Report

Environmental dynamics during the onset of the Middle Stone Age in eastern Africa

See allHide authors and affiliations

Science  06 Apr 2018:
Vol. 360, Issue 6384, pp. 86-90
DOI: 10.1126/science.aao2200

The Middle Stone Age in Africa

The Olorgesailie basin in the southern Kenya rift valley contains sediments dating back to 1.2 million years ago, preserving a long archaeological record of human activity and environmental conditions. Three papers present the oldest East African evidence of the Middle Stone Age (MSA) and elucidate the system of technology and behavior associated with the origin of Homo sapiens. Potts et al. present evidence for the demise of Acheulean technology that preceded the MSA and describe variations in late Acheulean hominin behavior that anticipate MSA characteristics. The transition to the MSA was accompanied by turnover of large mammals and large-scale landscape change. Brooks et al. establish that ∼320,000 to 305,000 years ago, the populations in eastern Africa underwent a technological shift upon procurement of distantly sourced obsidian for toolmaking, indicating the early development of social exchange. Deino et al. provide the chronological underpinning for these discoveries.

Science, this issue p. 86, p. 90, p. 95

Abstract

Development of the African Middle Stone Age (MSA) before 300,000 years ago raises the question of how environmental change influenced the evolution of behaviors characteristic of early Homo sapiens. We used temporally well-constrained sedimentological and paleoenvironmental data to investigate environmental dynamics before and after the appearance of the early MSA in the Olorgesailie basin, Kenya. In contrast to the Acheulean archeological record in the same basin, MSA sites are associated with a markedly different faunal community, more pronounced erosion-deposition cycles, tectonic activity, and enhanced wet-dry variability. Aspects of Acheulean technology in this region imply that, as early as 615,000 years ago, greater stone material selectivity and wider resource procurement coincided with an increased pace of land-lake fluctuation, potentially anticipating the adaptability of MSA hominins.

Environmental drivers of African hominin evolution over the past several million years have been attributed to hydroclimate extremes (arid or moist conditions) and habitat variability (15). Testing for environment-evolution connections requires (i) evidence for geographic and temporal correspondence between novel hominin behaviors and shifts in climate, landscape, and overall biota, and (ii) an understanding of how those behaviors offered selective benefits to hominins as they foraged for resources affected by changing environmental parameters. Linkages between environmental change and evolution are anticipated because climate parameters, such as precipitation amount and variability, influence selection gradients globally (6), and because annual- and decadal-scale variability in climate has a demonstrable influence on survival, reproduction, and demography (7). In human hunter-gatherers, environmentally induced variability and unpredictability in food return rates tend to favor wider mobility, information gathering, and investment in social resource exchange networks—a combination that promotes foraging efficiency, reduces risk, and thus improves fitness (810). Largely unexplored in human origins research are the environmental conditions and adaptive context in which a substantial technological shift and resource exchange networks emerged, as evidenced in the early Middle Stone Age (MSA) (11).

Long sequences of African middle Pleistocene strata that yield multiproxy environmental data associated with a dense archeological archive are rare. The Olorgesailie basin (Fig. 1), containing sedimentary strata dated between 1.2 million years (1.2 Ma) and 499,000 years (499 ka) ago and between 320 and 36 ka ago, preserves an archeological record that spans the Acheulean and MSA. The basin thus presents an opportunity to investigate the environmental context of this behavioral transition and to consider its evolutionary implications in relation to changes in landscape dynamics, climate, flora, and fauna.

Fig. 1 Context of the Acheulean-to-MSA transition in the Olorgesailie basin.

(A) Sequence of major periods of erosion and deposition with inferred paleoenvironments in the Olorgesailie and Oltulelei Formations, associated with Acheulean and MSA archeological sites, respectively. Color code: orange, aggrading sediment; brown, stable land surface; red, burned zone. (B) Insolation during the interval of Acheulean and MSA sites in the Olorgesailie basin. Shaded areas are the three intervals of prolonged (>72 ka), highly variable climate predicted by eccentricity pertaining to the late Acheulean and oldest MSA sites in the basin [climate variability phases H5, H4, and H2 (5)]. (C) Summary of land-lake oscillations and erosional phases associated with Olorgesailie Acheulean and MSA sites. (D) Comparison of earlier Acheulean (purple), later Acheulean (red), and MSA (blue) sites in terms of (top) maximum size of flaked pieces (values for one-tailed t test) and (bottom) percentage of distant stone materials (exotic) relative to local rock sources (exotic stone counts for earlier versus later Acheulean, χ2 = 448.93, P < 0.001; for later Acheulean versus MSA, χ2 = 535.30, P < 0.001; see table S6).

