This dissertation research examines variability in Natufian settlement and subsistence patterns. The Natufian is a late Epipaleolithic culture that flourished in the Levant during the late Pleistocene and early Holocene. There has been considerable research interest in the nature of Natufian settlement and subsistence over the last fifty years. This is due to the apparent pivotal position the Natufian holds in the developmental sequence from mobile hunting and gathering societies to sedentary village communities subsisting on domestic plants and animals. The research has two major components. The first aspect of the research focused on the range of activities and the nature of settlement intensity and duration during the Natufian occupation at one site, Beidha. The second aspect of the research entailed examining data for settlement and subsistence patterns throughout the Levant during this time period in order to gain insight into pan-Levantine variation. Research on the Beidha settlement has provided information on the range of chipped stone manufacturing processes carried out at the site, the nature of the tools produced, the range of activities that these tools imply, and the distribution of these elements between different areas of the site. The minimal evidence of spatial differences between provenience units is indicative of a site that was occupied as a relatively short-term camp site. The interpretation of the site as a short-term settlement, that was occupied repeatedly, fits with the lack of evidence for permanent features such as houses and storage facilities, as well as the lack of large groundstone objects and burials. Analysis of available data on Natufian settlement and subsistence has indicated that there are regional differences in settlement and subsistence patterns, most notably between steppe and desert sites versus sites of the forest and coast. In addition, there appear to be patterns in variability within the steppe and desert area with respect to the permanence of sites and the range of activities carried out in them. One set of sites is characterized by moderate settlement permanence and intensity, while the other set of sites is characterized by less permanent occupation and more specialized activity focused primarily on hunting.

The Anthropological Papers of the University of Arizona is a peer-reviewed monograph series sponsored by the School of Anthropology. Established in 1959, the series publishes archaeological and ethnographic papers that use contemporary method and theory to investigate problems of anthropological importance in the southwestern United States, Mexico, and related areas.

The research reported here arose from perceived lacunae regarding archaeological understanding of mesolithic settlement in eastern Scotland. Historically this area, for a number of reasons, has seen 1ittle archaeological research in comparison to the maritime west of the country, a bias that requires redressing. The characteristics, problems and potentials of available data are assembled for the first time and critically assessed. Discussion of methodologies appropriate to this material is developed, and small-scale fieldwork undertaken within this framework presented. Any introduction of a new range of data is, in part, a construction of that data, and the particular interpretative and thematic stresses of the thesis arise from the argument that narratives of gatherer-hunter communities in the past have objectified those groups, consequently hindering comprehension of them. To this end an approach to a social archaeology of the mesolithic is developed, stressing the importance of examining skills and routines that, through thei; extension in particular contexts, may have structured an agent's experience of landscapes in the past. In order to flesh out these arguments and introduce the material evidence in more detail, a series of overlapping case studies is developed exploring in turn, the relationships between mesolithic folk and woodlands, the significance of salmon fishing, the inhabitation of the coast, and stone tool procurement, production and discard. These varied narratives incorporate the results of a range of small-scale desktop projects and fieldwork designed to test the potential of this approach to a social archaeology of the period. Whilst these studies are at present fragmentary, it is contended that they demonstrate that accounts of gatherer-hunter communities in the east of Scotland can aspire to a meaningful level of engagement with human lives in the past. The project scholarship was funded by Historic Scotland.

