Matt Caruana

Recent studies have identified differences in handaxe reduction strategies within the Acheulian assemblages from Amanzi Springs, with operational sequences that involve a variety of giant core methods to produce large flake blanks, as well as being made directly on cobbles. Despite these different blank selection patterns, there is a general standardisation in the final morphology of handaxes from Area 2 (~530-<408 ka). This study uses three-dimensional geometric morphometric, descriptive statistics and diacritical analyses to explore large flake usage at the site, and its implications in handaxe morphology and manufacture. Our results demonstrate that Amanzi knappers used large flake blanks with standardised characteristics and morphologies to shortcut challenging technical aspects of handaxe production. Despite previous descriptions of handaxes being large and unstandardised in appearance, Middle Pleistocene knappers at Amanzi Springs were able to anticipate challenges of the locally available raw materials by producing a range of large flake blank morphologies to overcome knapping mishaps.

Renewed research at Amanzi Springs has increased resolution on the timing and technology of the Acheulian industry in South Africa. The archeology from the Area 1 spring eye has recently been dated to MIS 11 (~404-390 ka), and analyses revealed significant technological variability when compared to other southern African Acheulian assemblages. We expand on these results in presenting new luminescence dating and technological analyses of Acheulian stone tools from three artifact-bearing surfaces exposed within the White Sands unit of the Deep Sounding excavation in the Area 2 spring eye. The two lowest surfaces (Surfaces 3 and 2) are sealed within the White Sands and dated between~534 to 496 ka and~496 to 481 ka (MIS 13), respectively. Surface 1 represents materials deflated onto an erosional surface that cut the upper part of the White Sands (~481 ka; late MIS 13), which occurred before the deposition of younger Cutting 5 sediments (<408-<290 ka; MIS 11e8). Archaeological comparisons reveal that the older Surface 3 and 2 assemblages are predominated by unifacial and bifacial core reduction and relatively thick, cobble-reduced large cutting tools. In contrast, the younger Surface 1 assemblage is characterized by discoidal core reduction and thinner large cutting tools, mostly made from flake blanks. Typological similarities between the older Area 2 White Sands and younger Area 1 (404-390 ka; MIS 11) assemblages further suggest long-term continuity in site function. We hypothesize Amanzi Springs represent a workshop locality that Acheulian hominins repeatedly visited to access unique floral, faunal, and raw material resources from at least~534 to 390 ka.

We present the first techno-functional examination of backed knives from the southern African Acheulean. Our results suggest that they were opportunistically produced, although they demonstrate a unique ergonomic design that may have increased their efficiency in subsistence activities. Moreover, the frequency of backed knives at Wonderboom may be associated with possible meat harvesting at a nearby gap (Wonderboompoort) in the Magaliesberg range, which formed a bottleneck for animal herds migrating across major biome boundaries in the deep past. The Wonderboom knives might have been made on an ad hoc basis to augment butchery practices.

Wonderboom remains largely excluded from discussions about the Earlier Stone Age of southern Africa, despite having one of the largest Acheulean assemblages for the region. With this contribution, we revisit its large cutting tool assemblage (namely the handaxes, cleavers, picks and knives) and investigate some of the tool manufacturing preferences of the hominins who used the site during the Earlier Stone Age. We also provide an inter-site comparison of handaxes with another later and two early Acheulean sites from South Africa, wherein the Wonderboom assemblage clusters with the later Acheulean and is distinguished from the early Acheulean assemblages. This tentatively places Wonderboom within the South African Acheulean chronology and provides the first characterization of later Acheulean LCT production strategies for the region that includes the Cradle of Humankind.