MSA technologies are characterized by small points including weapon armatures, increased reliance on prepared core technologies requiring greater planning depth (12), and procurement of rocks from distant sources, among other traits indicative of complex symbolic, technological, and socioeconomic behaviors (1315). By contrast, in addition to handaxes and other distinctive large tools (16), the 700-ka Acheulean sequence of the Olorgesailie Formation (Fm) exhibits highly localized use of stone material; 98% of rock used in tool manufacture was accessed within 5 km (table S6) (17). The well-documented Acheulean technology of Olorgesailie (29 stratigraphic levels; Fig. 1A) is entirely replaced by MSA technology by ~320 ka ago (18), with evidence for long-distance obsidian exchange networks and previously unknown use of coloring materials (11). This represents a substantial revision in African hominin behavior at or near the time of origin of Homo sapiens (19, 20).

Evidence from Olorgesailie enables direct comparison between the environments associated with Acheulean and MSA technology (21). A ~180-ka erosional hiatus between 499 and ~320 ka ago in the Olorgesailie sequence precludes direct examination of the transitional stages, but the environmental and faunal differences before and after this time interval indicate a marked increase in climatic variability, landscape instability, and aridity. The rich Acheulean sequence of the Olorgesailie Fm is preserved in aggrading lacustrine floodplains and low-energy channels and paleosols representing relatively stable depositional conditions for ~700 ka (i.e., 1.2 Ma to 500 ka ago) (22, 23). However, an increasing pace of environmental change is recorded in the upper Olorgesailie Fm beginning ~800 ka ago, including more frequent land-lake oscillations and fire-reddened siliceous zones that reflect intense moist-arid fluctuation (Fig. 1, A and C) (24). Novel elements of Acheulean technology, some of which resemble innovations seen in the later MSA record, are recorded in this environmentally dynamic upper portion of the Olorgesailie Fm (text S2 and S3). Reliance on smaller artifacts is apparent, and the proportion of artifacts made from nonlocal stone (e.g., chert, obsidian, quartzite) is 3 to 6 times that observed in older Acheulean sites, and within the range found at the later MSA sites (Fig. 1D and tables S5 and S6).

Sedimentation of the Olorgesailie Fm came to an end ~499 ka ago, the age of the last recorded Acheulean technology in the basin (Fig. 1A and fig. S7). Approximately 180 ka of landscape dissection caused by major structural changes in the basin was followed by deposition of the Oltulelei Fm, signifying a basin-wide shift in depositional regime (25). In its oldest unit (Olkesiteti Member), the early MSA is recorded in strata dated ~320 to 295 ka ago (18) in a sequence of diverse superposed lithologies indicative of rapidly shifting fluvial channel, floodplain, and spring deposition. The first records of the Olorgesailie MSA are therefore associated with dynamic paleolandscapes in which linked but semi-independent basins were subjected to repeated erosion and deposition across an area of ~60 km2 north and northwest of Mt. Olorgesailie. Multiple phases of channel cutting and filling within the lower Olkesiteti Member of the Oltulelei Fm further indicate repeated shifts in local base level, likely related to a combination of faulting and arid-moist climate variability (25).

MSA horizons at BOK (sites BOK-1E/BOK-3, BOK-2, and BOK-4, from lowest to highest), dated ~320 to 295 ka ago (18), occur in a succession of sedimentary fills of cross-cutting channels incised 2 to 4 m deep, indicating repeated cycles of erosion, filling, and development of floodplains (Fig. 2A). Silicified plant stems, roots, and siliceous tuffs are associated with the largest archeological assemblage at site BOK-2 (11) and are interpreted as evidence for localized springs and associated moist habitats during the artifact-rich channel-filling phase (25). The BOK sites are located within tens of meters of a confluence where the main drainage flowed toward a gap in a nearby lava ridge. This confluence and nearby springs were a likely water source for hominins and other organisms. Site GOK-1 occurs below an erosional unconformity on top of a 5-m-thick red soil, and the eroded top of this soil is at least 1 m below a tuff dated ~277 ka ago (18). Thus, GOK-1 is possibly as old as or older than the BOK sites. The red paleosol interfingers with tufa deposits, indicating associated spring activity, shallow ponds, and wetlands; subsequent MSA archeological levels at GOK occur in overlying floodplain deposits and paleosols (11, 25).

Fig. 2 Landscape and environmental setting of the >300,0000-year-old MSA sites at Olorgesailie.

