The Makgadikgadi Basin in central Botswana is a key region for understanding Stone Age archaeology and human evolution. However, paleoenvironmental reconstructions have so far offered only fragmentary insights into past hydroclimatic and environmental conditions. We present the first environmental record from Sowa Pan, paleolake Makgadikgadi, covering the last 82-20ka (Marine Isotope Stages 5-2). Sedimentological and biomarker lipid analyses reveal four distinct lake phases, reflecting major hydrological shifts and sediment provenance. From 82 to 75ka, the lake was shallow, with lowest lake levels temporally coinciding with Middle Stone Age archaeological material (silcrete lithics, mainly unifacial and bifacial points) on the western lakebed. Between 75 -58ka, the lake refilled, likely fed by river inflow sourced from the Angolan highlands. A second low stand likely occurred around 58-56ka. From 52 to 37ka, geochemical proxies indicating a change in lake inflow with less influence of the Kwando catchment. During this time, the lake had high primary productivity and biomarkers for terrestrial herbivores are recorded in the sediments. After an accumulatory hiatus, the youngest unit (ca. 21ka) marks the formation of the Sowa Spit, characterized by high sand input and minimal organic preservation. This record refines the chronology of lake-level changes and reveals that desiccation phases provided ecologically viable landscapes for human activity. It also identifies shifting sediment and nutrient sources that shaped lake dynamics during the late Pleistocene.

Framing the Middle Stone Age of bed V at Mumba Rock Shelter, Tanzania, using electron spin resonance and uranium series

The Middle Awash study area of Ethiopia's Afar Rift features a composite stratigraphic thickness of >1 km. Near the top of this succession lie sediments of the lower Halibee member, comprising the Faro Daba and Chai Baro beds. The former are radioisotopically dated to ~100,000 y in age and contain abundant fossils and associated lithic artifacts representing the Middle Stone Age (MSA). Geological, paleontological, and archaeological datasets recovered from these sediments enlarge a sparse later Pleistocene record of African human evolution, a time before anatomically modern populations of our species expanded into Eurasia. The Faro Daba occurrences comprise the richest, least disturbed, and most spatially extensive of many open-air MSA-bearing localities in the study area and beyond. Protected atop a resistant underlying conglomerate, the largely horizontal outcrops of the soft, eroding fossiliferous Faro Daba sediments provide spatially extensive access to in situ assemblages of artifacts and fossils. Sedimentology, faunal composition, and combustion features are consistent with a wooded depositional environment with seasonal flooding distant from the main river channel. Archaeological and paleontological assemblages indicate minimal postdepositional disturbance of primary lithic tool manufacture and discard during ephemeral human occupations on this floodplain. Among the recovered fossils are three partial human skeletons with taphonomic evidence of different postmortem pathways.

The selection and acquisition of suitable raw material constitute the first steps in stone tool technology. Previous ethnographical and archaeological research suggests that hominins in the Pleistocene primarily collected their stone materials while carrying out other activities. Direct provisioning for this purpose alone remains an outlier and is rarely demonstrated. Archaeological excavations coupled with multidisciplinary analyses at Jojosi in South Africa demonstrate that early modern humans undertook specific, repeated visits to a raw material source over tens of thousands of years for the exclusive purpose of obtaining hornfels. This rare, stratified, open-air locality features uniquely preserved lithic assemblages with abundant refits dating from ~220 ka to ~110 ka for the reduction and export of a single tool stone. The scope of these knapping activities is underscored by millions of Middle Stone Age hornfels artefacts paving the modern landscape. The consistent, specialised procurement of a single raw material at Jojosi already during the Middle Pleistocene challenges the standard model of embedded procurement for this period. These findings further show that key capacities of Homo sapiens, including increased long-term planning and behavioural plasticity in the interaction with the material world, emerged early in their evolutionary history.

A Middle Stone Age occupation identified at Baden-Baden in the grasslands of the Free State, South Africa.