The study of core reduction, or how little raw material was transformed into flakes which could be used for tools, is fundamental to the interpretation of most archaeological lithic assemblages. A distinction can be made in this regard between the "manner" of core reduction, or the mechanics of flake removal, and the "method" of its reduction, which focuses upon the sequence of core preparation and tool blank production. The emphasis in current lithic studies has been upon the former aspect of the core reduction process. Evidence for the method of flaking cores in prehistory has not been as widely or thoughtfully addressed and, consequently, a comprehensive theory and methodology for its study is lacking. This dissertation addresses itself to the problems of reconstructing core reduction sequences from archaeological assemblages of chipped stone. It introduces the theoretical background and associated methodology that is necessary to approach the study of the method of core reduction, without the aid of backfitting or the assumptions involved in replicative studies. This approach is based not only on the cores discarded after reduction but also on an interpretation of those features of the flakes that can inform upon their role and place in the core reduction sequence. The potential of the method is assessed through an analysis of blank production at several Middle Paleolithic sites in Northern Bosnia, Yugoslavia. Of these, the sites of Zobiste and Visoko Brdo form the basis of an interpretation of the core reduction strategies practiced in this area during the early Upper Pleistocene. This strategy is shown to be a result of the varying interaction between the nature of the raw material source, the intent of the knapper, and the principles of the lithic reduction process itself. The new data presented and the perspective achieved from the study of the core reduction sequence will be useful for future studies of the Middle Paleolithic in this area and in the comparison of these industries with other regions of the Old World.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. Johannesburg, 2014. / The Middle Stone Age rock shelter site Sibudu contains abundant evidence of pyrotechnology (the controlled use of fire) such as combustion features, ashy layers and lenses and burned bone. I study the use, reuse and discard of fire at Sibudu and the layers Brown/Grey mix and Brown under Yellow Ash 2(i) from the post-Howiesons Poort (post-HP) industry (c. 58 000 years BP) and the layers Pinkish Grey Sand and Pinkish Grey Sand 2 from the Howiesons Poort (HP) techno-tradition (c. 65 000 years BP) are selected for detailed studies. I describe properties of selected combustion features, such as size, form, number of hearth strata, contents (bone, charcoal and stone) and pH values. The post-HP combustion features generally contain a higher proportion of charcoal and are more acidic than the HP combustion features, suggesting post-depositional differences. Twenty actualistic experiments including 39 experimental fires constitute the largest component and contribution of this thesis. The experiments were conducted in two cycles; the first cycle of experiments burned (in different fires) one each of the wood taxa Casuarina equisetifolia, Dichrostachys cinerea, Eucalyptus globulus or Acacia erioloba and the second cycle of experiments burned Dichrostachys cinerea exclusively. Variables such as wood mass, topsoil horizon and number of sequential fires were carefully controlled. The surface and subsurface temperatures of the experimental fires were recorded. The experimental hearths were excavated using similar techniques as those used at Sibudu. Surface temperatures vary greatly even under similar environmental conditions, but subsurface temperatures are more predictable. Five kilograms of wood is sufficient to produce high temperatures for several hours, and slowly adding logs to a fire ensures even temperatures. Sibudu's hearths have a basal black layer with charcoal, but experimental fires do not. Instead they produced 2 – 5 distinct strata, and long-burning fires produced more strata than short-burning ones. Experimental ash dumps lack strata. Areal extent of a hearth is dependent on the wood mass burned, but its depth is dependent on the wood taxon burned. I present spatial analysis of charcoal, bone and knapped stone from the selected Sibudu layers using Kernel density estimation. Higher densities of bone, charcoal and stone are present in the post-HP layers than in the HP. Specialist studies of wood taxa, micromorphology and organic petrology are also included. No specific firewood taxa were preferred, but herbaceous plants were burned at 65 000 years ago, suggesting short fires. This thesis provides a foundation for future research on fire-related behaviour at Sibudu and other Middle Stone Age sites. For example, higher acidity was recorded in the post-HP than in the HP combustion features; this could be caused by different diagenesis and geoarchaeological studies need to be conducted to resolve the issue.. In addition, all the primarily deposited archaeological combustion features in this study were underlain by a basal black layer rich in charcoal, but such layers did not develop in the experimental combustion features and I cannot explain this difference. Thus it is necessary to conduct more studies of the variables affecting the formation and preservation of combustion features. One aspect to study experimentally is the range of attributes controlling surface temperatures of campfires. Amongst the experiments described here, experimental fires burning five kg wood produced maximum surface temperatures in the range of 132 - 848°C, which demonstrates that surface temperatures are unpredictable. The implication is that producing particular temperatures of a fire is not straight-forward and that prehistoric fire users needed planning abilities and a thorough knowledge of pyrotechnology. Both these skills require mental abilities like those of modern people and they also require a sophisticated understanding of environmental conditions.