Long-distance stone transport and pigment use in the earliest Middle Stone Age

See allHide authors and affiliations

Science  06 Apr 2018:
Vol. 360, Issue 6384, pp. 90-94
DOI: 10.1126/science.aao2646

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


Previous research suggests that the complex symbolic, technological, and socioeconomic behaviors that typify Homo sapiens had roots in the middle Pleistocene <200,000 years ago, but data bearing on human behavioral origins are limited. We present a series of excavated Middle Stone Age sites from the Olorgesailie basin, southern Kenya, dating from ≥295,000 to ~320,000 years ago by argon-40/argon-39 and uranium-series methods. Hominins at these sites made prepared cores and points, exploited iron-rich rocks to obtain red pigment, and procured stone tool materials from ≥25- to 50-kilometer distances. Associated fauna suggests a broad resource strategy that included large and small prey. These practices imply notable changes in how individuals and groups related to the landscape and to one another and provide documentation relevant to human social and cognitive evolution.

The Middle Stone Age (MSA) comprises a diverse group of African industries characterized by the absence or rarity of large cutting tools (LCTs, such as Acheulean handaxes and cleavers), the presence of prepared core (Levallois) technologies, and regional and/or temporal variation in weapon armatures (13). Later MSA industries contain varying frequencies of small retouched bladelets and geometric artifacts (4), small unifacially and bifacially worked points, and larger lanceolate and bone points (5, 6). African MSA sites are associated with the oldest known fossils attributed to Homo sapiens, which date between ~195 and ~350 thousand years (ka) ago (79). After 130 ka ago, late Pleistocene MSA sites, mostly from the north and south temperate zone extremities of the continent, document some of the oldest beads (10, 11), complex geometric designs (12, 13), lithic projectile armatures interpreted as arrow or dart tips (14, 15), paint production (16), heat treatment of lithics (17), and expanded resource use, including fishing and shellfish collecting (18, 19).

Although these late Pleistocene behaviors are thought to reflect advanced cognitive abilities for learning, imitation, working memory, planning, and recursive technological innovation, their middle Pleistocene or earlier antecedents have remained obscure. Evolutionary explanations for the middle Pleistocene enlargement of relative brain size, largely overlapping the modern H. sapiens range, include a shift to larger social groups or networks (20); greater foraging, technological, and social adaptability in the face of more rapid and unpredictable climate fluctuations (21); and the development of symbolic expression (22).

Middle Pleistocene evidence from >200 ka ago for the emergence of these new cognitive and social behaviors is limited to a small number of isolated sites and site complexes, mostly in tropical and subtropical Africa and also in the northwest (8, 2326). The most extensive record in both chronological coverage and individual excavations is from the Kapthurin Formation in Kenya (2730).

Olorgesailie is a ~65-km2 paleolake basin in the southern Kenya rift where more than 75 years of research have documented Acheulean sites in the Olorgesailie Formation, dating from 1.2 million years ago to 499 ka ago (31, 32). From 2001 to 2011, the Smithsonian team investigated sites in the younger Oltulelei Formation (33, 34). Seventeen excavations were conducted in Localities B and G in the Oltulelei Formation, including 11 in the Olkesiteti Member (Fig. 1). The five oldest sites are described here (35). The oldest (~305 to 320 ka old) and the youngest (between ~295 and ~298 ka old) (32) yielded some of the earliest evidence for MSA lithic industries lacking LCTs (Fig. 2 and fig. S5).

Fig. 1 Composite geological sections of Localities G and B early MSA sites, showing archaeological sites, artifact levels, and dated tuffs.

Artifact levels are indicated with red dashes. Bottom left, location of the Olorgesailie Basin; center, relationship of Localities G and B to Mt. Olorgesailie; right, Locality B site locations.

Fig. 2 MSA artifacts from BOK sites.

(A) Levallois point (Olorgesailie basalt). (B) Levallois point (obsidian). (C to F) Small obsidian MSA points with flat invasive retouch (acute-angled bases in side view). (G) Single-platform blade core. (H and I) Obsidian Levallois cores. Point (F) refits onto core (I). Artifact (D) and (F) have bases thinned by retouch. Artifacts (A) and (E) are from BOK-4; the others are from BOK-2.