Amanzi Springs is a series of inactive thermal springs located near Kariega in the Eastern Cape of South Africa. Excavations in the 1960s exposed rare, stratified Acheulian-bearing deposits that were not further investigated over the next 50 years. Reanalysis of the site and its legacy collection has led to a redefined stratigraphic context for the archaeology, a confirmed direct association between Acheulian artefacts and wood, as well as the first reliable age estimates for the site. Thermally transferred optically stimulated luminescence and post-infrared infrared stimulated luminescence dating indicates that the Acheulian deposits from the Amanzi Springs Area 1 spring eye formed during Marine Isotope Stage (MIS) 11 at 404-390 ka. At this time, higher sea levels of~13-14m would have placed Amanzi Springs around 7 km from a ria that would have formed along what is today the Swartkops River, and which likely led to spring reactivation. This makes the Amanzi Springs Area 1 assemblage an unusual occurrence of a verified late occurring, seaward, open-air Acheulian occupation. The Acheulian levels do not contain any Middle Stone Age (MSA) elements such as blades and points that have been documented in the interior of South Africa at this time. However, a small number of stone tools from the upper layers of the artefact zone, and originally thought of as intrusive, have been dated to~190 ka, at the transition between MIS 7 to 6, and represent the first potential MSA identified at the site.

Recent excavations at Penhill Farm and Amanzi Springs have reinvigorated interest in the Acheulean archaeological record of the Eastern Cape Province, South Africa. While this research now provides valuable detail on hominin adaptations in environments that differ from the interior, few Acheulean assemblages in this region have been recorded or thoroughly analysed. Here we compare Acheulean handaxes from Penhill Farm and the Amanzi Springs Area 1 locality to help characterise the expression of this technocomplex in the Eastern Cape. We employ a multivariate analysis of allometry to highlight the relationship between shape variance in relation to the size of handaxes, which further provide perspective on shaping processes. Results demonstrate high levels of techno-morphological variability that may distinguish Acheulean handaxes in the Eastern Cape region from sites elsewhere. We further argue that morphological variation in handaxes from Penhill Farm and Amanzi Springs may have also been influenced by site function, discard behaviours and group mobility patterns. These data refute the notion that the Acheulean technocomplex represents a single, homogenous technological entity, but rather was a flexible tradition that was influenced by region-specific factors.

Studies investigating the production of later Acheulian handaxes in South Africa remain rare.
Acheulian assemblages in this region are typically fragmented, which hinders our ability to
investigate a full range of manufacturing stages at any one given site. Insights can therefore
only be gained through comparative research involving multiple assemblages. To explore this
issue, handaxes from Amanzi Springs and Cave of Hearths are contrasted to identify their
position within a common reduction continuum. Results suggest that the Amanzi Springs
handaxes represent early shaping phases focused on defining bifacial edges, while the Cave of
Hearths reflects the later stages of thinning. Multivariate statistical analyses are used to
extrapolate reduction trends that provide perspective on how later Acheulian handaxes in
South Africa were shaped.

Abstract The Drimolen Palaeocave System in the ‘Fossil Hominid Sites of South Africa’ UNESCO World Heritage Site is well known for numerous remains of early hominins such as Paranthropus robustus and early Homo. These hominin fossils, along with bone tools and notably diverse accumulation of non-hominin primates and fauna, have all been excavated from the &#39;Main Quarry&#39; area of the site where extensive lime-mining took place. Here we report the first radiometric age of 1.712 ± 0.269 Ma for hominin bearing deposits associated with the DNH7 Paranthropus robustus cranium in the Main Quarry area of the site, which is consistent with recent biochronological estimates. This age is similar to recent estimates for Swartkrans Member 1 Hanging Remnant (somewhere between 2.3 and 1.8 Ma) which also contains Paranthropus and early Homo. Simultaneously, we integrate the newly radiometrically dated Main Quarry deposits with a new fossil deposit, the Drimolen Makondo, discovered in 2013, that is situated some 50 m up the hill to the west from the Main Quarry. It has experienced only limited disturbance from mining but much more extensive erosion. Preliminary excavations and analysis have revealed that the Makondo infill is older than the Main Quarry, dating to 2.706 ± 0.428 Ma. Its greater age is confirmed by biochronology. The Makondo thus overlap with the suggested end of deposition of Australopithecus bearing Sterkfontein deposits, although it is yet to yield any hominin remains. These new dates for the two Drimolen Palaeocave System deposits indicates that, contrary to prior age estimates, the Drimolen site as a whole records the critical hominin and faunal turnover in South African palaeocommunities that occurred around 2.3–1.7 Ma. Finally, as the Drimolen Makondo represents a rare example of a pre-2 Ma fossil bearing deposit in the Gauteng exposures of the Malmani dolomite, we also integrate our results into the greater South African record of palaeodeposit formation (most of which occur between ∼2.0 and 1.0 Ma). An analysis of the age of palaeocave infillings across the Malmani dolomite suggests that, as is classically the case with karst, the height within the dolomite is broadly correlated to their age, although with some notable exceptions that are likely related to localised geological features. Our analysis also indicates that most caves have undergone some form of secondary karstification related to a younger phase of cave formation, contrasting with models that suggest the cavities all formed at the same time and that infill is related to erosion and the opening up of cave passages. As such, the reason that few pre-2 Ma deposits have been identified in the Gauteng exposures of the Malmani dolomite is probably because these older caves have been eroded away. Identifying such early caves is critical in understanding whether earlier hominins may have once existed in South Africa or if erosion of older deposits (or an absence of speleogenesis at this time) has made such early periods absent from the geological record.