Wind(ow) of change: The end of the Middle Stone Age and the beginning of the Later Stone Age at Umhlatuzana rockshelter showcasing concurrent technological and techno-economic shifts

The use of Nassariidea shells as personal ornaments is attested to an increasing number of Middle Stone Age (MSA) archaeological sites in northern and southern Africa. The chronological extent of this behavior is constantly moving back in time; currently, the oldest evidence has been identified at the Bizmoune cave site in Morocco back to the MIS 6. Although these evidences make it possible to refine the spatial and temporal distribution of this behavior, shell beads remain rare in Middle Stone Age assemblages and are generally composed of several beads, or at best dozens, for each of these sites. This restricts our understanding of the behaviors specifically related to the collection, selection and preparation phases of shells, and potentially limits our understanding of their use. In this article, we studied shell beads from MSA layer US 8 from the coastal archaeological site of El Mnasra Cave (Rabat-Témara, Morocco). This collection corresponds to the largest MSA shell bead assemblage in Africa (272 Tritia cf. gibbosula, 6 Tritia corniculum and 3 Columbella rustica in US 8 with 154 of them showing smoothing of the perforation edge, facet of abrasion, or traces of pigment). The shell bead assemblage of El Mnasra presents features previously observed at other MSA sites, connecting it to a North African cultural context; however, the size of the El Mnasra shell bead assemblage, and the presence of shell sources near the site, allows us to identify specific features that could be related to particular modes of use as ornaments. These specific features include the prevalence of un-perforated shells, some of which show use-wear, that could have been fixed on items without having been perforated. These results provide new insights into the wide range of variants and originalities of shell bead uses over a relatively “short” chronological phase, between 115 and 94 ka and can be correlated with the multistep evolutionary scenario proposed for South Africa. The archaeological documentation presented here shows that El Mnasra Cave provides a significant contribution to the study of culturing the Palaeolithic body in North Africa.

ABSTRACT Levallois technology is a particularly skilful and cognitively intensive lithic technology where preparatory core shaping is aimed at producing predetermined flake blanks. It has significant implications for hominin evolution and dispersals, yet remains notoriously difficult to define relative to other prepared cores. Significant variability in prepared core techno-morphology is often rendered into a handful of rarely agreed upon discrete categories. We ask whether existing classifications of cores are immutable, discrete groups, or whether there are unbroken continua of prepared core variability. With a bespoke set of 3D computational methods, we directly quantify the criteria used to define Levallois technology. Applied to a diverse sample of prepared cores from five African MSA sites, these methods establish continua of prepared core preparation, recurrence and exhaustion. Quantifying this continuous variability in how blanks are removed from cores can help holistically address past hominin technological decision making, investment and flexibility.

Homo sapiens emerged in Africa around 300 − 200 thousand years ago (ka). Although the earliest H. sapiens fossils are associated with the Middle Stone Age (MSA), lithic technologies considered diagnostic of the MSA have been found alongside Acheulian technology in eastern Africa and the interior of southern Africa by ~ 500 − 400 ka, suggesting a deep evolutionary history of our species in these regions. The southern coastal plain of South Africa, geographically separated from the interior by the Cape Fold Belt and Great Escarpment, has one of the best documented records of the MSA in Africa; however, only a single site is older than 125 ka and little is known about the origins of the MSA in this region. Here, we report a stratified sequence of Acheulian to MSA lithic assemblages from the open-air site of Amanzi Springs covering the period between ~ 379 to 95 ka. We show that the MSA emerged around 230 ± 18 ka, significantly earlier than previously documented along the southern coast. The pattern of technological change also differs to the interior, with no diagnostic MSA elements found in the late Acheulian, although persistent methods of flake production indicate a gradual transition and continuity into the MSA. The relatively late emergence of the MSA along the southern coast highlights the variable and complex nature of demographic and behavioural change during this period, with regionally distinct technological trajectories extending into the Middle Pleistocene in southern Africa.