(A) Simplified stratigraphic panel showing superposed channel deposits (Olkesiteti Member of the Oltulelei Fm: blue-green to blue-gray) that preserve the archeological record in Locality B (BOK archeological sites), viewed toward approximately west-southwest (toward the drainage outlet). White, Olorgesailie Fm; overlying orange and gray, channels of the upper members of the Oltulelei Fm. (B) Carbon isotope values of pedogenic carbonates from Olorgesailie paleosols (this study) and from elsewhere in eastern Africa [compiled in (37)] over the past 800 ka; vegetation cover categories are from (38). (C) Percentage of phytolith morphotypes of woody and herbaceous taxa, grass subfamilies, and palm and sedge (Cyperaceae) taxa associated with the MSA site BOK-2. (D) Chord distance, a measure of faunal dissimilarity for taxonomic abundance data (33), between the BOK fauna and census counts from 33 modern African wildlife reserves [data from (34)]. Modern faunas are arranged from lowest chord distance (top: most similar to BOK fauna) to highest chord distance (bottom: least similar to BOK fauna). Color coding of modern wildlife reserves: white, open and arid/semi-arid; green, woodland; purple, closed/humid (3436). The BOK fauna is most similar to those from the open and arid settings of Makgadikgadi (Botswana) and Etosha (Namibia).

Coinciding with the large-scale change in depositional regime, the mammalian fauna associated with the Olorgesailie MSA indicates substantial turnover [Jaccard distance = 0.85; i.e., only seven taxa (15%, n = 46) shared between the Acheulean and MSA assemblages] (Fig. 3). This impressive biotic shift included local extinction of previously dominant large-bodied specialized grazing lineages (e.g., Paleoloxodon recki, Equus oldowayensis, Theropithecus oswaldi) prevalent in the Olorgesailie Acheulean sequence, their replacement by related taxa of smaller body size (Fig. 3) (26), and an example (springbok, Antidorcas marsupialis, known from Florisbad, South Africa ~100 to 400 ka ago) of apparent range extension covering as much as 2800 km (table S1 and fig. S5). The springbok’s temporary presence in eastern Africa supports phylogeographic evidence that this region’s dynamic environmental history contributed to elevated turnover, which included local extinctions and recolonization from southern African sources (27). This reorganization of the large mammalian community, which was more extensive than other examples of Pleistocene faunal turnover in eastern Africa [e.g., (28)], is first recorded in southern Kenya between ~330 and 320 ka ago (26, 29) (Fig. 3), roughly contemporaneous with the onset of the MSA sequence and consistent with the last appearances of extinct middle Pleistocene large-bodied mammals in eastern Africa (29).

Fig. 3 Acheulean-associated and MSA-associated mammalian taxa of the Olorgesailie basin.

The hyper-grazer alcelaphin Damaliscus hypsodon became prominent in the southern Kenya rift by ~330 ka ago (29) and is also found in fauna dated >305 ka ago in Olorgesailie MSA sites. Equus, including a previously unidentified large species that we attribute to Equus aff. E. capensis, and the bovids Oryx and D. hypsodon indicate arid, open terrain suitable for grazing; Tragelaphus, Diceros, and Madoqua indicate the presence of bush/shrubland; and Hippopotamus and Crocodylus attest to perennial bodies of water (text S1, tables S1 and S2, and figs. S3 and S4). Chord distance measures of dissimilarity to mammalian faunas from modern African wildlife reserves show that the BOK fauna most closely matches modern faunas from open, arid settings (Fig. 2D and fig. S2). BOK small mammal taxa (e.g., Pedetes) also imply semi-arid, open grassland to bushland habitats with little woody cover, and the presence of ostrich and francolin at the BOK sites indicates an open, dry, and grassy component of the Olorgesailie paleolandscape (text S1).

Carbon isotope data from the limited number of pedogenic carbonate horizons preserved at Olorgesailie point to the development of grasslands spanning the transition between the latest Acheulean and the earliest MSA. In accord with paleosol data from elsewhere in eastern Africa, Olorgesailie δ13C values show a steady increase in grassy environments during the past 800 ka (Fig. 2B and text S1). The pedogenic carbonate isotope data are consistent with the presence of arid grassland-adapted fauna but do not show a sharp shift in vegetation commensurate with the faunal turnover. We consider it likely that the faunal change was a response to a combination of aridity and food/water resource variability that resulted, for example, from shifting rainfall seasonality and tectonically caused increases in topographic diversity (25)—factors that may have had little influence on the carbon isotopic composition of vegetation in the local soils. A phytolith assemblage obtained from the BOK archeological sites, however, indicates a local habitat that was more moist and vegetated. The BOK phytolith sample (Fig. 2C and text S1) contains short-grass (Chloridoideae) forms, consistent with the faunal and carbon isotope data, but also a high proportion of woody, herbaceous, and moisture-associated tall grass (Panicoideae) and sedge morphotypes, indicative of the vegetation at or adjacent to the focus of hominin activity. Together, the environmental and geomorphological evidence implies that the MSA sites at BOK were situated in a predominantly grassland habitat near a reliable water source that also supported woody vegetation.