Experimental research examining Later Acheulian handaxes often equates their production with expert knapping skill. Yet knapping mishaps, including step and hinge fractures, mismanaging cross-sectional proportions and transverse breaks are common throughout manufacturing stages and may represent an important source of morphological variability. Recent investigations of handaxes from the ~400 ka Area 1 locality at Amanzi Springs have suggested that the large and asymmetric proportions of these tools may relate the abundance of step and hinge fractures and excessive thick edges that impeded reduction. This study uses both geometric and traditional morphometric techniques to explore the effects of knapping mishaps on the reduction and morphology of Amanzi Springs handaxes. Results suggest that interpretations of Acheulian technological skill can benefit from examining the occurrence of production errors, which may account for some of the morphological differences in handaxes observed on both diachronic and geographic scales.

Figure S1 displays PCA results for planview outlines extracted from the handaxes discussed in the text. The first principal component score (PC1) captures variation related to elongation, where the positive side of the axis represents forms that are more pointed (i.e. longer compared to their width), while the negative end of PC1 represents forms that are more ovate-like. The PC2 axis captured variance relating to tip and base shape, where the positive end represents ovular forms suggesting that tips and bases have been reduced in equal proportions. The negative end of PC2 represents forms that have been more intensely reduced in the tip proportion when compared to bases. Overall, the results of the planview PCA graph show significant overlap of the groups, regardless of their chronological affiliation within the Acheulian industry. This corroborates the result of past research, suggesting that planview shape does not discriminate handaxes from Early versus Late periods (McNabb and Cole 2015; Li et al. 2018; Caruana and Herries 2020). Figure S2 displays the results of a CVA analysis of planview outlines using the first seven principal component scores, which also shows significant overlap. Table S1 shows cross-validated results from this analysis, the classification accuracy is only 32.32% (Kappa= 0.302), further confirming the low discriminatory power of planview shape in correctly identifying Early from Late Acheulian handaxes.

The refinement of handaxes, defined as increasing planview symmetry and profile thinness, has been used to distinguish Early and Late Acheulian assemblages. However, recent studies have found that this is not a ubiquitous trend throughout the Acheulian industry. Yet, research suggests that Late Acheulian handaxes differ from earlier forms in the complexity and extent of thinning procedures. To test the discriminatory power of thinning in distinguishing Early and Late Acheulian handaxes, cross-sectional shapes are compared through geometric morphometric techniques. Results suggest that analysing cross-sectional outlines can identify variance relating to reduction processes, which differentiate handaxes from Early and Late Acheulian assemblages.

Late Acheulian handaxes in South Africa seem to increase in morphological variability when compared to earlier forms, which contradicts longstanding notions that these tools become more refined in symmetry and thinness through time. Recent investigations into the symmetry of Early and Late Acheulian handaxes in this region have found no significant differences in planview or profile perspectives. However, shifting analytical focus onto cross-sectional shapes provides valuable insight into morphological features that distinguish handaxes from Early from Late Acheulian periods. This is related to an increase in the intensity of thinning processes in the Late Acheulian, which potentially represents a global trend. The results of this study suggest that cross-sectional shape variation is a powerful discriminating factor when comparing handaxes from different periods of the Acheulian in South Africa, although testing larger datasets is required to confirm these findings.