These MSA sites occur in a topographically complex paleolandscape marked by repeated channel erosion and filling, periodic influxes of volcaniclastics, tectonic shifts, and spring-deposited tufas (33, 34). The oldest MSA sites in Locality B (BOK-1E and BOK-3; Fig. 1) directly overlie channel-deposited gravels representing initial aggradation after the long (~180 ka) erosional hiatus between the Olorgesailie and Oltulelei Formations. These gravels postdate the ~499-ka capping age of the former and predate the ~305-ka-old tephras overlying the BOK-1E occupations (32). Site BOK-2 is slightly higher stratigraphically and underlies the ~305-ka-old tephras and a ~302-ka-old tuff, whereas BOK-4 is, as noted above, slightly younger (between ~295 and ~298 ka old) (Fig. 1).

In eastern Locality G, the Olorgesailie Formation is absent in most outcrops. We infer that the well-developed red paleosol containing the GOK-1 lithic industry formed after Olorgesailie Formation deposition ceased <500 ka ago. Archaeological site formation in the GOK-1 red soil was followed by a widespread erosional unconformity. A U-series date of 277 ± 1.8 ka old in deposits overlying the unconformity provides an upper limit for the age of GOK-1. Subsequent volcaniclastic units above the unconformity at GOK-1 are dated at ~220 ka ago (32). The lithic industry shares comparable percentages of diagnostic MSA elements (table S3) with the two oldest sites in Locality B. No LCTs were found in the extensive GOK-1 excavations. Together, the stratigraphic context and flaked stone artifacts suggest that the age of GOK-1 is comparable to or possibly older than that of the early Locality B MSA sites (Fig. 1).

All early Locality B sites are in fine-grained silty or sandy sediments that represent low-energy sedimentation in channel-fill settings above the basal gravels. The absence of preferential orientation of long axes and abundance of artifacts <2 cm in maximum dimension (table S2), together with large cores, grindstones, and hammerstones (table S3), indicate that the artifact assemblages were deposited by hominins in or near inactive channels, followed by only minor postdepositional sorting or winnowing. Sets of refitted obsidian artifacts at BOK-2 and BOK-4 (figs. S3 and S4 and table S2) further indicate the integrity of the artifact assemblages.

Most artifacts at all but one of the Olkesiteti Member sites were made on local volcanic rocks with 1 to 8% nonlocal obsidian and chert (table S4). The small amount (0 to 2%) of green, brown, and white chert has not been sourced but is not available locally. Even within the lavas, a preference for fine-grained rocks is evidenced by the predominance of Mt. Olorgesailie basalt, which lacks visible phenocrysts and was available at outcrops within 2 km.

At BOK-2, however, 42% of the plotted artifacts (table S5) are obsidian, which has no known local usable outcrop or conglomerate source. Compositional analysis of 688 obsidian artifacts from BOK-2 and BOK-4 was carried out by several techniques, including portable x-ray fluorescence (pXRF) and neutron activation analysis (NAA), to determine probable sources, based on a large database of known provenience. About 78% of the Olorgesailie samples are attributed to seven source groups located in direct lines from 25 to 50 km from the BOK sites in five different directions. A small number of samples were from more distant sources (Fig. 3). Given the rugged terrain of the southern Kenya rift, the walking distance would have been considerably longer than the straight-line distance. For four occupation horizons at BOK-2, the source diversity and distance are smallest in the oldest level and largest in the youngest (Fig. 3, bottom panels).

Fig. 3 Location of major obsidian sources.

Sources matched chemically to Olorgesailie MSA artifacts by pXRF and NAA at the University of Missouri Research Reactor (MURR) and/or by pXRF at the National Museums of Kenya (NMK). The bottom panels are chemical plots that show the use of multiple sources in basal and upper levels at BOK-2 and BOK-4. Source group 2 is an unknown source. ppm, parts per million.

