The finding of 3.3-million-year-old stone flints, cores, hammers and anvils in Kenya suggests that the first stone tools were made by human ancestors that pre-dated the earliest known members of the genus Homo. See Article p.310
The earliest production of stone tools by human ancestors marked the dawn of an innovative behavioural strategy that transformed the ecology, social systems and culture of humans through social learning of skills and technology. Direct evidence for this behaviour is found in the archaeological record, and its earliest appearance has been gradually pushed deeper into time over the past five decades1. On page 310 of this issue, Harmand et al.2 report findings from Lomekwi 3 (LOM3), a site on the western side of Lake Turkana, Kenya, that extend the record of material culture by some 700,000 years, to 3.3 million years ago.
Palaeoanthropologists have long predicted, on evolutionary and archaeological grounds, that the first stone tools should be older than the previously oldest known instances, which date to around 2.6 million years ago (Ma). Once thought to be unique behaviours of hominins (the group that includes humans and their close extinct ancestors), tool use and toolmaking are now well documented in extant taxa of non-human primates — mainly chimpanzees, but also orangutans, gorillas and certain monkeys. Parsimony suggests that hominin tool use and toolmaking were practised by the last common ancestor of chimpanzees and hominins3, some 7–5 Ma, but that the making of tools from stones is unique to hominins, with the notable exception of chimpanzee nut-cracking hammer stones4; the tools of other primates are mostly twigs or other plant matter. However, the stone tools at sites dated to 2.6 Ma or slightly later, which are assigned to the archaeological complex known as the Oldowan1, seem to be too well made to have been the first experiments of early humans in producing sharp-edged stone flakes by free-hand core-reduction techniques (when a stone block, or core, is held in the hand and hit with a hammer stone)5,6.
Thus, core reduction as a technique for producing sharp edges for cutting plant and animal material was hypothesized to have first emerged among early hominin ancestors, but the time of its first appearance remained undefined in the absence of direct evidence. Furthermore, archaeologists did not have a 'template' with which to search for these proposed very early stone tools, and open questions remained regarding the selective processes that might have led to the emergence of stone toolmaking.
Harmand and colleagues' findings inform us on the first two issues. Extensive geological mapping of the project area, and the presence of well-dated volcanic ash layers, together with palaeomagnetic correlations, constrained the age of the geological layer in which the authors had found the tools to 3.31–3.21 Ma. This range was narrowed down to 3.3 Ma on the basis of sedimentation rates. Because the sediments in these layers are fine-grained, and a flake found by the authors could be fitted back onto the core from which it had been detached, it is unlikely that the tools accumulated through stream activity or that substantial disturbance of the sediments occurred after the tools had been discarded.
The small collection of stone tools at LOM3 is unlike those from known Oldowan localities, which contain mainly flakes. Most of the LOM3 items (around 76%) are cores, anvils, hammer stones and worked cobbles, indicating that the main activities were associated with pounding against anvils rather than hand-held core reduction. Some items bear traces of having been used for multiple techniques — as an anvil or hammer stone, or as a core to produce flakes — which hints at their use for exploiting variable resources (Fig. 1). All types of object in LOM3 are larger than their counterparts in the later Oldowan sites. The flakes are more massive than the accidentally produced flakes found in chimpanzee nut-cracking localities, but the anvils and hammers are within the size range of chimpanzee nut-cracking kits4,7,8.
The surface characteristics of the items indicate the use of forceful blows during the application of both pounding and hand-held core reduction. However, whether the hand anatomy and precision grip that are needed for tool use and toolmaking were already present in the various species of hominin that existed around 3 Ma is debated9,10,11. The earliest known12 appearance of the genus Homo is at 2.8 Ma, suggesting that the LOM3 toolmakers may have belonged to other lineages of early hominin.
Harmand and colleagues contrast the postulated multiple uses for the LOM3 items with the generally single-purpose tools used by extant non-human primates. The authors suggest that the LOM3 tools could represent a technological stage between a hypothetical pounding-oriented stone-tool use by hominins earlier than those at LOM3 and the flaking-oriented behaviour of later Oldowan toolmakers. Primatologists may take issue with the first statement and argue that tool use in primates is multifaceted13. Evolutionary theorists may prefer less gradualist interpretations, and archaeologists could argue that one must not exclude the possibility that, at the beginning of each discrete episode of its use, each stone object was perceived merely as available raw material. The cognitive implications in this last case would differ from those offered by Harmand and colleagues. Therefore, we should focus on the evidence14,15 for core reduction as a marker of new cognitive abilities and a new technological path on which hominins embarked at 3.3 Ma.
The age and nature of the finds from LOM3 call for a re-evaluation of models16,17 that tie together the timing and patterns of environmental change, hominin evolution and the origins of technological behaviour around 2.5 Ma. However, caution is warranted. Our understanding of ancient hominins and their cognitive, cultural and social capacities is only as good as the available archaeological and fossil data. Similar to animal bones from Dikika, Ethiopia, that date to at least 3.39 Ma and arguably bear stone-inflicted cut marks18, the stone tools from LOM3 are at present an isolated occurrence. To maintain that either of these instances marks innovations in hominin behavioural evolution, the temporal gaps must be filled in with more data. In this respect, the LOM3 discoveries stand to have an immediate impact on human-origins research in eastern Africa by providing the long-needed search template for early stone tools.
Moreover, until now, the search for 'older than the Oldowan' archaeological sites has focused on a few areas that contain sediments dated to between 2.9 and 2.6 Ma, with the aim of establishing a sequence with known Oldowan sites. The discoveries at LOM3 allow research also to focus on the time range 3.4–2.9 Ma, which so far has not been tapped for evidence of material culture. And why not dig deeper in time? LOM3 may not be the final — or rather, the first — word on the roots of human technology.