Prepared core technology is a defining feature of the Middle Stone Age (MSA) in Africa, yet classic preferential Levallois cores are rare in many southern African assemblages. The recent identification of Nubian Levallois cores – a well-defined prepared core strategy – at sites in the South African interior raises broader questions about how current terminologies may be shaping interpretations of Levallois in southern Africa. The open-air site of Tweefontein presents a large Nubian core assemblage together with a range of other prepared core forms, providing a unique opportunity to assess how Nubian cores fit within the broader spectrum of prepared core variability. Using 3D geometric morphometrics and other quantitative methods, we examine variation in core morphology and technology and evaluate the distinctiveness of specific core types. Our results indicate that Nubian Levallois at Tweefontein represents one extreme of a continuum of preferential Levallois reduction, which occurs alongside distinct radial and opposed-platform reduction strategies. While the southern African Nubian cores are geographically, temporally and culturally separated from other instances of this technology, our approach demonstrates the broader potential of 3D analysis in moving beyond typological categories to capture local technological adaptations. This is especially relevant to the study of prepared cores in southern Africa, where currently heterogeneous terminology hinders meaningful comparison between assemblages.

ABSTRACT This paper reports on a new archaeological research project in the Stormberg region of South Africa’s Eastern Cape Province that began in 2018. After four prospecting surveys, an archaeological site — Marshill Rockshelter — was selected that could offer a long diachronic chronostratigraphic framework from the late Pleistocene to the Holocene in order to understand the deep past of this region and the long-term occupation of a mountain environment in the piedmont of the Maloti-Drakensberg Mountains. This paper presents the first archaeological analyses of this site’s stratigraphy, geochronology, faunal pollen, ceramic and lithic assemblages and rock art. Marshill has a long and complex stratigraphic sequence with three clusters of dates: one in the last 2000 years, a second around the terminal Pleistocene (10–15 kya) and a third at >20 kya. The archaeological materials found are in accordance with these initial ages and provide unique snapshots into the region’s past. A Middle Stone Age archaeological assemblage shows similarities with other sites in the Grassland, Savanna and Fynbos Biomes of southern Africa, although with the lithic evidence currently available it is difficult to establish stronger comparisons. Holocene layers, in contrast, produced microlithic Later Stone Age assemblages oriented towards the production of small flakes, most likely from unifacial cores. Initial faunal and pollen analyses accord with historical records in indicating an open grassland environment with thicketed valleys and gorges. The occurrence of domestic livestock in the upper Holocene layers suggests links to pastoralist groups and a herding economy. Preliminary ceramic analyses show several strategies for surface treatment, firing and modelling, while the textural groups defined document the use of fibre temper. Finally, the site’s rock paintings were created in a shaded and hard-edge technique with both ‘classic’ and more recent subject matter. In sum, the archaeological record from Marshill Shelter demonstrates a long but intermittent series of human occupations extending from the deep past to the historical period.

ABSTRACT Southern African Middle Stone Age research has largely focused on the region’s coastal caves and rock-shelters, while its semi-arid interior remains under-investigated due to the assumption of limited stratified deposits. Nevertheless, as recent studies prove, this region also possesses potential for contributing to Palaeolithic archaeology. This study presents surveys and excavations in southern Botswana that documented Pleistocene quartzite lithic artefacts near quartzite outcrops and pans (palaeolakes). Excavations at Maralaleng Pan and Itireleng yielded stone artefacts on and below the surface, providing new information about Pleistocene habitation in this region. This paper presents the abundance, condition and spatial distribution of lithic artefacts within an open-air context, connecting them to potential occupational windows within this landscape. Generally, damaged and weathered artefacts are scattered at quartzite outcrops close to pans. Flakes are the most abundant artefact type in every unit, while cores are less common and mostly present at the raw material outcrops. A likely scenario is that the artefacts were manufactured at the raw material outcrops. Various spots of lithic concentrations are influenced by natural processes and are also indicators of occupational windows in a changing landscape. This study aims to contribute to a deeper understanding of human development in the southern Kalahari during the Pleistocene.

