Prehistory

A 4.5 Ma record of fluvial and lacustrine deposits is well exposed at Gona, in the Afar Depression of Ethiopia. We use isotopic values of pedogenic carbonate and fossil teeth to reconstruct Plio–Pleistocene environmental change at Gona. An increase in δ13C values of pedogenic carbonates since 4.5 Ma points to a shift from woodlands to grassy woodlands in the early Pliocene, −10.4 to −3.9‰ (VPDB), to more open but still mixed environments in the late Pleistocene, −3.0 to −1.4‰ (VPDB). This pattern is also seen in isotopic records elsewhere in East Africa. However, at 1.5 Ma the higher proportion of C4 grasses at Gona is largely a result of a local facies shift to more water-limited environments. The wide range of δ13C values of pedogenic carbonate within single stratigraphic levels indicates a mosaic of vegetation for all time intervals at Gona that depends on depositional environment. Elements of this mosaic are reflected in δ13C values of both modern plants and soil organic matter and Plio–Pleistocene soil carbonate, indicating higher amounts of C4 grasses with greater distance from a river channel in both the modern and ancient Awash River systems. δ18O values of pedogenic carbonates increase up-section from −11.9‰ in the early Pliocene to −6.4‰ (VPDB) in the late Pleistocene. The wide range of δ18O values in paleovertisol carbonates from all stratigraphic levels probably reflects short-term climate changes and periods of strong evaporation throughout the record. Based on the comparison between δ18O values of Plio–Pleistocene pedogenic carbonates and modern waters, we estimate that there has been a 6.5‰ increase in mean annual δ18O values of meteoric water since 4.5 Ma. δ18O values of pedogenic carbonate from other East African records indicate a similar shift. Increasing aridity and fluctuations in the timing and source of rainfall are likely responsible for the changes in δ18O values of East African pedogenic carbonates through the Plio–Pleistocene.

Published evidence of Oldowan stone exploitation generally supports the conclusion that patterns of raw material use were determined by local availability. This is contradicted by the results of systematic studies of raw material availability and use among the earliest known archaeological sites from Gona, Afar, Ethiopia. Artifact assemblages from six Pliocene archaeological sites were compared with six random cobble samples taken from associated conglomerates that record pene-contemporaneous raw material availability. Artifacts and cobbles were evaluated according to four variables intended to capture major elements of material quality: rock type, phenocryst percentage, average phenocryst size, and groundmass texture. Analyses of these variables provide evidence of hominid selectivity for raw material quality. These results demonstrate that raw material selectivity was a potential component of Oldowan technological organization from its earliest appearance and document a level of technological sophistication that is not always attributed to Pliocene hominids.

Newly recorded archaeological sites at Gona (Afar, Ethiopia) preserve both stone tools and faunal remains. These sites have also yielded the largest sample of cutmarked bones known from the time interval 2.58–2.1 million years ago (Ma). Most of the cutmarks on the Gona fauna possess obvious macroscopic (e.g., deep V-shaped cross-sections) and microscopic (e.g., internal microstriations, Herzian cones, shoulder effects) features that allow us to identify them confidently as instances of stone tool-imparted damage caused by hominid butchery. In addition, preliminary observations of the anatomical placement of cutmarks on several of the recovered bone specimens suggest that Gona hominids may have eviscerated carcasses and defleshed the fully muscled upper and intermediate limb bones of ungulates—activities that further suggest that Late Pliocene hominids may have gained early access to large mammal carcasses. These observations support the hypothesis that the earliest stone artifacts functioned primarily as butchery tools and also imply that hunting and/or aggressive scavenging of large ungulate carcasses may have been part of the behavioral repertoire of hominids by c. 2.5 Ma, although a larger sample of cutmarked bone specimens is necessary to support the latter inference.

Inter-site technological variation in the archaeological record is one of the richest potential sources of information about Plio-Pleistocene hominid behavior and evolution. However, appropriate methods for describing and comparing Oldowan assemblages have yet to be agreed upon, and interpretation of the early record remains highly controversial. Particularly salient is disagreement over whether the Oldowan is a single technological phenomenon or is more accurately divided into multiple regional and/or chronological traditions, perhaps including a less developed Pre-Oldowan phase in the late Pliocene. Some of this disagreement reflects theoretical and methodological differences between research traditions and some is more directly evidential. Here we present a framework for describing and interpreting Oldowan variation and apply it to three Pliocene assemblages (EG-10, EG-12, and OGS-7) from Gona, all dated to c. 2.6 million years (Ma). Results indicate proficient knapping and a full range of Oldowan reduction strategies in these earliest known occurrences, consistent with the idea of an Oldowan “technological stasis” from 2.6–1.6 Ma. Patterns of variation in raw material selection and predominant reduction strategy at each site clearly indicate the importance of cultural transmission in the Oldowan, but confounding ecological and economic variation continue to render interpretation in terms of multiple tool making traditions or species inappropriate. We propose that cultural transmission and ecological adaptation should be recognized as complementary, rather than mutually exclusive, mechanisms in future attempts to explain Oldowan technological variation.