Large obsidian flakes and cores weighing 60 to 90 g, refit sequences (figs. S3 and S4), low cortex ratios, and the presence of >46,000 obsidian pieces <2 cm in size at BOK-2 and 226 pieces at BOK-4 (tables S2 and S4) indicate that obsidian was brought in as raw material and worked on-site rather than imported as finished artifacts. Average cortex ratios on whole flakes are low (0 to 25%) for both local and exotic common materials in the Locality B sites (table S4). Long-distance obsidian transport at Olorgesailie precedes previously documented occurrences (30) by ~80 to 100 ka.

The distinct artifact levels at BOK-2 represent multiple reoccupations, also evident at BOK-1E and GOK-1, indicating repeated visits to focal landscape points in the basin, likely related to resource availability.

Analysis of lithic technology (Fig. 3, fig. S5, and table S3) at all five sites supports attribution to the MSA rather than to the Acheulean. First, no large cutting tools were recovered in the excavations, although a few surface finds of core axes and picks with trihedral points and unfinished bases were recovered on a red-soil surface below GOK-1; such artifacts are typical of early post-Acheulean industries of middle Pleistocene age in eastern and central Africa (36) (fig. S1). Second, diverse technological sequences recovered in situ included prepared cores, especially Levallois, or asymmetrical cores (Fig. 2, H and I, and fig. S5); other core strategies included the production of blade and bladelet cores (Fig. 2G), bipolar cores on small chert and obsidian nodules, and discoidal and multiplatform cores. Third, the desired end products were predominantly small- to medium-sized flakes (≤5-cm average length), especially flakes that were relatively thin (table S4). Fourth, flat invasive retouch was used to shape some of the points and the more rounded “ovates.”

Classic MSA end products at the Locality B sites reflect innovation and standardization that were not present earlier. These include pointed forms (Fig. 2 and table S3) both unretouched and retouched at BOK-2, BOK-4, and BOK-1E. Retouched points were preferentially made on obsidian at BOK-2 and on chert at BOK-4. The bases of some points have been thinned or modified in a manner typical of hafting (Fig. 2, D and F). Perforators, side and end scrapers, notched and denticulate pieces, and other retouched forms are present at all three sites. At BOK-2, a series of short end scrapers on obsidian flakes occur, as well as very small scrapers on flakes (fig. S2). In summary, the industries of these early sites fit within the scope of African industries grouped as MSA in the diversity of artifact forms and technological approaches to tool manufacture.

More than 2000 faunal remains of 23 larger (>2 kg) mammalian taxa, micromammals, reptiles, birds, and fish were recovered from the Locality B sites in association with the lithics; these remains carry implications for reconstructing both human behavior and paleoenvironments and are presented in detail in (34). Surface markings document human use. Carnivore remains are rare and do not suggest a den concentration, and dens or burrows were not present in the excavated sediments. The most common taxa include equids, suids, eland, kudu, an extinct alcelaphin (Damaliscus hypsodon), springbok, bat-eared fox, springhare, and root rat. The presence of many relatively smaller taxa suggests direct predation by humans rather than scavenging, as small taxa remains are unlikely to survive initial consumption by primary predators. Comparable recent African forager reliance on small mammals, including small carnivores, is documented in table S15.

Given the age of the sites, of particular interest are mineral pigments from two of the five sites: BOK-1E and GOK-1. Within the upper artifact horizon at BOK-1E were 86 rounded concretionary lumps of soft dark mineral. One nodule analyzed by scanning electron microscopy energy-dispersive x-ray spectroscopy consists mainly of calcite with fluorite inclusions and veins of manganese oxide, a dark brown–to-black potential pigment, whereas the nodule’s surface layer contains diatoms and possibly cyanobacteria, indicating formation in water (Fig. 4C) (supplementary text S6). Test squares extending both north and south from the excavation indicated that, although the sediment enclosing the artifacts continues in both directions, the dark mineral lumps occur only in association with artifacts. This association, together with the lack of evidence for standing water around the artifact horizon, suggests that these specimens were collected and brought to the site by hominins, although they were too friable to preserve grinding striations (supplementary text S6).

Fig. 4 MSA pigments from Olorgesailie.

