical and cultural development, the PYS sequence reveals a forms, technological production strategies, and subsistence pattern of intermittent presence of different technological traits behaviors. It seems that the Middle and Late Pleistocene of Africa and symbolic artifacts that have been associated with the MSA is best characterized by diverse Homo sapiens populations, and LSA. adopting a range of survival strategies and new forms of social communication on an intermittent, ad hoc basis in different 41, 42 environmental and climatic contexts . It is this adaptive Paleoecology. Numerous proxies point to broad perseverance plasticity that truly defines the expansion and development of of tropical forest and grassland environments throughout the humans as a global species. sequence (Supplementary Note 5). This is indicated, for example, by the consistently high (> 25 species per sample) terrestrial Methods mollusk diversity, with most of the mollusk species requiring Excavation. Our excavation is in a large rockshelter in the first chamber of the humid shady conditions. collapsed-roof cave, near to the main entrance. Three seasons of excavation Sedimentology and magnetic susceptibility data indicates a between 2010 and 2013 have exposed a 3 m deep sequence, with the trench measuring 3.5 × 2 m at the top stepping into 2 × 1 m at the base. The trenches were shift to drier conditions following the early occupation in Layers excavated using the single context method according to the Museum of London 17–19 ~78–73 ka (Fig. 2) (see also Supplementary Note 5). This is protocols with the addition of total excavated sediment volume recorded for each supported, albeit with a lag, by the stable isotope data which context. Any animal burrows were excavated by hand and their contents discarded. shows higher stable carbon (δ C) in Layers 10–12 relative to In situ excavated deposits were dry-sieved on site through a 5 mm mesh. A program of bulk soil sampling for flotation (0.5 mm) and wet-sieving (1 mm) 13–16 and 17 (Supplementary Tables 12 and 13), indicative of an to recover archeobotanical, zooarcheological, and paleoenvironmental increase in the presence of C resources, most likely in the form of microremains was also implemented (see below). Where possible, a minimum grassland, in the diets of fauna being exploited at the site in the sample volume of 60 litres per context was maintained. Smaller contexts (<60 L) region c. 48 ka. Visually, δ O also tracks this trend but there is were sampled in their entirety. Nineteen stratigraphic layers were identified in the section, each corresponding to a particular excavation context. A continuous no statistically significant difference between Layer 17 and Layers column bulk sediment sample was taken for palaeoenvironmental analyses with 13–16 or Layers 10–12 (Supplementary Tables 14 and 15). 100 g samples taken at every 2 cm of depth. Deposits exposed in PYS Trench 4 From ~67–48 ka, paleoenvironmental proxies at PYS show that (2013 excavation) were logged on-site following standard sedimentological 43–45 the local environment underwent relatively little variation until procedures . Sediment color, texture, composition, structures, postdepositional the final occupation of the cave at ~0.5 ka (Fig. 2). The wide δ C disturbance and the nature, and geometry of layer boundaries were recorded for 4 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE each layer resolved by the excavators. Layers were grouped into higher-order, Fourteen C measurements were produced at the Oxford Radiocarbon multilayer lithostratigraphic units, as shown in Supplementary Fig. 1. Accelerator Unit (ORAU) (n = 10) and by Beta Analytic (n = 4). Most charcoal samples were prepared using the standard acid–base–acid (ABA) protocol . While for the younger material (<30 ka BP) this is usually sufficient, it has been shown Micromorphology. A set of 23 undisturbed micromorphology sediment samples that ABA does not efficiently decontaminate older charcoal samples when were collected from the excavated profile in clear polyurethane boxes. In view of compared with the more rigorous protocol: acid–base oxidation/stepped the large dimensions and stratigraphic complexity of the excavated trench, sam- combustion (ABOx-SC) . Paired ABA and ABOx-SC preparation was used on one pling was at a reconnaissance scale, concentrating on layer boundaries and dis- sample from context 413 C. For this particular sample, ABA and ABOx tinctive features. Sample boxes were labeled, photographed and plotted on the methodologies produced identical AMS ages. This is probably due to the profile drawing before removal from the profile (Supplementary Fig. 1). Out of this exceptional state of preservation of the charcoal in this part of the sequence as well sample set, 10 samples from the Pleistocene part of the profile were processed for as due to the fact that it is not very old sample. For material from lower levels micromorphological analysis at the Thin Section Micromorphology Laboratory, where only ABA dates from Beta Analytic exist these should be considered University of Stirling (sample code PYS; thin sections manufactured by George minimum ages only. This is demonstrated by the much older ages returned from MacLeod). Samples were air-dried and impregnated with polyester (polylite) resin ostrich eggshell samples from similar contexts. following standard procedures (http://www.thin.stir.ac.uk/). Ca. 30 μm thick, It is not possible to interpret radiocarbon ages reliably without calibration, due uncovered, large format thin sections (7.5 × 11 cm) were manufactured from the to variation in the concentration of radiocarbon in the