Joseph V. Ferraro

Earliest Archaeological Evidence of Persistent Hominin Carnivory

The emergence of lithic technology by ∼2.6 million years ago (Ma) is often interpreted as a correlate of increasingly recurrent hominin acquisition and consumption of animal remains. Associated faunal evidence, however, is poorly preserved prior to ∼1.8 Ma, limiting our understanding of early archaeological (Oldowan) hominin carnivory. Here, we detail three large well-preserved zooarchaeological assemblages from Kanjera South, Kenya. The assemblages date to ∼2.0 Ma, pre-dating all previously published archaeofaunas of appreciable size. At Kanjera, there is clear evidence that Oldowan hominins acquired and processed numerous, relatively complete, small ungulate carcasses. Moreover, they had at least occasional access to the fleshed remains of larger, wildebeest-sized animals. The overall record of hominin activities is consistent through the stratified sequence – spanning hundreds to thousands of years – and provides the earliest archaeological evidence of sustained hominin involvement with fleshed animal remains (i.e., persistent carnivory), a foraging adaptation central to many models of hominin evolution.

Oldest Evidence of Toolmaking Hominins in a Grassland- Dominated Ecosystem

Background Major biological and cultural innovations in late Pliocene hominin evolution are frequently linked to the spread or fluctuating presence of C4 grass in African ecosystems. Whereas the deep sea record of global climatic change provides indirect evidence for an increase in C4 vegetation with a shift towards a cooler, drier and more variable global climatic regime beginning approximately 3 million years ago (Ma), evidence for grassland-dominated ecosystems in continental Africa and hominin activities within such ecosystems have been lacking. Methodology/Principal Findings We report stable isotopic analyses of pedogenic carbonates and ungulate enamel, as well as faunal data from ∼2.0 Ma archeological occurrences at Kanjera South, Kenya. These document repeated hominin activities within a grassland-dominated ecosystem. Conclusions/Significance These data demonstrate what hitherto had been speculated based on indirect evidence: that grassland-dominated ecosystems did in fact exist during the Plio-Pleistocene, and that early Homo was active in open settings. Comparison with other Oldowan occurrences indicates that by 2.0 Ma hominins, almost certainly of the genus Homo, used a broad spectrum of habitats in East Africa, from open grassland to riparian forest. This strongly contrasts with the habitat usage of Australopithecus, and may signal an important shift in hominin landscape usage.

Oldowan Technology and Raw Material Variability at Kanjera South

Advances in the study of Oldowan research have suggested that the earliest tool-makers had the technological capabilities usually suggested in later time periods. Work in West Turkana and Gona research areas suggests that Pliocene hominins had a concise understanding of stone fracture mechanics and had a clear conception of how to reduce cores in a manner that maintained flaking surfaces. Here we investigate if these same patterns existed at the Pliocene site of Kanjera South in Western Kenya. Technological analyses suggest that although many of the technological capabilities described for other Oldowan sites are present in the Kanjera South assemblage, specific aspects of the context of the site (raw material variability) produced a different expression of these behaviors. The most obvious difference between the Kanjera South site and other Oldowan sites is that as reduction continues several different reduction patterns can be seen. This suggests that a reduction sequence or core reduction mode is not an immutable formula and can change depending on its context.

The Environmental Context of Oldowan Hominin Activities at Kanjera South, Kenya

The earliest archaeological traces and two new hominin genera (Homo and Paranthropus) appear in the late Pliocene of Africa. These first appearances may reflect novel hominin adaptive responses to shifting resource bases over geological time and/or an increasingly seasonal distribution of food over the annual cycle. Whereas regional environmental change has been documented during the Plio-Pleistocene of East Africa, it is difficult to resolve relative proportions of specific habitats at a given place and time, how these may have changed over time, and the explicit nature of particular habitats. Detailed reconstructions of paleohabitats based on paleontological, geological and geochemical evidence are necessary in order to better understand the interplay between environmental change and hominin biological and behavioral evolution.