(A) From GOK-1, large ochre block (specimen 6) with two opposing holes (white arrows) and abraded red area. (A1) Smaller lower hole with chop marks made by sharp, straight-edged tool (white arrows), a gouge mark (blue arrow) in lower left corner, and vertical walls left and right. (A2) Top view of the large ochre block. (A3) Magnified view of area enclosed by dashed rectangle in (A2). Larger upper hole with chop marks (white arrows) and gouge mark (lower left). (A4) Magnified view of abraded red area showing striations. (B) From GOK-1, smaller block with red streaks and notch in upper left (specimen 7). (B1) Magnified view of notch showing chop mark (white arrow). (C) Surface section of a BOK-1E lump, showing diatoms (white arrows), possible cyanobacterium (yellow arrow), and manganese infillings. (D and E) Pigment examples from BOK-1E.

In the GOK-1 excavation, two rocks that have bright red streaks and patches were recovered in the red soil, one (specimen 7) in the uppermost unit and one (specimen 6) in the middle unit. Specimen 6 has two opposing holes, initially thought to be of geological origin. Microscopic examination revealed that the red streaks on this stone are grinding striations (Fig. 4, panel A4), and geochemical analysis of the rock by laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) (fig. S8, A and B) indicated that it is mainly composed of iron mineral. Rather than a grindstone that was used to produce pigment, this is interpreted as a lump of ochre pigment that had been ground. Further visual examination with a free-angle zoom microscope documented chop marks in the two opposing holes, one large and one small, as well as in the notches of the second rock (specimen 7), indicating that the holes and notches are anthropogenic (Fig. 4). Geochemical comparison of the GOK-1 red rocks with potential local sources indicates that the rocks are exotic to the immediate region and not derived from the nearest potential source located on Mt. Olorgesailie (fig. S9). Although earlier ochre has been occasionally reported, evidence of anthropogenic transport and utilization is rarely documented [but see (1, 37)]. The GOK-1 rocks are thus among the oldest-known clearly worked pigments and the oldest evidence of a possible attempted perforation. Although ochre can have utilitarian functions, experimental data have shown that these functions can also be satisfied by less-brightly colored minerals (38), suggesting that transport of exotic bright red and black rocks may have been valued for their bright color and used as symbolic communicators of identity or status.

The evidence from the five oldest MSA sites at Olorgesailie has multiple implications for understanding human behavioral evolution. First, this evidence indicates that distinctive technological features of the African MSA reflecting innovation, standardization, and new cognitive abilities had already been developed in eastern Africa before 300 ka ago. Second, the repeated occurrence of exotic raw materials from multiple distant sources as cores, knapping products, and finished artifacts and the quantities involved suggest a new behavior in the human repertoire: the formation of networks of exchange or procurement over a substantial area. At least two sources were used in the oldest level at BOK-2, and four to six sources were used in the later horizons of this site and at BOK-4 (Fig. 3). The formation of extended networks and social connections is an important universal aspect of ethnographically documented hunter-gatherer adaptations to unpredictable environments and is effective strategy for risk mitigation (39, 40). Expanded social networks are one possible explanation for the middle Pleistocene increase in relative brain size (20), although evidence supporting this social expansion >300 ka ago was not previously available. Along with improved lithic technology, particularly stone-tipped weapons, and a more diverse economic strategy suggested by the faunal remains, larger-scale social relationships would have provided a buffer against increased climate variation and resource unpredictability (34, 39, 40).

In a Kalahari environment comparable to that implied by the BOK sites, fauna, isotopes, and other indicators (34), modern hunter-gatherer families maintain contacts with exchange partners up to a 100-km distance but generally range over only a 20-km radius annually (39). Thus, the distances implied by the range of obsidian procurements somewhat exceeds the likely home range in the Olorgesailie environment of a small family-based band but are consistent with exchange-network sizes of modern people in a similar environment. The absence of this pattern from the underlying Acheulean levels (34) indicates that this major transition in human social and cognitive behavior occurred before 300 ka ago but postdated 500 ka ago. Finally, the apparent association of early pigment use and expanded social networks implied by the repetitive use of exotic raw materials from multiple sources suggests that the pigment itself may also have been of social and symbolic importance.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S10

Tables S1 to S15

References (4176)