atmosphere through time. hardened impregnated blocks. All terrestrial C measurements were calibrated using IntCal13 and SHCal13, the 55, 56 Thin sections were observed with a polarizing microscope at magnifications of most recent internationally agreed calibration curves available . As PYS lies ×12.5 to ×400, using plain polarized (PPL), cross-polarized (XPL), and oblique close to the equator, a 68.2/31.8 northern/southern hemisphere split was used in incident light (OIL). Relative abundance of sediment/soil components was the calibration curve, taking into account the position of the site relative to the 46, 47 estimated using standard semi-quantitative estimation charts . Key sediment inter-tropical convergence zone. The radiocarbon ages obtained from PYS are constituents larger than ca. 50 μm (mineral grains; pedoclasts; biogenic particles, summarized in Supplementary Table 1. etc.) were point-counted using a 0.5 × 0.5 cm grid overlay printed on clear acetate. Point-counted biogenic particles included bone fragments (from small cave OSL. Optically stimulated luminescence (OSL) samples were obtained by ham- vertebrates and larger vertebrate fauna—the latter possibly including human prey); mering metal tubes into section faces following cleaning. The samples were sealed small vertebrate coproliths; indeterminate isotropic particles; shell fragments; using adhesive tape. Following transport to the Royal Holloway University of charcoal (both woody and non-woody tissue); burnt plant pseudomorphs; ash and London Luminescence Laboratory, samples were processed under subdued orange ash intraclasts; quartz debitage (Supplementary Fig. 2). light. The outer 5 cm of sample (presumed as being exposed to sunlight) was removed and retained for background radioisotope concentration determination. Paleomagnetism. A sub-sample of the continuous column sample from PYS was Quartz was extracted from the part of each sample not exposed to sunlight transported to The Australian Archaeomagnetism Laboratory for preparation and following burial. As the bedrock at PYS is primarily composed of carbonate, analysis. Once in the laboratory samples were dried to standardize water content, samples were initially wet-sieved to isolate the 212–180 µm size fraction. This crushed with a non-magnetic mortar, and pestle to become homogenized and then removes large bedrock clasts from the sample before acid treatment, meaning that packed into standard 8cc palaeomagnetic plastic cubes. The approach to the the possibility of incorporating grains liberated by dissolution of the bedrock was analysis follows that of Herries and a number of analyses were run to establish minimized. the magnetic mineralogy of the samples. These included mass specific low-field The volume of 1 M HCl and H O were used to remove carbonates and organic 2 2 susceptibility (χlf), mass specific high-field susceptibility (χhf), mass specific fre- matter from the 212–180 µm fraction, respectively. The samples were then re- quency dependant susceptibility (χfd) and saturation isothermal remanent mag- sieved at 180 µm and quartz extracted from the >180 µm fraction using density netization acquisition curves and backfields. This was done to understand the separations at 2.62 and 2.70 g/cm followed by a HF acid etch (23 M HF for 60 min mineralogy driving magnetic susceptibility change through identifying ferrimag- followed by 10 M HCl rinse). The resulting, etched samples were sieved at 150 µm netic vs. anti-ferromagnetic minerals on the basis of their coercivity, establish the to remove partially dissolved grains. All samples were then stored in opaque concentration of magnetic minerals present within each sample, and examine grain containers prior to measurement. size and domain state trends in the sequence. All OSL measurements were carried out using a Risø TL/OSL-DA-15 57 58, 59 The magnetic susceptibility of each sample was measured using a Bartington automated dating system , fitted with a single-grain OSL attachment . Single- MS2 susceptibility-meter connected to an MS2B sensor. Samples were measured at grains were stimulated using a 10 mW Nd: YVO4 solid-state diode-pumped green 2 57 0.47 kHz (low, χlf) and 4.7 kHz (high, χhf) to attain both low and high frequency laser (532 nm) focused to yield a nominal power density of 50 W/cm . All susceptibility values. These measurements were then used to compute the infrared (IR) stimulation was carried out using an IR (870 nm) laser diode array frequency dependence of the magnetic susceptibility (χfd) and then expressed as a yielding a power density of 132 mW/cm . OSL passed through 7.5 mm of Hoya percentage (χfd%) using the formula stated by Dearing et al. . Isothermal U-340 filter and was subsequently detected using an Electron Tubes Ltd 9235QB15 remanent magnetizations (IRM) were induced up to 1 T (representing the photomultiplier tube. saturation IRM: SIRM) using a magnetic measurements pulse magnetizer Irradiation was carried out using a 40 mCi 90 Sr/90Y beta source providing ~6 (MMPM10) and measurements made on an AGICO JR6 magnetometer. Forward- Gy/min. This source is calibrated relative to the National Physical Laboratory, field measurements were taken at 20, 500, 600 mT, and 1 T and back-field Teddington 60Co γ-source (Hotspot 800) . Due to the spatial inhomogeneity of measurements at 20, 40, 100, 150, 200, 250, and 300 mT. Full IRM acquisition beta emitters across the active face of our 90 Sr/90Y beta source we calibrated the curves and backfields were also produced for each layer of the site. HIRM dose rate to each individual grain position on a single-grain disc using the 50 62 measurements were also taken using the method described by Liu et al. . Soft IRM method reported by Armitage et al. . For more detail on measurement and quality measurements showing the concentration of ferrimagnetic minerals were taken at criteria see Supplementary Note 3. 