Report

Chronology of the Acheulean to Middle Stone Age transition in eastern Africa

See allHide authors and affiliations

Science  06 Apr 2018:
Vol. 360, Issue 6384, pp. 95-98
DOI: 10.1126/science.aao2216

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

The origin of the Middle Stone Age (MSA) marks the transition from a highly persistent mode of stone toolmaking, the Acheulean, to a period of increasing technological innovation and cultural indicators associated with the evolution of Homo sapiens. We used argon-40/argon-39 and uranium-series dating to calibrate the chronology of Acheulean and early MSA artifact–rich sedimentary deposits in the Olorgesailie basin, southern Kenya rift. We determined the age of late Acheulean tool assemblages from 615,000 to 499,000 years ago, after which a large technological and faunal transition occurred, with a definitive MSA lacking Acheulean elements beginning most likely by ~320,000 years ago, but at least by 305,000 years ago. These results establish the oldest repository of MSA artifacts in eastern Africa.

The earliest development of the Middle Stone Age (MSA) has been obscure because stratigraphic records that span from the Acheulean, which is characterized by distinctive large cutting tools (LCTs), to the early MSA are rare and poorly constrained in time. Here we present precise geochronology based on 40Ar/39Ar and U-series dating of middle Pleistocene archaeological sites in the Olorgesailie basin, southern Kenya rift (detailed site stratigraphy is shown in Fig. 1; composite stratigraphy is shown in fig. S8), that document the stone technology, fauna, and environments of the Acheulean-to-MSA sequence (1, 2).

Fig. 1 Locations, stratigraphic relationships, and stratigraphic placement of localities, archaeological sites, and dated tuffaceous beds.

(A) Olorgesailie Formation, Member 11 (M11), in Localities A and B. (B) Oltulelei Formation in Locality B and BOK sites. Mb., member. (C) Oltulelei Formation in Locality G and GOK sites. Inset maps show the location of Olorgesailie (top) and stratigraphic locations for (A), (B), and (C). Positions and ages of dated samples are shown in italics. Stratigraphic relationships are based on field mapping of channel margins and lithologies (13). Dashed lines indicate inferred correlations between beds, whereas solid lines indicate correlations that are more traceable and secure. Colored lines in (B) correspond to geological boundaries in Locality B: black between the Olorgesailie and Oltulelei Formations, green and orange between three units of the Olkesiteti Member of the Oltulelei Formation, brown at the base of the Oltepesi Member, and red at the base of the Tinga Member. Numbers above columns are latitude-longitude positions and stratigraphic section designations. Lateral section positions are not to scale.

As originally defined (3), the MSA lacked LCTs and bladelet technologies, although more recent work recognizes the presence of LCTs in some early MSA assemblages and of bladelets especially after 70,000 years (70 ka) ago. For much of the 20th century, the MSA was considered irrelevant to human evolution and dispersal, largely because the chronometric techniques and climate correlations used in African contexts placed it at the end of the Pleistocene, contemporary with the richer record of European cave art and elaborate lithic tool kits. Human behavioral evolution during the middle Pleistocene (780 to 130 ka ago) and into the late Pleistocene before 30 ka ago was thus poorly understood (4).

Based on newer techniques, the well-constrained oldest MSA sites in eastern and eastern-central Africa now have estimated ages of ≥286 to ≤182 ka (Table 1), whereas the earliest fossils attributable to Homo sapiens sensu stricto in eastern Africa have ages of ~160 to 200 ka [Table 1 and table S4; all K-Ar and 40Ar/39Ar ages are recalculated for consistency with K decay constants and revised standard ages (5)]. Although recent investigations in northern Africa suggest the coeval evolution of H. sapiens sensu lato and the MSA in the middle Pleistocene (6, 7), the earliest MSA predates current evidence for the first appearance of H. sapiens sensu stricto in eastern Africa and likely constitutes the context in which the anatomical and behavioral characteristics of our species evolved.

Table 1 Eastern African MSA and Acheulean middle Pleistocene sites with published K-Ar or 40Ar/39Ar age estimates from ~600 to ~180 ka, excluding new Olorgesailie Formation (Member 11) analyses reported in this paper.