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. Johannesburg, 2016. / ABSTRACT Sibudu rock shelter was occupied by hunter gatherer groups of modern humans during the Middle Stone Age between 77 and 38 000 (38 ka) years ago. Sibudu is in the summer rainfall region and is situated about 15 km inland of the eastern coast of South Africa. The site includes technological industries such as Still Bay and Howiesons Poort that are distinguished by specific stone tool types. Woody taxa identified from archaeological charcoal in this site, indicate changes in vegetation and wood use. Archaeological charcoal specimens, specifically from hearths on three putative occupation floors, were analysed for behavioural information about wood selection and use. Two of the stratigraphic layers, BYA2 (i) and SPCA, are approximately 58 ka old and the MOD layer is approximately 49 ka old. All three layers contain tools from the post-Howiesons Poort industry. Identifications of woody taxa were made by means of a comparative charcoal reference collection and the InsideWood database. Charcoal was studied under stereo and reflective light microscopes. Woody taxa identified confirmed that the coldest, driest phase of occupation was at approximately 58 ka. Bushveld woods, including five Acacia types, have been identified in charcoal from MOD layer, ~49 ka. The vegetation mosaic was different from today’s coastal forest and savanna near Sibudu. The availability of natural wood around Sibudu, visible in the type of wood burnt in hearths, has changed since 58 ka and 49 ka for either climatic or anthropogenic reasons or for a combination of these. In all three layers, the wood bundles include tinder, fuel, and a selection of woody plants that today are known to have medicinal bark, leaves and wood. People may have selected wood for burning properties (for example, temperature, light and coal production) and for medicine. Woods suitable for making firesticks for starting fires by means of friction are present. Wood from Spirostachys africana (tambotie) occurs in one hearth in each of the three layers. Since Spirostachys africana is normally avoided in cooking fires as it is poisonous, this suggests that the wood was selected deliberately and burnt for insecticidal smoke or other medicinal purposes. Tarchonanthus parvicapitulatus (syn. T. camphoratus) charcoal occurred in layer SPCA. This adds to the previous evidence of the use of aromatic plants at Sibudu, where sedge bedding was topped with aromatic, insecticidal Cryptocarya woodii (river-quince) leaves (Wadley et al. 2011. Science 334 (6061): 1388-1391). Wood use is different between hearths and surrounding sediments and between occupations dating 58 ka and 49 ka. Keywords: phytomedicine, Sibudu, Spirostachys africana, Acacia, charcoal analysis, wood anatomy / LG2017

Ochre is found at many Middle Stone Age (MSA) sites and its use is often attributed by archaeologists to enhanced mental abilities and symbolism. However, the links between the visible uses of ochre, cognition and symbolism have not been clearly demonstrated. Here it is argued that by understanding ochre processing technology and some of the stages involved in using ochre, one can determine the skill, knowledge and cognitive abilities required to execute those activities. In order to understand the usetraces found on ochre, and to enable the identification of them, as well as the types of ochre used, experiments were first performed with geological ochre specimens. Ochre pieces collected from the Sibudu surrounds were used experimentally for a variety of grinding, scoring and rubbing activities. All use-traces created on the ochre piece during an activity were macro- and microscopically examined, recorded and compared. Experimental ochre pieces ground against a coarse or fine-grained slab develop parallel striations. Grinding results in significant changes to the surface shape of the ochre, and often results in faceted edges. Scoring can be performed with the intention to create powder, or to create a design. The incisions created from scoring often do not reach all the edges of the used surface and they regularly have frayed terminations. A frayed incision termination shows that the incision was created by multiple scoring strokes. When ochre is scored to manufacture powder the incisions that are generated are parallel groups of grooves with erratically oriented grooves as well. Grooves created through both grinding and scoring have microstriations within them and they show a range of profile shapes. The most common use-wear from rubbing ochre on soft materials is smoothing, edge rounding and polish. Microstriations and metallic lustre occasionally form during rubbing. The