References and Notes

  1. Materials and methods are available as supplementary materials.
Acknowledgments: We are grateful to the National Museums of Kenya (NMK) and the Ministry of Sports, Culture, and the Arts for support of this project. Special thanks to M. Kibunjia, I. O. Farah, F. K. Manthi, E. N. Mbua, M. Muungu, J. Mwangi, C. Ogola, E. Ndiema, P. Kiura, C. M. Kilonzi, S. M. Musyoka, J. N. Mativo, A. M. Kioko, C. Kirwa, J. M. Munyiri, and F. Bante for their contributions at the NMK. J. M. Nume, B. Kanunga, and C. M. Kilonzi were in charge of the excavation teams. We thank S. Nchilalo, E. ole Keri, and S. ole Keri for guidance to obsidian and ochre sources, D. Wawrzyniak for assistance with obsidian compositional tables, J. Watson and N. Little for initial analysis of the pigments, and L. Dussubieux for laboratory assistance and use of facilities at The Field Museum of Natural History (Chicago) for ochre compositional studies. The following individuals also contributed to field and laboratory research: W. G. Sharp, J. T. Faith, C. M. Haradon, C. A. Tryon, E. M. Williams, K. Ranhorn, N. E. Levin, A. G. Henry, F. Gomez, A. Bauernfeind, R. Biermann, R. Kinyajui, R. B. Owen, and R. W. Renaut. Funding: Kenya field and Kenya and U.S. laboratory work were supported by the Peter Buck Fund for Human Origins research (Smithsonian Institution) (R.P.) and by NSF Human Origins: Moving in New Directions (HOMINID) award BCS-0218511 (R.P.). Field and laboratory work by A.M.Z. on obsidian and ochre sourcing was supported by NSF Graduate Research Fellowship 2011116368 (A.M.Z.), NSF Doctoral Dissertation Research Improvement Grant BCS-1240694 (A.S.B. and A.M.Z.), NSF Integrative Graduate Education and Research Traineeship (IGERT) award DGE-0801634 to George Washington University, and Wenner-Gren Foundation Dissertation Fieldwork Grant 8623 (A.M.Z). Laboratory analyses on dating samples were funded by NSF grant EAR-1322017 to A.L.D. Field and laboratory research on Kenya obsidian sourcing was funded by NSF awards to the University of Missouri Research Reactor Center (BCS-0802757) and to J.R.F. and S.H.A. for the Kenya obsidian project (BCS-0814304). Laboratory work on the ochre by F.d’E. was funded by the Research Council of Norway (SFF Centre SapienCE), project number 262618, and the Université de Bordeaux LaScArBx research program (ANR-10-LABX-52). J.E.Y. was supported by the NSF and by the Smithsonian Institution Olorgesailie project. Author contributions: A.S.B. and J.E.Y. led the survey, excavation, and analysis of the MSA sites and preparation of relevant tables and figures; R.P. directed the overall Olorgesailie Project and analyzed the fauna and nonobsidian volcanic materials; A.K.B. and A.L.D. conducted the geological study and sampling of the Oltulelei Formation; A.L.D. conducted 40Ar/39Ar dating; J.E.Y. identified artifact raw materials and carried out pXRF analysis at the NMK; D.E.L. conducted the refitting studies; S.H.A. led the Kenyan field study of obsidian sources and participated in the source chemistry analysis; J.R.F., S.H.A., and K.F. conducted geochemical studies of sources and artifacts; F.d’E. conducted the first microscopy that demonstrated grinding striations on the ochre; A.M.Z. assisted with the excavations and led the study of ochre sources, including geochemical compositional study of the ochre pieces by LA-ICPMS at the Field Museum of Natural History (Chicago); S.W. conducted microscopy on the anthropogenic modifications of the ochre pieces; J.P. analyzed dark pigment in polished section; E.G.V. produced the lithic drawings; and J.B.C. conducted the microfauna study and prepared the figures, including artifact photography and computer assembly. A.S.B. was the principal author, with major sections contributed by A.K.B., S.H.A., A.M.Z., and K.F., and R.P. edited the paper. Competing interests: The authors declare 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 archaeological, paleontological, and obsidian-source reference collections and field records are archived in the NMK, Department of Earth Sciences.

Stay Connected to Science

Navigate This Article