Ages are recalculated using revised K decay constants (17) and astronomically calibrated standard ages [Fish Canyon Tuff sanidine, 28.201 Ma (18); Alder Creek Rhyolie sanidine, 1.1848 Ma (19)].

View this table:

A summary of dates (Table 1) shows that Acheulean and transitional industries incorporating large tools (Sangoan and Fauresmith) disappeared in Kenya before 285 ka ago or earlier but persisted in areas of Ethiopia, Sudan, and southern Africa into the end of the middle Pleistocene, after the first appearance of H. sapiens sensu stricto. Although dates have been published for 12 middle Pleistocene South African sites, only Florisbad, Kathu Pan, and possibly Sterkfontein and Wonderwerk—all in northern South Africa—have dated MSA horizons older than marine isotope stage 6 (MIS6; ~191 ka ago) (8), and six lack LCTs. Coastal MSA cave horizons older than MIS6 appear in northwestern Africa but are absent from the rich later record of coastal South African caves.

After our team’s investigations of the Acheulean sites, fossil remains, and geology of the Olorgesailie Formation, dated at ~1.2 to 0.5 million years (Ma) old (912), research since 2002 has extended into the Oltulelei Formation (13), which has rich accumulations of MSA artifacts. In this work, we set out the chronostratigraphic controls for the end of the Acheulean and its replacement by MSA sites in the Olorgesailie basin, which contains the longest stratigraphic sequence of Acheulean and overlying MSA sites in eastern Africa. Other recent and concurrent manuscripts establish the stratigraphic and regional geological context of strata younger than 500 ka at Olorgesailie (13), characterize the MSA sites (1), and examine the paleoenvironmental setting, faunal turnover, and landscape dynamics associated with the MSA in the Olorgesailie basin (2).

The MSA sites of Olorgesailie are exposed within a ~65-km2 early to late Pleistocene sedimentary basin in the southern Kenya rift, north of the deeply eroded late Pliocene–early Pleistocene central volcano of Mt. Olorgesailie (14) (Fig. 1, inset). The Pleistocene deposits lie on older lavas dated at 2.66 ± 0.06 Ma old (all chronologic uncertainties are reported at the 2σ level) (9) and the regionally extensive Magadi Trachyte (1.4 to 0.8 ka old) (15). Normal fault movement contemporaneous with sedimentation has resulted in the separation of the Olorgesailie basin into three sub-basins with shifting histories of terrigenous and lacustrine sedimentary environments (13).

The basin strata consist of the older, areally extensive and horizontally stratified Olorgesailie Formation, overlain unconformably by channel fills and sheet deposits of the Oltulelei Formation. The latter formation has an upper age limit of <36 ka (13) and a lower age limit established here of ~320 ka. The ~180-ka hiatus between the formations was a time of deep erosion (>60-m-deep valley formation) and transport of basin materials into the Koora Graben.

Members 1 to 14 of the Olorgesailie Formation encompass the longest series of excavated Acheulean sites in Africa. Members 8 to 14 fall in the early part of the middle Pleistocene and contain Acheulean industries characterized by bold percussion flaking and varying percentages of LCTs. Recently discovered sites in Members 11 and 12 provide evidence of variations in the Acheulean that are relevant to MSA origins (2). Here we provide evidence establishing the age of the Member 11 Acheulean sites and of the oldest MSA sites that lack LCTs, which have been excavated in the Olkesiteti Member of the Oltulelei Formation (1).

The chronological constraints on the MSA sites at Olorgesailie are mainly based on the 40Ar/39Ar dating technique applied to single phenocrysts of K-feldspar from volcaniclastic deposits. Two variants of the 40Ar/39Ar method were used; an early phase of the dating effort was based on single-crystal total-fusion analyses, whereas later analyses advanced to single-crystal incremental heating procedures (5). These approaches together were applied to a total of 28 samples from about 20 tuffaceous units. We also applied the U-series dating method to siliceous rhizoliths in growth orientation preserved in a vitric tuff bed (13). Samples were obtained from the vicinity of archaeological sites in Localities B and G and archaeological and geological sites in Locality A (Fig. 1 and table S2 provide sample locations and further details).