The Cradle of Humankind World Heritage Site (CoHWHS), Gauteng Province, South Africa, preserves a remarkable palaeontological and archaeological record. Since the 1950s, stone tool assemblages have been excavated from well-known cave sites in this region, offering valuable insight into the behavioural and cognitive capabilities of early human ancestors. Modern research perspectives draw on a wealth of information to understand taphonomic, site formation, and technological aspects of these assemblages, and to interpret their significance in human evolution. At a point when high-resolution spatial data can be captured at sites in the Cradle region, reviewing the history of how current interpretive models of archaeological assemblages have developed is important for forging future directions in research. This history highlights the evolution of classification and analytical trends that have led to current multidisciplinary approaches to archaeology in southern Africa. It also illuminates the fact there has been a narrow focus on palaeoanthropological cave systems, with less information derived from archaeological assemblages found on the landscape above. As such, cave sites in the Cradle have provided perspectives on the Earlier Stone Age, while the Middle and Later Stone Ages have unfortunately been under-represented. Future directions in research should focus on increasing chronological, environmental and spatial resolution of the archaeological record in this region, which requires the collation of a wide array of data from cave and open-air sites.

The Cradle of Humankind World Heritage Site (CoHWHS), Gauteng Province, South Africa, preserves a remarkable palaeontological and archaeological record. Since the 1950s, stone tool assemblages have been excavated from well-known cave sites in this region, offering valuable insight into the behavioural and cognitive capabilities of early human ancestors. Modern research perspectives draw on a wealth of information to understand taphonomic, site formation, and technological aspects of these assemblages, and to interpret their significance in human evolution. At a point when high-resolution spatial data can be captured at sites in the Cradle region, reviewing the history of how current interpretive models of archaeological assemblages have developed is important for forging future directions in research. This history highlights the evolution of classification and analytical trends that have led to current multidisciplinary approaches to archaeology in southern Africa. It also illuminates the fact there has been a narrow focus on palaeoanthropological cave systems, with less information derived from archaeological assemblages found on the landscape above. As such, cave sites in the Cradle have provided perspectives on the Earlier Stone Age, while the Middle and Later Stone Ages have unfortunately been under-represented. Future directions in research should focus on increasing chronological, environmental and spatial resolution of the archaeological record in this region, which requires the collation of a wide array of data from cave and open-air sites.

Raw material selection and transport behaviors are important constraints on Oldowan lithic production strategies. While patterns in toolstone preference at Oldowan sites suggest that physical properties relating to fracture mechanics played a key role in determining selectivity, the durability of rock types was also a critical concern. However, few studies have examined the influence of raw material performance, during subsistence activities, on procurement strategies
in the Oldowan industry. Here we report on an actualistic butchery experiment used to measure attrition rates in quartz and quartzite flakes for the purpose of examining raw material selectivity in Oldowan sites from South Africa. Interpretations are focused on the two largest assemblages in this region, Sterkfontein Member 5 East and Swartkrans Member 1 and why over ninety percent of artefacts from these sites reflect a technological preference for the use of vein quartz.

Sterkfontein Member 5 East (Oldowan Infill) and Swartkrans Member 1 (Lower Bank) represent the largest concentrations of Oldowan artefacts in southern Africa, and yet they vary significantly in terms of raw material use and typological frequencies. While previous research has described these differences in detail , questions remain as to the cause and implications of this variability. To increase resolution on this matter, this study implements quantitative methods to investigate lithic production strategies at these sites. Results expand upon previous findings concerning differences in raw material use and knapping methods and how these patterns relate to core reduction and flake production. Explanations for these patterns focus on a dichot-omy between efficiency and expediency in lithic production, which differ from previous interpretations. As such, variability between these assemblages may relate more to mobility patterns in early hominins within this region and immediate needs for tool use.