collection of utilised experimental ochre formed a comparative collection for the examination of the Sibudu ochre. The main body of this research comprises a study of the Middle Stone Age ochre assemblage from Sibudu, KwaZulu-Natal, South Africa. Sibudu has a large Middle Stone Age ochre assemblage of over 9000 pieces of ochre from layers dating between ~77 ka and ~37.6 ka. All pieces were examined to determine the types of ochre used and to inspect all use-traces present on the pieces. The assemblage comprises 5449 ochre pieces iv >8 mm, including 682 pieces with markings from use. The pre-Still Bay (~77 ka) and Howiesons Poort (~65–62 ka) layers have the highest percentage of utilised pieces. Bright-red ochre was preferentially selected for use throughout most of sequence. There is evidence of the preferential selection of specific types of ochre for use in the Sibudu assemblage. Shale and pieces with medium hardness values are common throughout the sequence. Grain sizes change through time – pieces with clayey grain sizes are favoured during the Still Bay and Howiesons Poort, whereas silty grain sizes are preferred in the younger Middle Stone Age occupations. High frequencies of bright-red amongst the utilised pieces, coupled with high frequencies of yellow or orange pieces with no evidence of use, suggest that colour choices were deliberate and not a product of postdepositional heating. Chemical analysis of a sample of utilised pieces indicates that they all contain iron, silicon, aluminium and calcium; many pieces contain hematite and some contain maghemite. Use-traces were divided into activity categories, based on experimental results. Combined grinding and rubbing is the most frequent activity for which ochre pieces were used. Grinding and rubbing use-wear also occurs independently on many pieces. Scored pieces are rare, but are more frequent in the pre-Still Bay (~77 ka) industry than elsewhere in the sequence. Some of the incisions appear to be deliberate engravings, and parallel lines and fan-like marks are the most often repeated patterns. Use-traces acquired during powder-producing activities predominate, implying a desire to create ochre powder. Powder-producing activities were mainly performed with bright-red pieces, while minimal scoring is mainly present on brown-red pieces. Pieces with mica inclusions are not common, but were frequently used for powder-producing activities. Once the activities performed with ochre were established, thought-and-action sequences, or cognigrams, were constructed. This helped establish the steps involved in each activity and the temporal and physical distance between the commencement of a task to its completion. Inferential sequences were constructed to establish the procedures and knowledge needed to complete an activity, thereby establishing the cognitive prerequisites. Cognitive interpretations are made using the concept of enhanced executive functions of the brain. The construction of the inferential thoughtv and-action sequences showed that the various ways that ochre was used have different cognitive requirements. Powder-production alone is not an indicator of complex cognitive processes, although some planning, foresight and knowledge of materials is required. Some of the powder was used in the creation of hafting adhesives, which is a cognitively demanding process requiring attention-switching ability, response inhibition and abstract thought. Grinding ochre and then rubbing the piece on a soft material for the direct transfer of powder does require some complex mental abilities, such as multi-tasking and switching attention. Scoring a piece of ochre with a sharp tool does not necessitate enhanced executive functions, but some engravings demonstrate foresight, intentionality and an awareness of space and symmetry that may demonstrate abstract thought. This research provides a complete description of the Middle Stone Age ochre assemblage at Sibudu, and establishes the way that ochre was used at the site. This contributes to the debate on the advent of enhanced behaviours in the past by providing insight into the cognitive abilities required by the ochre users. It offers a method of analysing ochre use in the past by drawing on cognitive theory and the visible applications of ochre. Both simple and complex cognitive abilities were required for ochre activities at Sibudu. The requirement for cognitively complex abilities in some of the ochre-related activities at Sibudu suggests that the people living there during the MSA had advanced mental capabilities like modern humans living today. This research shows how ochre use can be employed as a proxy for cognitive capabilities, and can therefore shed light on the evolution of the modern mind.