The latest known Acheulean archaeological sites in the Olorgesailie Formation and the remains of the distinctive large-bodied grazing fauna associated with these sites are located in Members 11, 12, and 14 (2). Previous 40Ar/39Ar ages constrained Member 11 to a 62-ka interval between underlying Member 10 at 671 ± 8 ka old and overlying Member 12 at 609 ± 6 ka old, with an age of 499 ± 2 ka for Member 14 (9). New 40Ar/39Ar age determinations for Member 11 tuffaceous deposits further constrain the archaeological materials in Member 11' to the lower part of the 6-ka interval from 615 to 609 ka ago (Fig. 1A). The 615-ka-old constraint is based on two ages obtained on pumice from a poorly sorted silty sandstone containing and underlying the artifacts (615.7 ± 6.9 ka from Locality A and 614.7 ± 5.2 ka from Locality B; a third result of 626.6 ± 7.7 ka from Locality A was rejected as being too old; supplementary text S3).

Archaeological sites that we dated in the Olkesiteti Member of the Oltulelei Formation in Locality B occur within a 4- to 6-m-thick sequence of siltstones, sandstones, and conglomerates, with occasional pumice gravels and vitric tuffs (Fig. 1B). The stratigraphically lowest MSA sites lie above an eroded surface of Member 9 of the Olorgesailie Formation (757 to 671 ka old). 40Ar/39Ar ages were obtained from three sites (BOK-1E, BOK-2, and BOK-4; Fig. 1B) (1), all within a lateral distance of 130 m.

MSA site BOK-1E underlies a succession of vitric tuffs with concordant ages of 303.7 ± 8.4, 306.1 ± 9.5, and 264.1 ± 1.3 ka (section B10-01/B08-06, Fig. 1B). These are in turn overlain unconformably by much younger pumice gravel of the Oltepesi Member of the Oltulelei Formation, which yielded an age of 158.5 ± 3.8 ka. Although the minimum age of the site is established by these dated tuffs at >305 ka, the maximum age is not similarly constrained and must be estimated through sedimentological arguments. We suggest an estimate of ~320 ka, based on considerations of fluvial sedimentology. This stratigraphic interval consists of a single upward-fining cycle with intervals of pedogenesis, and, for this fluvial system, ~10 to 20 ka (i.e., 15 ka) of accumulation is considered a reasonable estimate (16).

Site BOK-2 (section B03-03, Fig. 1B) is similarly constrained by an overlying pair of vitric tuffs, which are potentially correlative to those above BOK-1E. Their ages are 305.2 ± 6.8 and 302.2 ± 6.3 ka. The lowest dated tuff lies less than a meter above the BOK-2 artifact levels, without indication of a major unconformity. Site BOK-4 occurs within a filled channel incised into the BOK-1E/BOK-2 stratigraphy (section B06-02, Fig. 1B). 40Ar/39Ar ages were obtained on seven pumice gravels intercalated with the artifact horizons within a ~2.5-m interval. For convenience, we grouped the gravels into a lower bed, a middle sequence of five dated gravels, and a capping gravel with ages of 297.0 ± 3.2, 298.4 ± 1.5, and 295.1 ± 6.4 ka, respectively. The entire episode of more than 4 m of channel cutting and filling at BOK-4 was completed in just ~6 to 7 ka after deposition of sites BOK-1 and BOK-2. In brief, the BOK sites began accumulating as early as ~320 ka ago and terminated by ~295 ka ago.

MSA sites in Locality G are distributed in a roughly north-south orientation over a distance of ~1.1 km in strata of the Olkesiteti Member of the Oltulelei Formation (Fig. 1C). Most of the tuff samples that date these sites also are from this member. Six pumiceous tuff units were analyzed by the 40Ar/39Ar method, with additional age control from U-series dating.

The U-series sample, consisting of silica and carbonate replacing roots or stems in the oldest tuff unit in Locality G, yielded the precise age of 277.1 ± 1.8 ka (unit G-T1 in section G06-08, Fig. 1C). This tuff occurs in strata that overlie an erosional unconformity above the MSA archaeological material at GOK-1, which occurs in a thick (~2.5 m), well-developed reddish paleosol (section G08-01, Fig. 1C). Several factors suggest that this site is considerably older than the U-series date: its occurrence in a thick paleosol, the presence of an erosional unconformity at the top of the paleosol marking a cut-and-fill sedimentary cycle, and >1 m of strata beneath the dated tuff in section G06-08. These elements suggest that GOK-1 may be of antiquity equal to or greater than the BOK MSA sites.