This article presents the first quantitative geometric and spatial analysis of engraved ostrich eggshell (EOES) fragments from the Howiesons Poort (HP) technocomplex of the African late Middle Stone Age (MSA), to evaluate whether the EOES demonstrates genuine formal structuring and visuo-spatial organization. By considering their 'non-accidental properties'-such as curvature, parallelism, and co-termination-which remain consistent across different viewpoints, as well as their metric properties, including angular inclinations, based on empirical thresholds, we show that the HP dataset systematically employs salient geometric features. These features are combined and embedded through complex cognitive operations, including the iteration and alignment of parallel lines, rotation of lines generating intersections with variable angular openings, and translation of specific elements nested within organized spatial layouts. These engravings therefore constitute an early material expression of complex graphic representation, attesting to a species-specific human capacity for organizing geometric thought. Overall, the patterns reflect a system of rules through which Homo sapiens in the HP organized visual forms, revealing the cognitive foundations of structured graphic behavior.

Cooperation is fundamental to human societies. While several basic theoretical mechanisms underlying its evolution have been established, research addressing more realistic settings remains underdeveloped. Drawing on the hypothesis that intensified environmental fluctuations influenced early behavioral evolution in humans during the Middle Stone Age in Africa, we examine the effects of environmental variability and human mobility on the evolution of cooperation. In our model, the variability is represented by randomly moving resource-rich spots across a two-dimensional space, and the mobility is represented by resource-seeking migration of agents. The agents interact cooperatively or competitively for resources while adopting their behavioral strategies from more successful neighbors. Through extensive simulations of this model, we reveal three key findings: (i) with sufficient agent mobility, even modest environmental variability promotes cooperation, but further variability does not enhance cooperation; (ii) with any level of environmental variability, agent mobility promotes cooperation; and (iii) these effects occur because the joint effect of environmental variability and agent mobility disrupts defector groups in resource-rich areas, forming cooperator groups at those sites. Although previous studies examined environmental variability and mobility separately, to the best of our knowledge, this is the first study to analyze their joint effects on the evolution of cooperation. These findings suggest that environmental variability can promote cooperative group formation without enhanced cognitive abilities, providing new insights into the evolution of human cooperation and, by extension, sociality. • Modeling environmental variability as a driver of Middle Stone Age human cooperation. • Environmental variability and agent mobility can promote the evolution of cooperation. • Environmental variability can disrupt stable defector groups in resource-rich areas. • Agent interactions and mobility can foster cooperative group formation. • Cooperation can evolve without requiring enhanced cognitive abilities.

Abstract This paper presents excavation results from Nyabusora, northern Tanzania, conducted by M. Posnansky and W.W. Bishop (1959) and M. Posnansky (1961). Only preliminary reports have previously been published. It synthesises the site’s history, incorporating previously unpublished analyses and information from Posnansky’s original field notes, and presents new 2014 field survey results and new archival research. Nyabusora holds particular significance as the only Early to Middle Stone Age (ESA/MSA) site in the region to have yielded both lithic and faunal remains, which gain new relevance in light of recent developments in ESA/MSA archaeology in eastern Africa. Nyabusora’s ‘Sangoan’ lithic assemblage is now largely decontextualised and associated finds have been lost, so this study presents the only available lithic and faunal analyses, alongside interpretations of the stratigraphic sequence and site. Such stratified assemblages are exceptionally rare and are generally attributed to the Middle Pleistocene. This research enhances understanding of Plio-Pleistocene landscape evolution in the Kagera River and western Lake Victoria-Nyanza Basin. It contributes important new data on ESA/MSA lithic variability and, via ongoing investigations by Basell within the Kagera catchment, offers huge potential for clarifying Middle Pleistocene palaeoenvironments.