20 mT. S-ratios (IRM-300 mT/SIRM and IRM-100 mT/SIRM) were calculated to establish variation in the grain size of ferromagnetic minerals. Remanence of Bayesian methods. The absolute age determinations were used to construct an age coercivity values were also calculated to further establish the mineralogy, grain size, model using Bayesian software (OxCal 4.3 ) and the INTCAL13 curve. The and domain states of the magnetic minerals present within the samples. determinations were input as values in fraction modern (fM) plus or minor fM errors at 1σ (R_F14C in OxCal). In order to determine whether there are pro- blematic determinations that do not agree with the prior framework, an outlier Charcoal abundance. From the continuous column sample described in SM3, detection method was applied. When there is a lack of agreement with the prior subsamples of 1 cm were extracted for charcoal abundance analysis. The addition framework, significant outlier results allow us to quantify the degree of difference. of sodium hexametaphosphate to a beaker containing the sample was used to Values excessively higher than the prior outlier probabilities applied are auto- disaggregate the samples and aid in the separation of the organic material and the matically down-weighted in the models. A posterior outlier probability of 0.5 clay particles . The contents of the beaker was then passed through a 125 μm sieve means that the radiocarbon likelihood of the sample is only included in half of the and the trapped content transferred to a gridded Petri dish. Pieces of charcoal were runs of the model. The two Beta dates were assigned a 0.3 value to reflect the identified using comparative collections and the total charcoal count was deter- inadequate chemical protocol applied in the decontamination of these samples mined through visual inspection and manipulation with a metal probing needle prior to measurement. under a Zeiss Stemi 2000-C optical stereomicroscope (×10–40 magnifications). The Only age estimates older than 20,000 years ago were included in the model. The microscopic charcoal size identified represents charcoal that was locally produced inclusion of younger ages does not affect the older ages at all. Younger ages were during fires within the catchment area of the site . excluded because of the large span of the dates; the model runs with a broad resolution of 50–100 years for the older data, whereas 20 years would be more Radiocarbon dating. Items selected for radiocarbon dating were a human bone appropriate for the younger data. In each Bayesian model, a start and end and a charred sorghum seed (identified by Alison Crowther) from the upper part of boundary was added in order to bracket the archeological phases within the the sequence; as well as unidentified charcoal, and ostrich eggshell pieces including sequence. The posterior distributions of these boundaries facilitated determination a bead, from the middle part of the sequence. These dating samples were either of probability distribution functions (PDF) for the beginning and ending of these recovered during excavation or taken from the section at the end of excavation. phases of activity. Due to the presence of a depositional hiatus represented in the NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 5 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 later part of Layer 17, we used two boundaries between the end of 17 and start of one articular surface and/or key landmark, enabling identification to element and Layer 16. In addition, due to uncertainties related to the reliability of sample OSL-7 usually to taxon or at least taxonomic group (e.g., “bird,”“small carnivore”). (see text above) these age estimates were not included in the model. MaxID specimens were identified in all contexts in both trenches. Second, minimally identifiable (minID) bones include limb shafts and axial fragments that can generally be identified at least to carcass size; these were identified in nine high- Lithics. A total of 30,420 lithic artifacts were recovered through the excavation priority contexts from Trench 4, in order to obtain a sample for taphonomic seasons at PYS between 2010 and 2013. The analyses proceeded by classifying all analysis. Third, nonidentified (NID) bones were separated from minID specimens lithics in accordance with stratigraphic context, raw-material type, and technolo- and weighed in these selected contexts. On average across all contexts in Trench 4, gical class. All artifacts were subsequently weighed and counted according to these 10% of the assemblage was maximally identifiable. When minID specimens categories. Cores and retouched artifacts were further classified in accordance with identified in nine selected contexts were included the average identification rate reduction strategy and typology. For unretouched flakes, blades, Levallois flakes, rose to 37%. 