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. Johannesburg, 2016. / This study is within the context of the Howiesons Poort Industry of the Middle Stone Age. This is a dynamic period of increasing behavioural and material complexity. In the lithic assembles, this can be seen in a strong bias towards the selection of high-quality fine-grained rocks. This has often been interpreted as evidence for long distance travel, reciprocal exchange, or even increased mobility. This study aims to determine what influence the mechanical properties of rock types exerted on the Middle Stone Age assemblages at Sibudu. This requires a consideration of the distribution of rock types across the landscape surrounding Sibudu Cave. The study was limited to hornfels and dolerite as these rock types dominate the Sibudu assemblage, and quartz and quartzite that were sometimes used at the site are exceedingly difficult to collect in large enough sample sizes to conduct experiments. It was important to carry out tests on the mechanical properties of hornfels and dolerite and to characterise them. Hardness, roughness, elasticity and brittleness dictate the ease of knapping as well as the durability of flaked tools and these rock properties can be measured by the mechanical tests described here. To understand how these properties affect the assemblage in practice, dolerite and hornfels flakes were produced and used experimentally for cutting and scraping leather. The edge damage produced was compared. Finally, preliminary analysis was undertaken of square C4, layer PGS, which forms part of the oldest Howiesons Poort layer at Sibudu. The information and insights gained from the mechanical tests and experimental work were used to interpret the role of mechanical properties for the archaeological sample of PGS. The results show that all rocks used at Sibudu are local, and do not support any models of long distance travel/trade, reciprocal exchange, or models of increased mobility. Dolerite and hornfels form the bulk of the assemblage at all times (except briefly in the post-Howiesons Poort), and there is a bias towards the selection of fine-grained rocks during the Howiesons Poort. Dolerite can be characterised as hard, tough, elastic, and rough, while hornfels is hard, brittle, and fine-grained. These properties affect knapping and the qualities of a tool’s edge. The properties of hornfels allow for knapping accuracy and predictability, and it is better suited to blade production and cutting. However, tool edges are not robust. Dolerite is not as easy to knap, but produces tools with a robust edge that are particularly suited to scraping. Each rock type appears to have fulfilled a different function at Sibudu. Most rock studies geochemically source rocks, establish models of rock procurement or show trends in rock selection for artefact classes. Mechanical studies of rocks have typically formed part of heat treatment debates (Brown et al 2009; Domanski & Webb 1992, 1994; Webb & Domanski 2008). Through the combined approach of mechanical testing, experimental knapping and tool use, and lithic analysis, this research provides a context for possible rock procurement choices at a time in the past when many African sites reveal a changing pattern of rock selection. / LG2017

A Masters Dissertation submitted to the faculty of Science, University of the Witwatersrand in fulfillment of the requirements for the Masters` degree. Johannesburg 2016. / The Middle Stone Age (MSA) levels (c. 100 – 72 ka) (thousand years) at Blombos Cave, South Africa, contain buried remnants of human existence and their activities (these include ochre-stained shell beads, processed ochre pieces, bone tools, stone tools, and ochre processing toolkits). A preliminary study on the occupation layers at Blombos Cave was done in order to chemically and physically characterise the sediment deposits. The interpretation of the information preserved in these sediments, as a result of human activities, was achieved by using the following characterisation techniques: FT-IR, ICP-OES/MS, CHNS analysis as well as pH and EC measurements. Fluctuations in the bio-essential trace elements (K, P, Mg, Ca, S, Fe, Al, Cr, Na, Cu and C) and in the markers of in situ burning (P, K, C, Ca and Mg) through the MSA levels suggested a non-contiunous anthropogenic occupation of the cave. Further, the dark coloured layers were indicative of periods when heavy human activities, such as fire use, were encoutered. These layers were characterised by high levels of P, K, C, Ca and Mg, also elevated pH and EC levels were observed. The calcite origin for these sediment deposits was confirmed to be pyrogenic. The layers found to contain large bone, ochre and shell densities (with no evidence of fire use) had high levels of Fe, P high Ca and low EC. The calcite origin for these sediment deposits was geogenic. Biogenic calcite was only observed in the more recent CA and CC layers. Substantial amounts of ochre assemblages were recovered from the MSA levels at Blombos Cave. Among these was an ochre processing toolkit recovered from the CP layer (c.100 ka). Ochre is abundant in African archaeological sites after 165 ka and is likely to have played a symbolic role in the lives of prehistoric people. It contains an iron-oxide mineral as well as accessory minerals. When mixed with a binder (such as fat or water) ochre can be used as a pigment. A large number of fragmented bone remains were found in the M1 and M2 levels. It is suggested that the majority of the fragmentation occurred while the bones were in a fresh state; implying marrow extraction by the site occupants. The exploitation of bone marrow was crucial in order to extract the fat and use it as a binder during the pigment production. Some of the broken and marrow-extracted bones were heated and used as fuel during seasons when wood was scarce while the other bones were deliberately engraved for symbolic intent. The characterisation of the ochre processing toolkit was achieved by mineralogical analysis and elemental fingerprinting. FT-IR analysis