Bovid ecomorphology and hominin paleoenvironments of the Shungura Formation, lower Omo River Valley, Ethiopia

The Shungura Formation in the lower Omo River Valley, southern Ethiopia, has yielded an important paleontological and archeological record from the Pliocene and Pleistocene of eastern Africa. Fossils are common throughout the sequence and provide evidence of paleoenvironments and environmental change through time. This study developed discriminant function ecomorphology models that linked astragalus morphology to broadly defined habitat categories (open, light cover, heavy cover, forest, and wetlands) using modern bovids of known ecology. These models used seven variables suitable for use on fragmentary fossils and had overall classification success rates of >82%. Four hundred and one fossils were analyzed from Shungura Formation members B through G (3.4-1.9 million years ago). Analysis by member documented the full range of ecomorph categories, demonstrating that a wide range of habitats existed along the axis of the paleo-Omo River. Heavy cover ecomorphs, reflecting habitats such as woodland and heavy bushland, were the most common in the fossil sample. The trend of increasing open cover habitats from Members C through F suggested by other paleoenvironmental proxies was documented by the increase in open habitat ecomorphs during this interval. However, finer grained analysis demonstrated considerable variability in ecomorph frequencies over time, suggesting that substantial short-term variability is masked when grouping samples by member. The hominin genera Australopithecus, Homo, and Paranthropus are associated with a range of ecomorphs, indicating that all three genera were living in temporally variable and heterogeneous landscapes. Australopithecus finds were predominantly associated with lower frequencies of open habitat ecomorphs, and high frequencies of heavy cover ecomorphs, perhaps indicating a more woodland focus for this genus.

Late Pliocene Oldowan excavations at Kanjera South, Kenya

The appearance of Oldowan sites c. 2.5 million years ago signals one of the most important adaptive shifts in human evolution. Large mammal u butchery, stone artefact manufacture and novel transport and discard behaviours led to the accumulation of the first recognized archaeological debris. Although the earliest instances of these behaviours are 2.5 million years ago, most of what we know about Oldowan palaeoecology and behaviour is derived from localities more than half a million years younger, particularly c. 1.8 million-year-old sites from Bed I Olduvai Gorge, Tanzania (Potts 1988). Sites from Kanjera South, Homa Peninsula, southwestern Kenya, yield dense concentrations of artefacts in association with the oldest (c. 2.2 million years) substantial sample of archaeological fauna known thus far from Africa. This study is the first to use a wide range of traditional and innovative techniques to investigate Oldowan hominin behaviour and site formation processes before 2 million years ago.