64 65 and bipolar flakes were counted. Levallois , bipolar , laminar, and discoidal Taxonomic identifications were made on the basis of extensive reference strategies as well as resultant blanks were all documented to varying frequencies in collections housed at the National Museums of Kenya (Nairobi) Osteology Unit. a number of the layers. Calculations of the minimum number of individuals (MNI) were made using the resulting database following completion of this analysis and took into account specimen laterality, size, and where relevant, age estimates. It should be stressed Beads, osseous artifacts, and ocher. Over two hundred potential beads, bone tools, engraved bone and stone objects, and pigment lumps recovered during that limb shafts were only studied in selected contexts and therefore could not be excavation, were examined under a low power reflected light microscope in search used to calculate the minimum number of elements (MNE) or the resulting MNI for anthropogenic modifications. When necessary, sediment was carefully removed values for layers or phases, as is standard practice in contexts where density- under the microscope with a soft brush or a wet tooth pick. This resulted in the mediated attrition has occurred. It is therefore possible that MNI estimates will be retention of 159 pieces bearing compelling traces of manufacture and use, too low in some instances. unmodified or marginally modified shell fragments probably used as beads, and modified and unmodified lumps of iron-rich rock and sediments, possibly used to extract ocher powder. Stable carbon and oxygen isotope analysis of mammalian tooth enamel. Stable The retained artifacts were examined at magnifications between ×4 and ×40, isotope analysis of mammalian tissues has frequently been used to assess the diets and photographed with a motorized Leica Z6 APOA microscope equipped with a 71–73 and ecologies of East African fossil fauna . This work primarily relies on the Leica Application Suite (LAS) and Multifocus module, and Leica Map DCM 3D distinction between the C -or C -photosynthetic pathways at the base of East 3 4 computer software. The Multifocus module enables the acquisition of extended African foodwebs. In the context of tropical and sub-tropical forest ecologies, this depth of field images. Once the digital images had been complete for different distinction can be used to assess the degree of faunal reliance on C forest resources heights, algorithms in the software compile them into a single composite images as opposed to C plant resources available in open habitats, with C plants being 4 4 that significant extends the depth of field, and provides clarity in viewing the entire 13 74–76 enriched in C relative to C plants . object. This distinction is further enhanced by the “canopy effect” whereby vegetation The selected areas of one Conus shell were scanned using a Sensofar Sneox growing under a closed forest canopy is strongly depleted in C (with scanning confocal microscope with a ×20 objective. The resulting files were correspondingly lower measured δ C) due to low light and the presence of large analyzed with Mountains 7.2 software. For further details on the analytical 77, 78 amounts of respired CO . This results in the tissues of animals consuming protocols for beads, osseuous artifacts, and ocher see Supplementary Note 4 and forest vegetation, as well as forest herbivores, having lower δ C values than the references therein. A full description of all osseous artifacts, beads, and used animals pending some, or all, of their time consuming open-habitat foodstuffs ocher will be reported elsewhere, but Fig. 3 shows examples of each of these artifact e.g. . types from PYS excavation, and Supplementary Figure 19 gives an overview of their Stable oxygen isotope measurements from mammalian enamel can yield further distribution through the sequence. 80, 81 paleoecological information about water and food . Given a constant source of water, plant water δ O will primarily reflect the impacts of relative humidity on Phytolith analysis. Seventeen sediment samples from the continuous stratigraphic leaf water evapotranspiration, with decreasing humidity resulting in increased δ O 82–84 column described in SM3 were processed for phytolith analysis. Of these, the lower values . In a tropical or sub-tropical setting, increased forest cover, and the six samples were barren. We followed extraction protocols employed on Middle resulting shade and increased humidity, will lead to decreased evapotranspiration 66 67 18 79 Stone Age sites from adjacent countries and modern topsoils . The procedure and therefore decreased δ O, especially on the forest floor . As faunal tooth 18 18 included sieving, drying, deflocculating, acid/base treatment, and sequential density enamel δ O primarily reflects water and food-water δ O, herbivores feeding and separation by manipulating the specific gravity of sodium polytungstate. Aliquot drinking in forests can be expected to have lower enamel δ O than those feeding mounting was with “Entellan New,” which allowed for microscopic inspection in open, irradiated areas. This is complicated by physiological and behavioral (×40) and 3D rotation before drying. The average count per slide was 238 phy- variables . Nevertheless, animals consuming plants and water in a shaded, forested toliths. The inferential baseline was grounded on East African phytoliths from setting will broadly reflect the corresponding lower levels of evapotranspiration in 68, 69 plants and soils . Preservation was adequate for morphometric analysis and their enamel δ O. type identification. Stable carbon and oxygen isotope analysis of faunal tooth enamel excavated from the various archeological Phases at PYS was undertaken in order to directly assess the diets and ecologies of animals being exploited by humans living at the Mollusk analysis. Macro marine molluskan remains were identified and counted site at different points in time. This should, in turn, provide information regarding by Patrick Faulkner using comparative modern reference collections. Since no fluctuations in the degree of forest cover surrounding PYS in the past. Faunal evidence for subsistence on these was