revealed that the general matrix of the ochre samples comprised of hematite (Fe2O3) or goethite (α-FeO(OH)) as the main chromophores and clay minerals (such as kaolinite (Al2Si2O5(OH)4), muscovite [(KF)2(Al2O3)3(SiO2)6(H2O)] and illite [K(Al4Si2O9(OH)3)]), calcite (CaCO3) and quartz (SiO2) as the main accessory minerals. PXRD analysis confirmed Fe2O3, Al2Si2O5 (OH)4, [K(Al4Si2O9(OH)3)] and SiO2 to be the predominant mineral phases in the ochre, implying this specific type of ochre was preferred during the production of the pigment. Fe2O3 contributed the red hue and the aluminosilicates their clayey properties making them good extenders of the pigment. Multivariate statistics and Fe ratios made it possible to identify elements important for differentiating the ochre recovered from the CP layer. Analysis of variance (ANOVA one-way) showed a statistically significant difference between the ochre residues in terms of trace elemental profiling. The variance suggested different geological origins for the ochre. FT-IR was used as a screening technique for any organic residues associated with the toolkit and GC-MS was used to identify the preserved organic residues. These were mostly lipids and terpenes. The identified bio-molecular markers; stearic acid and dehydroabietic acid were exploited to give insight on the origin of the residues. Stearic acid suggested the use of animal fat while dehydroabietic acid implied the use of a resin (potentially as a binder in the pigment). The characterisation investigations revealed that the prehistoric populations at Blombos Cave specifically sourced hematite and aluminosilicate- containing ochre pigments in order to fulfil their social and cultural demands. The MSA site occupants’ chemical understanding of these materials suggested they were technologically advanced. Keywords: Blombos Cave, MSA levels, ochre, mineralogical analysis and elemental fingerprinting. / LG2017

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. 2016 / The analysis of faunal remains from archaeological sites can reveal much about past behaviour, palaeoenvironments and bone technology. This is especially pertinent for the Middle Stone Age (MSA): a period that corresponds to both the behavioural and anatomical development of Homo sapiens. In this thesis, I examine the faunal remains from Blombos Cave (BBC) and Klipdrift Shelter (KDS), two significant MSA sites about 45 km apart along the present-day southern Cape coast. The focus of the analyses is on the Still Bay (SB) layers (c. 75 – 68 thousand years ago [ka]) at BBC and the Howiesons Poort (HP) layers (c. 65 – 59 ka) at KDS. The aim of this thesis is to explore subsistence behaviour and environmental conditions during the SB and HP in the fynbos region of the southern Cape. I also investigate the effects of trampling on bone and whether trampling and other types of taphonomic modification can be used to infer occupational intensity in cave and rock shelter sites. Taphonomic data indicate that significant differences exist between the SB and HP faunal assemblages at BBC and KDS, respectively. Carnivores and scavengers had a greater effect on the SB than on the KDS assemblage. Furthermore, subsistence activities at KDS focused on marrow extraction while filleting was probably an important strategy at BBC. Taxonomic analyses indicate that the representative fauna from both sites is generally consistent with what is expected in the fynbos biome but with some noticeable differences. Grazers, for example, are significantly more prevalent at KDS than BBC while seal is more common at BBC. Other zooarchaeological data also show differences in prey selection strategies. Diet breadth is more extensive at KDS than at BBC, although bovid mortality profiles at BBC are more juvenile-dominated. I propose that many of these contrasting patterns imply that subsistence intensification is more evident in the HP layers at KDS than in the SB at BBC. Moreover, the taphonomic data suggest that the SB at BBC was a low-intensity, sporadically occupied period in contrast to the high-intensity occupations, particularly during the middle layers of the HP at KDS. The results of the KDS analysis also indicate a shift in environmental conditions during the HP and show links between prey selection, the environment and occupational intensity during this period. Trampling experiments indicate that trampling can generally be distinguished from butcherymarks, although bioturbation in shelly deposits can sometimes result in marks that mimic cutiii marks. More importantly, experiments show that pitting and abrasion are a more significant indicator of trampling than lines that resemble cut-marks. Based on these experiments, I argue that trampling modification can be used to infer occupational intensity at archaeological sites. Trampling marks in the KDS assemblage, for example, correspond well with the high occupational periods as indicated by other taphonomic data. Trampling and taphonomic data support the notion that BBC was a low-intensity, sporadically occupied site during the SB. By incorporating taxonomic, taphonomic and novel methods of skeletal-part analyses, this study contributes to our knowledge of human subsistence and palaeoenvironments during the SB and HP in the southern Cape. This thesis strengthens and adds to other research that has demonstrated variability in subsistence behaviour during the MSA. / LG2017