Abstract Several syn and post-depositional processes are responsible for different degrees of site and assemblage formation and disturbance. Understanding the processes that lead to archaeological site formation is essential for interpreting past human activities, settlement patterns, and occupation dynamics. Among these processes, water flow is commonly recognized as a major factor influencing site and assemblage formation and integrity. Lithic taphonomy can add valuable information to the understanding of these natural processes. To better evaluate how different flint types react to both mechanical and chemical stresses, we carried out controlled and sequential experiments on knapped flint artifact replicas. These were realized on two different flint varieties, recognized in the archaeological assemblage of a Middle Stone Age open-air site located at Wadi Lazalim (Southern Tunisia). Our methodological framework is based on a sequential tumbling experiment that simulates the remobilization of artifacts by water. The resulting post-depositional surface modifications (PDSM) were observed and recorded. In addition, a sample of flint replicas underwent chemical alteration in a controlled environment, allowing us to document and measure both patina formation and changes in surface roughness. These experimental results were then compared with a sample of archaeological materials, providing the basis for a preliminary and exploratory hypothesis concerning the taphonomic processes at the site. Here, we observed that most of the alterations appear to be linked to patina formation, whose variable degree and kind are more strongly correlated with differences in raw material texture rather than to the co-occurrence of artefacts from different chronological spans.

Abstract An enduring challenge for the human evolutionary sciences is to integrate the palaeoanthropological record of human evolution and speciation with the archaeological record of change and differentiation in hominin lifeways. The simplest hypothesis, and therefore an attractive hypothesis, is that change is made possible by, and reflects, evolutionary change in the capacity of individual humans. The very long-term trend of increasing diversity and sophistication of technical and social lifeways (albeit with noise and periods of stasis) reflects long-term trends of increasing cognitive capacity linked to bipedality, followed by body size increase, encephalization and slow life history. We suggest instead that the long-term trend sees a gradual decoupling of human lifeways from the intrinsic capacities of individual people. We develop this view through an analysis of the Middle Stone Age and behavioural modernity, arguing that these depend on mosaics of social and individual factors, none clearly connected to specific evolved changes in individual humans.

Abstract The study of raw materials used for making stone tools allows archaeologists to gain insight into the behaviours of ancient people. Raw materials possess different properties, which influence the ability of knappers to flake, shape and use a stone tool. Measuring these properties experimentally requires specialised laboratories and measuring processes that are not straightforward. This makes such analyses a cost-intensive and often inaccessible way to investigate the past. Here, we present an alternative way to evaluate two mechanical properties, stiffness and hardness, using a single indentation test. We also test whether these two isolated measures allow making predictions on raw material selection and tool shape. We analysed tool-stones from the Middle Stone Age site Sibhudu on South Africa’s eastern seaboard. The site has yielded a rich assemblage of tools from different raw materials. We found that a single indentation test allows measuring hardness and stiffness reliably, simplifying the measurement protocol of raw material studies. We also observed weak correlations between those isolated properties and the shape of the finished tools. This has implications for future studies of Stone Age raw materials, proposing a simplified testing protocol. We also discuss the role of stiffness in stone knapping.

Humans are unique in their ability to build complex social networks that foster cooperation, knowledge sharing and innovation. Evidence from the African Middle Stone Age provides some of the earliest signs of these connections, alongside increasingly sophisticated behaviours. Archaeologists study past social interactions through various proxies, with stone tools playing a central role. Yet the extent to which stone tools reliably reflect cultural transmission and connectivity remains debated. Similarities in toolmaking can indicate knowledge exchange and social ties, but they may also result from convergent evolution, whereby different groups independently arrive at comparable solutions to similar challenges. Recent research from southern Africa and beyond shows that applying middle-range theories and integrating contextual data help distinguish cultural transmission from convergence. This approach sheds new light on how knowledge and practices spread in early human societies, revealing the deep roots of cooperation and collaboration that continue to shape human societies today.

The Quaternary landscapes, chronostratigraphy, and paleoenvironments of the Chalbi Desert, Kenya