found prior to Layer 6, they will be reported enamel samples were taken from all available PYS Layers. A broad selection of in detail elsewhere. PYS is an open-roofed cave with substantial input from the species was sampled for each Layer based on availability. Where possible, up to five external environment. The paleoenvironmental samples recovered from PYS members of each species/genus were sampled per layer grouping (Supplementary contained terrestrial mollusk fauna representing the area in and immediately Data 1). Faunal samples were identified to species and/or genus level using the outside the cave over a long time period . There is no evidence that any of these substantial reference collection available at the National Museums of Kenya, species have been transported to the site through natural or anthropogenic Nairobi. The full list of faunal samples and tooth identifications analyzed in this processes. study are shown in Supplementary Data 1. Excluding large marine shells that form part of the archeological record, the Air-abrasion was used to remove any adhering detrital material from the teeth aquatic component of the assemblage is extremely small and is less significant than or tooth fragments to be studied. Gentle abrasion with a diamond-tipped drill was that seen in faunas from the modern coastal forests on Zanzibar. As a result, performed along the full length of the buccal surface of the tooth or tooth fragment transport by fluvial activity or regular flooding of the area can be ruled out. The in order to maximize the period of formation represented by the resulting isotopic diverse community of snails could not have been sustained in a closed cave analysis. The resulting enamel powder was pretreated using standard, published environment but would have required continual external input in the form of leaf protocols in order to remove any organic or secondary carbonate contaminates. litter and the snails themselves or their shells. The non-marine mollusk record This consisted of a wash in 1.5% sodium hypochlorite for 60 min, followed by three from PYS therefore offers a long record of the local environment through time. rinses in purified H O. A volume of 0.1 M acetic acid was then added for 10 min 86, 87 prior to another three rinses in purified H O . Tetrapod analysis. The zooarchaeological analysis of PYS osseous remains took Gases were evolved from the treated samples using 100% Phosphoric Acid. t 13 18 place in 2012 (Trench 3) and 2014 (Trench 4) and produced a database of 5256 δ C and δ O of the resulting gases was measured using a Thermo Gas Bench 2 identified specimens (Number of Identified Specimens, NISP). A total of 6.4 kg of connected to a Thermo Delta V Advantage Mass Spectrometer at the Division of bone was excavated from Trench 3 and 14.7 kg from Trench 4. The Trench 4 study Archeological, Geographic and Environmental Sciences Bradford University. δ C included both taxonomic identification and taphonomic analysis, but microfauna and δ O measurements from samples were compared against international were not analyzed in Phase 1 due to time constraints. In Trench 3, microfauna were standards (NBS 19 and CO-9) registered by the International Atomic Energy identified to a very general level (e.g., “Muridae”). Agency (five of each for a run of 60). Replicate analysis of in-house OES and Initial sorting of faunal remains created up to three categories: first, maximally MERCK standards (six of each for a run of 60) suggests that machine measurement 13 18 identifiable (maxID) specimens include teeth and those bones that preserve at least error is c. ±0.1‰ for δ C and ±0.2‰ for δ O. 6 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 ARTICLE Analysis of variance (ANOVA) tests were performed on faunal enamel δ C 23. Groucutt, H. S. et al. Rethinking the dispersal of Homo sapiens out of Africa. and δ O to determine whether these isotopic parameters differed between groups Evolut. Anthropol.: Issues, News, Rev. 24, 149–164 (2015). of stratigraphic layers. The stratigraphic layers were grouped based on meaningful 24. Faith, J. T. et al. Paleoenvironmental context of the Middle Stone Age chronological and stratigraphic divisions as follows: 1–3, 4–6, 7–9, 10–12, 13–16, record from Karungu, Lake Victoria Basin, Kenya, and its implications and 17. Where variance was found to be significant following ANOVA testing, a for human and faunal dispersals in East Africa. J. Hum. Evol. 83,28–45 post-hoc Tukey pair-wise comparison was performed in order to determine which (2015). layer groupings were significantly different from each other in terms of δ C and 25. Tryon, C. A. et al. The Pleistocene prehistory of the Lake Victoria basin. Quat. 18 88 δ O. All statistical analyses were conducted using the free program R software . Int. 404, 100–114 (2016). 26. Blegen, N., Faith, J. T., Mant-Melville, A., Peppe, D. J. & Tryon, C. A. The Data availability. The authors declare that all data supporting the findings of this Middle Stone Age after 50,000 years ago: new evidence from the Late study are available upon request from the authors. The artifacts and faunal remains Pleistocene sediments of the eastern Lake Victoria basin, Western Kenya. from the Panga ya Saidi excavation are curated in the National Museum of Kenya, PaleoAnthropology 139, 139–169 (2017). Nairobi, under the site code PYS and the suffixes 10, 11, and 13 (denoting the year 27. Villa, P. et al. Border Cave and the beginning of the Later Stone Age in South of excavation). Africa. Proc. Natl Acad. Sci. USA 109, 13208–13213 (2012). 