The Olorgesailie MSA coincided with a predicted interval of prolonged wet-dry climate oscillation for eastern Africa (climate variability stage H2; Fig. 1B) (5). Basin lithofacies document an increased tempo of erosional phases and volcanic input to the Olorgesailie region (Fig. 1, A and C) as tectonic activity created semi-independent depositional zones during the past ~320 ka of Olorgesailie basin history (25). Tectonics also contributed to the topographic heterogeneity of the southern Kenya rift (30, 31). Because spatial and temporal habitat variability strongly influences resource patchiness and predictability, rapidly shifting environments at the outset of the Oltulelei Fm imply strong selective pressures on resource acquisition by hominins. Both the faunal turnover and MSA replacement of the Acheulean took place in this dynamic setting. Evidence of grassland expansion associated with the Olorgesailie MSA further implies that MSA hominins encountered periods of intensified resource uncertainty. Foraging unpredictability, with increased potential for resource scarcity, describes the conditions in which human hunter-gatherers broaden the spatial scale of the landscape they encounter through a combination of wider mobility, information and resource sharing, and the maintenance of resource exchange networks (8, 10, 32). These responses also characterize the Olorgesailie MSA, in contrast to the behavioral record associated with the older Acheulean of the region (11). We thus hypothesize that the emergence of the MSA and its wholesale replacement of the Acheulean in the Kenya rift by ~320 ka ago represents an evolutionary response to resource landscapes that were less predictable in time (as a result of amplified climate variability throughout eastern Africa) and were also more heterogeneous in space (as a result of local tectonic activity). Although we are not claiming that MSA technology originated in the Olorgesailie basin, the changing conditions and depositional sequence there document the environmental correlates associated with its appearance in southern Kenya.

Although immediate MSA precursors in the Olorgesailie basin cannot be investigated because of the erosional gap from 499 to 320 ka ago, recovery of Acheulean materials from intervals of enhanced climate oscillation recorded in Members 11, 12, and 14 of the upper Olorgesailie Fm (Fig. 1A: sites B11-500, AW12-10, and D14-10) indicate that Acheulean toolmakers between 615 and 499 ka ago (18) produced smaller cores and flakes, and also increased their access to distant stone sources and were more selective of stone material than earlier in the Acheulean (Fig. 1D and text S3 and S4). These aspects of foraging technology suggest that, at least in periods of strong environmental variability, Acheulean hominins broadened their mobility and range of resource acquisition within the limits of a technology typified by the manufacture of large cutting tools. These incipient similarities to the MSA point to variation within the Acheulean that, by 615 ka ago, appears to have laid a foundation for the later evolution of adaptable behaviors distinctive of H. sapiens.

Supplementary Materials

www.sciencemag.org/content/360/6384/86/suppl/DC1

Materials and Methods

Supplementary Text S1 to S4

Figs. S1 to S10

Tables S1 to S6

References (3969)

References and Notes

  1. See supplementary materials.
Acknowledgments: We thank the National Museums of Kenya and M. Kibunjia, I. O. Farah, F. K. Manthi, E. N. Mbua, M. Muungu, E. Ndiema, P. Kiura, C. M. Kilonzi, S. M. Musyoka, J. N. Mativo, A. M. Kioko, B. L. Pobiner, J. Moerman, and K. Meshida for support. J. M. Nume, B. Kanunga, and C. M. Kilonzi managed the excavation teams. We acknowledge Kenya Government permission granted by the Ministry of Sports, Culture and the Arts, and by NACOSTI permit P/14/7709/683. Funding: Supported by the Peter Buck Fund for Human Origins Research, Smithsonian Institution (R.P.), NSF HOMINID Program grant BCS-0218511 (R.P.), and NSF Archaeometry grant EAR-1322017 (A.L.D.). Author contributions: R.P. was principal author and directed the project; A.K.B. and A.L.D. led the geological research; A.S.B., J.E.Y., R.P., and C.A.T. led the excavations and recovery of archeological and fossil materials; R.P., J.T.F., J.B.C., C.M.H., and H.J.M.M. conducted faunal analysis; R.P. conducted the lithic analysis, reviewed by A.S.B. and C.A.T.; R.K. conducted phytolith sampling and analysis; N.E.L. carried out carbonate sampling and stable isotopic analysis; J.B.C. composed the figures; E.G.V. and J.B.C. drew and photographed the stone artifacts; R.B.O. and R.W.R. carried out analysis of spring deposits; and all authors contributed to the writing and review of the manuscript. Competing interests: The authors have no competing interests. J.E.Y. contributed to this article in his personal capacity; any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the NSF. Data and materials availability: All data are available in the paper and supplementary materials, and all archeological and paleontological collections and field records are archived in the Department of Earth Sciences, National Museums of Kenya.
View Abstract

Stay Connected to Science

Navigate This Article