Contesting early archaeology in California

The peopling of the Americas is a topic of ongoing scientific interest and rigorous debate1,2. Holen et al.3 add to these discussions with their recent report of a 130,000-year-old archaeological site in southern California, USA: the Cerutti Mastodon (CM) site, which includes the fragmentary remains of a single mastodon (Mammut americanum), spatially associated stone cobbles, and associated lithic debris that they claim indicates prehistoric hominin activity. In sharp contrast, we contend that the CM record is more parsimoniously explained as the result of common geological and taphonomic processes, and does not indicate prehistoric hominin involvement. Whereas further investigations may yet identify unambiguous evidence of hominins in California around 130,000 years ago, we urge caution in interpreting the current record. There is a Reply to this Comment by Holen, S. R. et al. Nature 554, http://dx.doi.org/10.1038/nature25166 (2018). Holen et al.3 claim prehistoric hominin involvement at the CM site based primarily on four lines of evidence: a reliable radiometric age; the presence of stone artefacts; clear evidence of tool-imparted percussion damage to the remains of the mastodon; and an undisturbed geological context. We take no issue with the published age for the site, but we believe that the latter three claims warrant further examination. The CM site stone artefacts are an assortment of cobbles and fractured cobble fragments excavated from a sandy silt matrix. There is no evidence of formal stone tool forms or diagnostic lithic microor macro-debitage. Instead, the CM site artefacts are identified by their proximity to the remains of the mastodon and their large size relative to the enveloping sediment. Additionally, surface features such as the presence of pitting and scratching on cobble surfaces, the presence of several cobble fragments with fracture morphologies reminiscent of hammerstone and/or anvil usage, and the presence of several refitting cobble fragments are interpreted as evidence of hominin percussive activities on-site. The lack of discarded formal tools and diagnostic lithic debitage is noteworthy, and is unusual relative to most archaeological assemblages that purport hominin processing of proboscidean remains (although see Haynes4). We also note that upslope of the site there are numerous alluvial fans, with clastic material a common occurrence. The cobbles and pebbles at the CM site can be derived from modest alluvial fan input with fines subsequently winnowed with lower energy fluvial erosion. Crucially, none of the criteria that Holen et al.3 use to define stone artefacts either requires prehistoric hominin involvement or meets the accepted criteria for falsifying natural ‘geofacts’5. The range of possible geological interpretations for the lithic assemblage highlights the critical issue of equifinality, in which an end product such as a shattered cobble may be generated via many and potentially unrelated means. It is a principle that applies to the bone record as well. We contend that Holen et al.3 presented equivocal evidence in support of tool-imparted percussion damage to the remains of the mastodon. The critical observations are of spiral fractures, cone flakes, impact flakes, bulbs of percussion, impact notches, negative flake scars, anvil-polished specimens, percussion-fractured specimens, and refitting specimens. These bone damage features are inferred to implicate sole agency by prehistoric hominins. As with the stone artefact record, issues of equifinality must first be addressed, including the question of whether other processes could produce such bone damage. Haynes6 presents compelling evidence of non-cultural and/or non-prehistoric processes producing comparable damage at the 24,000-year-old Inglewood Mammoth Site (IMS), Maryland, USA. As at the CM site, the IMS contains the remains of a single juvenile proboscidean recovered in situ from a sealed deposit of sandy clays with pebbles and cobbles6. Haynes6 provides descriptions and images of curvilinear and spiral ‘green-bone’ fractures on cranial, axial and appendicular specimens. Some of these fractures are recent in origin, probably related to heavy earthmoving equipment working on-site6. Other damage may reflect perimortem injuries sustained by the mammoth. No evidence of prehistoric hominin activities are noted or suspected for the site. Post-burial bone notches, impact points and impact scratches occurred on a number of specimens. We report a similar record of fractured proboscidean bones at the Waco Mammoth National Monument (WMNM), Waco, Texas, USA. The site contains the remains of at least 26 mammoths buried in fluvial sands, silts and clays, and dates from 66,800 to 51,300 years ago7. The WMNM was initially investigated as a potential archaeological site, although no evidence of prehistoric human activities was discovered. Figure 1 shows post-burial damage to WMNM mammoth long bones morphologically consistent with observations from the IMS and CM sites. This includes damage that resembles spiral fractures with associated sedimentary abrasion, hammerstone pseudo-notches8, negative flake scars, partially detached flakes and incipient notches, and bulbs of percussion. Such damage, including spiral fractures, is well known in the fossil record from as early as the Triassic period9 and can occur post-burial6. They neither require nor solely indicate prehistoric human agency4,6,8. Other proboscidean assemblages share a similar taphonomic signature with the WMNM, IMS and CM sites. Holen and others report various combinations of spiral fractures, impact points, bulbs of percussion and bone flakes at numerous other late Pleistocene mammoth death sites in the Americas10,11. As with the CM site, these latter assemblages uniformly lack unambiguous stone tools, cut marks, or any other unquestionable evidence of hominin activities, and most predate well-vetted geochronological and palaeogenomic evidence of the initial peopling of the Americas around 15.5 thousand years ago1,2,12–14. Moreover, it is not just what is present at the CM site, but also what is missing. Specifically, hammerstone striae and/or pits (HSSP)15 are noticeably absent despite reasonable cortical bone preservation, several hundred bone fragments, purported hammerstones, and purported anvil abrasions on both the bones and the cobbles. Experimental studies show that hammerstone-percussed proboscidean limb bone fragments should bear HSSP on greater than 30% of specimens created when using a hafted hammerstone and anvil12. The absence of HSSP at the CM site—a proposed percussed bone assemblage—cannot be explained using current experimental models and contradicts the assumption of hammerstone-wielding hominin involvement in bone breakage. Lastly, we question the assertion of an “undisturbed geologic context” at the CM site. Although the distance between some refitted finds necessarily suggests pre-burial breakage and scattering of some items (for example the molar fragments), other features of the record plausibly reflect subsequent forces modifying the assemblage over the last 130,000 years. As fluvial deposits slowly covered the remains, the bones of the mastodon would have remained semi-pliable for years6. Proboscideans or other large mammals subsequently using the muddy watercourse could potentially trample, displace, fracture, abrade and reorient (for example the semi-vertical tusk) the interred materials4,6. Later sediment compaction by metres of overburden and then eventual

Integrating Human Activities, Archeology, and the Paleo-Critical Zone Paradigm

Recent conceptual advances in the Earth sciences have led to an improved understanding of the dynamics governing the Critical Zone (CZ)—the interface where life meets rock and soil on land (Brantley et al., 2007; Nordt andDriese, 2013). Among the key insights is a renewed appreciation for the deeply intertwined and non-linear nature of the processes in play, where small changes in one or two variable values or in their interactions can have large and often non-intuitive consequences, including the emergence of complex phenomena (Brantley et al., 2017). This application of complex systems (CS) methods and theory to conceptualize the CZ, both in part and in whole, places a premium on valid model construction. The potential analytical consequences of inaccurately modeling variables, processes, or interactions here though is dramatic, especially in light of the non-linear nature and coupled dynamics of CSmodels. In this brief note, we offer our opinion to the CZ science community that CS modeling of Holocene CZ processes and records (i.e., the paleo-CZ [pCZ]; Beverly et al., 2017) may be further refined–and perhaps must be further refined–through increased integration of archeological data and theory.