28. d’Errico, F. et al. Early evidence of San material culture represented by organic artifacts from Border Cave, South Africa. Proc. Natl Acad. Sci. USA 109, Received: 23 June 2017 Accepted: 29 March 2018 13214–13219 (2012). 29. Tryon, C. A. & Faith, J. T. Variability in the Middle Stone Age of eastern Africa. Curr. Anthropol. 54, S234–S254 (2013). 30. Eren, M. I., Diez-Martin, F. & Dominguez-Rodrigo, M. An empirical test of References the relative frequency of bipolar reduction in Beds VI, V, and III at Mumba 1. Goodwin, A. & vanRiet Lowe, C. The Stone Age cultures of South Africa. Ann. Rockshelter, Tanzania: implications for the East African Middle to Late Stone South Afr. Mus. 27,1–289 (1929). Age transition. J. Archaeol. Sci. 40, 248–256 (2013). 2. Lombard, M. Quartz-tipped arrows older than 60 ka: further use-trace 31. Pargeter, J. Lithic miniaturization in Late Pleistocene southern Africa. evidence from Sibudu, KwaZulu-Natal, South Africa. J. Archaeol. Sci. 38, J. Archaeol. Sci.: Rep. 10, 221–236 (2016). 32. Tryon, C. A. et al. Middle and Later Stone Age chronology of Kisese II 1918–1930 (2011). 3. Henshilwood, C. S. Late Pleistocene techno-traditions in southernAfrica: a rockshelter (UNESCO World Heritage Kondoa Rock-Art Sites), Tanzania. review of the Still Bay and Howiesons Poort, c. 75–59 ka. J. World Prehistory PLOS One 13, e0192029 (2018). 25, 205–237 (2012). 33. Blome, M. W., Cohen, A. S., Tryon, C. A., Brooks, A. S. & Russell, J. The 4. Wadley, L. Those marvellous millennia: the Middle Stone Age of southern environmental context for the origins of modern human diversity: a synthesis Africa. Azania.: Archaeol. Res. Afr. 50, 155–226 (2015). of regional variability in African climate 150,000–30,000 years ago. J. Hum. 5. Jacobs, Z. et al. Ages for the Middle Stone Age of southern Africa: implications Evol. 62, 563–592 (2012). 34. Moernaut, J. et al. The seismic-stratigraphic record of lake-level fluctuations in for human behavior and dispersal. Science 322, 733–735 (2008). 6. Ambrose, S. H. Small things remembered: origins of early microlithic industries Lake Challa: hydrological stability and change in equatorial East Africa over in sub‐Saharan Africa. Archeol. Pap. Am. Anthropol. Assoc. 12,9–29 (2002). the last 140kyr. Earth Planet. Sci. Lett. 290, 214–223 (2010). 7. Klein, R. G. Archeology and the evolution of human behavior. Evolut. 35. Tryon, C. A. & Faith, J. T. A demographic perspective on the Middle to Later Anthropol. Issues News Rev. 9,17–36 (2000). Stone Age transition from Nasera rockshelter, Tanzania. Philos. Trans. R. Soc. 8. Mellars, P. Why did modern human populations disperse from Africa ca. B 371, 20150238 (2016). 60,000 years ago? A new model. Proc. Natl Acad. Sci. USA 103, 9381–9386 36. O’Connor, S., Louys, J., Kealy, S. & Samper Carro, S. C. Hominin dispersal and settlement east of Huxley’s Line: the role of sea level changes, island size, (2006). 9. Shea, J. J. & Sisk, M. L. Complex projectile technology and Homo sapiens and subsistence behavior. Curr. Anthropol. 58, S567–S582 (2017). dispersal into western Eurasia. PaleoAnthropology 2010, 100–122 (2010). 37. Mercader, J. Forest people: the role of African rainforests in human evolution 10. McBrearty, S. & Brooks, A. S. The revolution that wasn’t: a new interpretation and dispersal. Evol. Anth. 11, 117–124 (2002). 38. Roberts, P. et al. Direct evidence for human reliance on rainforest resources in of the origin of modern human behavior. J. Hum. Evol. 39, 453–563 (2000). 11. Conard, N. J. & Will, M. Examining the causes and consequences of short- late Pleistocene Sri Lanka. Science 347, 1246–1249 (2015). term behavioral change during the Middle Stone Age at Sibudu, South Africa. 39. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western Siberia. Nature 514, 445–449 (2014). PLOS One 10, e0130001 (2015). 12. Henshilwood, C. S. et al. A 100,000-year-old ochre-processing workshop at 40. Fiedel, S. J. & Kuzmin, Y. V. Radiocarbon date frequency as an index of Blombos Cave, South Africa. Science 334, 219–222 (2011). intensity of Paleolithic occupation of Siberia: did humans react predictably to 13. Marean, C. W. et al. Early human use of marine resources and pigment in climate oscillations? Radiocarbon 49, 741–756 (2007). 41. Lombard, M. & Parsons, I. What happened to the human mind after the South Africa during the Middle Pleistocene. Nature 449, 905–908 (2007). 14. Marean, C. W. Pinnacle Point Cave 13B (Western Cape Province, South Howiesons Poort? Antiquity 85, 1433–1443 (2011). Africa) in context: the Cape floral kingdom, shellfish, and modern human 42. d’Errico, F. et al. Identifying early modern human ecological niche expansions and associated cultural dynamics in the South African Middle Stone Age. Proc. origins. J. Hum. Evol. 59, 425–443 (2010). 15. Bouzouggar, A. et al. 82,000-year-old shell beads from North Africa and Natl Acad. Sci. USA 114, 7869–7876 (2017). implications for the origins of modern human behavior. Proc. Natl Acad. Sci. 43. Goldberg, P. & Macphail, R. Practical and Theoretical Geoarchaeology USA 104, 9964–9969 (2007). (Blackwell Publishing Ltd, Hoboken, NJ, 2006). 44. Tucker, M. E. Sedimentary Rocks in the Field: A Practical Guide(John Wiley & 16. Douka, K. et al. The chronostratigraphy of the Haua Fteah cave (Cyrenaica, northeast Libya). J. Hum. Evol. 66,39–63 (2014). Sons, Chichester, 2011). 17. Gliganic, L. A., Jacobs, Z., Roberts, R. G., Domínguez-Rodrigo, M. & Mabulla, 45. Ayala, G., Canti, M., Heathcote, J., Sidell, J. & Usai, M. Geoarchaeology: Using Earth Sciences to Understand the Archaeological Record (English Heritage, A. Z. New ages for Middle and Later Stone Age deposits at Mumba rockshelter, Tanzania: optically stimulated luminescence dating of quartz and Swindon, 2007). feldspar grains. J. Hum. Evol. 62, 533–547 (2012). 46. Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G.. & Tursina, T. Handbook for 18. Ambrose, S. H. Chronology of the Later Stone Age and food production in Soil Thin Section Description (Waine Research, Albrighton, 1985). 47. Stoops, G. Guidelines for Analysis and Description of Soil and Regolith Thin East Africa. J. Archaeol. Sci. 25, 377–392 (1998). 19. Brandt, S. A. et al. Early MIS 3 occupation of Mochena Borago Rockshelter, Sections (Soil Science Society of America Inc., Madison, WI, 2003). Southwest Ethiopian Highlands: implications for Late Pleistocene archaeology, 48. Herries, A. I. Archaeomagnetic evidence for climate change at Sibudu Cave. paleoenvironments and modern human dispersals. Quat. Int. 274,38–54 South. Afr. Humanit. 18, 131–147 (2006). (2012). 49. Dearing, J. A. et al. Frequency-dependent susceptibility measurements of 20. Pleurdeau, D. et al. Cultural change or continuity in the late MSA/Early LSA environmental materials. Geophys. J. Int. 124, 228–240 (1996). of southeastern Ethiopia? The site of Goda Buticha, Dire Dawa area. Quat. Int. 50. Liu, Q. et al. Environmental magnetism: principles and applications. Reviews of Geophysics 50, RG4002 (2012). 343, 117–135 (2014). 21. Miller, J. M. & Willoughby, P. R. Radiometrically dated ostrich eggshell beads 51. Bamber, R. Sodium hexametaphosphate as an aid in benthic sample sorting. from the Middle and Later Stone Age of Magubike Rockshelter, southern Mar. Environ. Res. 7, 251–255 (1982). Tanzania. J. Hum. Evol. 74, 118–122 (2014). 52. Lynch, J. A., Clark, J. S. & Stocks, B. J. Charcoal production, dispersal, and 22. Diez-Martín, F. et al. The MSA/LSA technological transition in East Africa. deposition from the Fort Providence experimental fire: interpreting fire New data from Mumba Rocksheiter Bed V and their implications in the origin regimes from charcoal records in boreal forests. Can. J. For. Res. 34, of modern human behaviour. J. Afr. Archaeol. 7, 147–173 (2009). 1642–1656 (2004). NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications 7 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04057-3 53. Brock, F., Higham, T., Ditchfield, P. & Bronk Ramsey, C. Current 83. Yakir, D., Berry, J., Giles, L. & Osmond, C. B. Isotopic heterogeneity of water pretreatment methods for AMS radiocarbon dating at the Oxford in transpiring leaves: identification of the component that controls the δ Oof Radiocarbon Accelerator Unit (ORAU). Radiocarbon 52, 103–112 (2010). atmospheric O and CO . Plant Cell Environ. 17,73–80 (1994). 2 2 54. Bird, M. et al. Radiocarbon dating of “old” charcoal using a wet oxidation, 84. Sheshshayee, M. et al. Oxygen isotope enrichment (Δ O) as a measure of stepped-combustion procedure. Radiocarbon 41, 127–140 (1999). time-averaged transpiration rate. J. Exp. Bot. 56, 3033–3039 (2005). 55. Reimer, P. J. et al. IntCal13 and Marine13 radiocarbon age calibration curves 85. Levin, N. E., Cerling, T. E., Passey, B. H., Harris, J. M. & Ehleringer, J. R. A 0–50,000 years cal BP. Radiocarbon 55, 1869–1887 (2013). stable isotope aridity index for terrestrial environments. Proc. Natl Acad. Sci. 56. Hogg, A. G. et al. SHCal13 Southern Hemisphere calibration, 0–50,000 years USA 103, 11201–11205 (2006). cal BP. Radiocarbon 55,1–15 (2013). 86. Sponheimer, M. et al. Hominins, sedges, and termites: new carbon isotope 57. Bøtter-Jensen, L., Andersen, C., Duller, G. & Murray, A. Developments in data from the Sterkfontein valley and Kruger National Park. J. Hum. Evol. 48, radiation, stimulation and observation facilities in luminescence 301–312 (2005). measurements. Radiat. Meas. 37, 535–541 (2003). 87. Lee-Thorp, J. et al. Isotopic evidence for an early shift to C4 resources by 58. Duller, G., Bøtter-Jensen, L. & Murray, A. Combining infrared-and green- Pliocene hominins in Chad. Proc. Natl Acad. Sci. USA 109,20369–20372 (2012). laser stimulation sources in single-grain luminescence measurements of 88. R Core Team. 2013. R: A language and environment for statistical computing. feldspar and quartz. Radiat. Meas. 37, 543–550 (2003). R Foundation for Statistical Computing, Vienna, Austria. 59. Duller, G., Bøtter-Jensen, L., Murray, A. & Truscott, A. Single grain laser luminescence (SGLL) measurements using a novel automated reader. Nucl. Acknowledgements Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater. At. 155, 506–514 Funding was provided by the SEALINKS project under a European Research Council (1999). (ERC) grant (no. 206148) awarded to N.B. Permission to conduct the research was 60. Armitage, S. & Bailey, R. The measured dependence of laboratory beta dose granted by the Office of the President of the Republic of Kenya through affiliation with rates on sample grain size. Radiat. Meas. 39, 123–127 (2005). the National Museums of Kenya (NMK). We are grateful for the support of the NMK 61. Ballarini, M., Wintle, A. & Wallinga, J. Spatial variation of dose rate from beta and the British Institute in Eastern Africa. P.R. was funded by NERC and the Boise Fund sources as measured using single grains. Anc. TL 24,1–8 (2006). (University of Oxford). S.J.A. and F.D. acknowledge support from the Research Council 62. Armitage, S. J. et al. The southern route “out of Africa”: evidence for an early of Norway, through its Centres of Excellence funding scheme, SFF Centre for Early expansion of modern humans into Arabia. Science 331, 453–456 (2011). Sapiens Behaviour (SapienCE) (no. 262618). FD and AP were funded by the ERC grant, 63. OxCal version 4.3.2 (2017). TRACSYMBOLS (no. 249587), and the Agence Nationale de la Recherche (ANR-10- 64. Boëda, E. Le concept Levallois, variabilité des méthodes (CNRS éditions, Paris, LABX-52), LaScArBx Cluster of Excellence. A.P.M. holds a Beatriu de Pinós postdoctoral 1994). fellowship (2014 BP-A 00122) from the Agency for Management of University and 65. de la Peña, P. A qualitative guide to recognize bipolar knapping for flint and Research Grants, Government of Catalonia. A.C. and H.S.G. were funded by the British quartz. Lithic Technol. 