Multiple 40Ar/39Ar results from pumice tuffs in Locality G above the level of the U-series date tightly constrain the age of the prolific MSA site GOK-1s (Fig. 1C). This site occurs within a few tens of centimeters of fine siltstone between the upper pair of a widespread group of three closely spaced tuffs, informally called the “Triple Tuffs” and assigned to the Oltulelei Tuffs (13). These tuffs yielded indistinguishable mean ages from 219.5 ± 3.6 to 221.8 ± 3.5 ka (Fig. 1C). Using a Bayesian chronology model for the Locality G samples (fig. S7), we estimate that the duration of the stratigraphic interval of GOK-1s was ≤1.5 ka.

MSA artifact sites GOK-3 and GOK-4 in Locality G are also constrained by the underlying Triple Tuffs and by two overlying pumice horizons within ~1.5 m of the artifact levels: a 3- to 10-cm pumice tuff unit at 215.3 ± 2.3 ka old (G-T6) and pumice gravel within the erosional base of a channel deposit at 207.1 ± 4.4 ka old (G-T7) (Fig. 1C).

This sequence of precise dates from the Olorgesailie basin in the southern Kenya rift, relying primarily on the 40Ar/39Ar technique but also using innovative U-series methodologies, provides secure chronostratigraphic constraints on the early establishment of MSA lifeways in eastern Africa and on the end of the Acheulean, which must have occurred by ~320 ka ago. Although the latest Acheulean site is found in Member 14 of the Olorgesailie Formation at ~499 ka old, earlier variations in the Acheulean showing a greater degree of stone raw material selectivity for toolmaking—including rocks from distant sources, thus potentially anticipating aspects of MSA behavior (2)—occur in Member 11 at ~615 to 609, Member 12 at ~609, and Member 14 at ~499 ka old. After a period of deep erosion at Olorgesailie between ~499 and ~320 ka ago, MSA technological innovations, including important cultural markers such as long-distance obsidian transport and pigment processing (1), were preserved in the Oltulelei Formation, beginning most likely by 320 and no later than 305 ka ago. These ages also imply that a major shift in the landscapes and mammalian biota of the southern Kenya rift (2) had occurred by 320 ka ago.

Our newly calibrated archaeological sequence provides a secure basis for comparison with the limited sample of other well-dated Acheulean and MSA occurrences in Kenya and elsewhere in Africa. Evidence for an increased pace of environmental change that now is well documented in the Olorgesailie sequence suggests that eastern Africa was important to the development of MSA technological, social, and cognitive innovations in human behavior before 300 ka ago, as a key region fostering the development of a more widespread set of behavioral changes related to the emergence of H. sapiens.

Supplementary Materials

www.sciencemag.org/content/360/6384/95/suppl/DC1

Materials and Methods

Supplementary Text

Figs. S1 to S8

Tables S1 to S10

References (3465)

References and Notes

  1. Materials and methods are available as supplementary materials.
Acknowledgments: We thank the National Museums of Kenya and the Kenya Government for permission granted by the Ministry of Sports, Culture and the Arts and NACOSTI (National Commission for Science, Technology and Innovation) permit P/14/7709/683. We thank J. Clark for assistance in preparation of the manuscript. Funding: Laboratory analyses were supported by National Science Foundation grant EAR-1322017 (A.L.D.). Fieldwork was funded by the Peter Buck Fund for Human Origins Research, the Human Origins Program (Smithsonian), and NSF HOMINID program grant BCS-0218511 (R.P.). Author contributions: A.L.D. contributed 40Ar/39Ar geochronology; W.D.S. contributed U-series geochronology; A.K.B. led stratigraphic studies, with input from A.L.D., R.P., and W.D.S.; and archaeological data were contributed by A.S.B., J.E.Y., and R.P. 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 necessarily reflect the views of the NSF. Data and materials availability: All geochronological data are available in the paper or in the supplementary materials. Samples and field notes are archived at the Berkeley Geochronology Center.
View Abstract

Stay Connected to Science

Navigate This Article