40, 316–331 (2015). Academy. Additional support has been provided by the McDonald Institute for 66. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Phytoliths Archeological Research (University of Cambridge) and the Max Planck Society. For from Middle Stone Age habitats in the Mozambican Rift (105–29 ka). J. Hum. assistance in the field and with artifact analyses, we wish to thank Jackson Mupe, Yahya Evol. 64, 328–336 (2013). Lenga, Emmanuel Mupe, Mohammed Lenga, Louise Green, John Mpangarusyia, Tim 67. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Soil McEnright, James Munene, Doreen Mutoro, Joyce Wathira, and Emma Le Vraux. phytoliths from miombo woodlands in Mozambique. Quat. Res. 75, 138–150 Annette Guenzel and Michelle O’Reilly are thanked for assistance with image (2011). production. 68. Mercader, J., Bennett, T., Esselmont, C., Simpson, S. & Walde, D. Phytoliths in woody plants from the Miombo woodlands of Mozambique. Ann. Bot. (Lond.). 104,91–113 (2009). Author contributions 69. Mercader, J. et al. Poaceae phytoliths from the Niassa Rift, Mozambique. N.B., C.S., R.H., and M.D.P. designed the research; C.S., P.R., M.E.P., J.B., R.H., A.T., N. J. Archaeol. Sci. 37, 1953–1967 (2010). K., A.C., S.J., W.A., R.T., S.A., K.D., F.D., A.P.M., B.R., T.W., J.M., C.C.M., R.C., and C.B. 70. Hunt, C. O. & Hill, E. A. in Molluscs in Archaeology: Methods, Approaches and collected the data; C.S., K.D., N.K., J.B., W.A., P.R., F.D., S.A., A.H., M.E.P., J.M., R.M., T. Applications Vol. 3 (ed. Michael J. Allen) 100–110 (Oxbow Books, Oxford, S.W., B.R., A.P.M., C.C.M., J.L.T., R.C., and P.F. analyzed the data; C.S., P.R., N.B., M.D. 2017). P., T.S.W., J.B., N.K., K.D., A.H., R.M., S.A., F.D., J.M., B.R., M.E.P., and H.S.G. wrote the 71. Levin, N. E., Simpson, S. W., Quade, J., Cerling, T. E. & Frost, S. R. Herbivore paper. enamel carbon isotopic composition and the environmental context of Ardipithecus at Gona, Ethiopia. Geol. Soc. Am. Spec. Pap. 446, 215–234 (2008). 72. Cerling, T. E. et al. Diet of Paranthropus boisei in the early Pleistocene of Additional information EastAfrica. Proc. Natl Acad. Sci. USA 108, 9337–9341 (2011). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 018-04057-3. 73. Uno, K. T. et al. Late Miocene to Pliocene carbon isotope record of differential diet change among East African herbivores. Proc. Natl Acad. Sci. USA 108, 6509–6514 (2011). Competing interests: The authors declare no competing interests. 74. Lee-Thorp, J. A., Van Der Merwe, N. J. & Brain, C. Isotopic evidence for dietary differences between two extinct baboon species from Swartkrans. Reprints and permission information is available online at http://npg.nature.com/ J. Hum. Evol. 18, 183–189 (1989). reprintsandpermissions/ 75. Kingston, J. D. & Harrison, T. Isotopic dietary reconstructions of Pliocene herbivores at Laetoli: Implications for early hominin paleoecology. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in Palaeogeogr. Palaeoclimatol. Palaeoecol. 243, 272–306 (2007). published maps and institutional affiliations. 76. Lee-Thorp, J. & Sponheimer, M. in Handbook of Paleoanthropology (eds Winifried Henke & Ian Tattersall) 441–464 (Springer, Berlin, 2015). 77. Farquhar, G. D., Ehleringer, J. R. & Hubick, K. T. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant. Biol. 40, 503–537 Open Access This article is licensed under a Creative Commons (1989). Attribution 4.0 International License, which permits use, sharing, 78. van der Merwe, N. J. & Medina, E. The canopy effect, carbon isotope ratios adaptation, distribution and reproduction in any medium or format, as long as you give and foodwebs in Amazonia. J. Archaeol. Sci. 18, 249–259 (1991). appropriate credit to the original author(s) and the source, provide a link to the Creative 79. Cerling, T. E., Hart, J. A. & Hart, T. B. Stable isotope ecology in the Ituri Commons license, and indicate if changes were made. The images or other third party Forest. Oecologia 138,5–12 (2004). material in this article are included in the article’s Creative Commons license, unless 80. Koch, P. L., Michener, R. & Lajtha, K. Isotopic study of the biology of modern indicated otherwise in a credit line to the material. If material is not included in the and fossil vertebrates. Stable Isot. Ecol. Environ. Sci. 2,99–154 (2007). article’s Creative Commons license and your intended use is not permitted by statutory 81. Krigbaum, J., Berger, M. H., Daegling, D. J. & McGraw, W. S. Stable isotope regulation or exceeds the permitted use, you will need to obtain permission directly from canopy effects for sympatric monkeys at Taï Forest, Côte d’Ivoire. Biol. Lett. 9, the copyright holder. To view a copy of this license, visit http://creativecommons.org/ 20130466 (2013). licenses/by/4.0/. 82. Flanagan, L. B., Comstock, J. P. & Ehleringer, J. R. Comparison of modeled and observed environmental influences on the stable oxygen and hydrogen isotope composition of leaf water in Phaseolus vulgaris L. Plant. Physiol. 96, © The Author(s) 2018 588–596 (1991). 8 NATURE COMMUNICATIONS (2018) 9:1832 DOI: 10.1038/s41467-018-04057-3 www.nature.com/naturecommunications | | |

Journal

Nature CommunicationsSpringer Journals

Published: May 9, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off