Citation
The Subsistence strategies of middle and later stone age foragers during interstadial / glacial transitions at Knysna, South Africa

Material Information

Title:
The Subsistence strategies of middle and later stone age foragers during interstadial / glacial transitions at Knysna, South Africa
Creator:
Keller, Hannah May
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( Master of arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Anthropology, CU Denver
Degree Disciplines:
Anthropology
Committee Chair:
Hodgkins, Jamie M.
Committee Members:
Beekman, Christopher
Cleghorn, Naomi
Warrener, Anna

Notes

Abstract:
Knysna Eastern Heads Cave 1 (KEH-1) demonstrates varying human occupation over the course of multiple ocean transgressions and regressions in the late Middle Stone Age and early Later Stone Age (46,000 to 18,000 Cal BP). The position of the sites on the shallow South African coastal shelf offered occupants potential access to riverine, coastal, and terrestrial resources, as high and low sea stands drastically altered the location of these resources. It has been hypothesized that the dearth of sites dated to MIS 2 on the south coast, a time when the ocean was 75-90 km distant, is due to a preference for dense and high protein marine resources, drawing early modern humans out onto the now-submerged Paleo Agulhas plain. Possible explanations for the recurring occupation at KEH-1 during this time are 1) that use of marine foods was only one of several subsistence strategies employed by early humans; 2) the intersection of terrestrial and riverine resources around Knysna rivaled coastal resources; and 3) competition and territoriality caused by dense marine resources on the coast forced groups into the interior. Zooarchaeological analysis of faunal remains deposited during the MIS 3-2 transition at KEH-1 examined variation in the intensity of long bone marrow processing as a proxy for changing nutritional stress. Approximately 2400 faunal specimens were analyzed across the sequence. Although the degree of post depositional fragmentation varies throughout the sequence, the high rate of long bone fragmentation (84% with an incomplete circumference) at KEH-1 is most likely caused by human processing, with minimal carnivore activity. The higher frequencies of green fractures (54%) when the coastline is distant are statistically significantly different (Χ=2804, DF=16, P=0) from the frequencies of green bone fractures when the coast is near (31 and 43%). These results suggest that the long bones were processed more intensely during glacial periods, despite the availability of diverse terrestrial resources.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Hannah May Keller. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
THE SUBSISTENCE STRATEGIES OF MIDDLE AND LATER STONE AGE
FORAGERS DURING INTERSTADIAL/GLACIAL TRANSITIONS AT KNYSNA,
SOUTH AFRICA
by
HANNAH MAY KELLER
B.A., University of Texas at Arlington
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment Of the requirements for the degree of Master of Arts Anthropology Program
2019


©2019
HANNAH MAY KELLER ALL RIGHTS RESERVED
1


This thesis for the Master of Arts Anthropology degree by Hannah May Keller Has been approved for the Anthropology Program
By
Jamie M. Hodgkins, Chair Christopher Beekman Naomi Cleghom Anna Warrener
Date: May 18th, 2019
11


Keller, Hannah May. (M.A., Anthropology Program)
The Subsistence strategies of Middle and Later Stone Age foragers during Interstadial/Glacial transitions in Knysna, South Africa
Thesis directed by Assistant Professor Jamie M. Hodgkins
ABSTRACT
Knysna Eastern Heads Cave 1 (KEH-1) demonstrates varying human occupation over the course of multiple ocean transgressions and regressions in the late Middle Stone Age and early Later Stone Age (46,000 to 18,000 Cal BP). The position of the sites on the shallow South African coastal shelf offered occupants potential access to riverine, coastal, and terrestrial resources, as high and low sea stands drastically altered the location of these resources. It has been hypothesized that the dearth of sites dated to MIS 2 on the south coast, a time when the ocean was 75-90 km distant, is due to a preference for dense and high protein marine resources, drawing early modem humans out onto the now-submerged Paleo-Agulhas plain. Possible explanations for the recurring occupation at KEH-1 during this time are 1) that use of marine foods was only one of several subsistence strategies employed by early humans; 2) the intersection of terrestrial and riverine resources around Knysna rivaled coastal resources; and 3) competition and territoriality caused by dense marine resources on the coast forced groups into the interior. Zooarchaeological analysis of faunal remains deposited during the MIS 3-2 transition at KEH-1 examined variation in the intensity of long bone marrow processing as a proxy for changing nutritional stress. Approximately 2400 faunal specimens were analyzed across the sequence. Although the degree of post-depositional fragmentation varies throughout the sequence, the high rate of long bone


fragmentation (84% with an incomplete circumference) at KEH-1 is most likely caused by human processing, with minimal carnivore activity. The higher frequencies of green fractures (54%) when the coastline is distant are statistically significantly different (X=2804, DF=16, P=0) from the frequencies of green bone fractures when the coast is near (31 and 43%).
These results suggest that the long bones were processed more intensely during glacial periods, despite the availability of diverse terrestrial resources.
The form and contract of this publication are approved. I recommend its publication.
Approved: Jamie M. Hodgkins
IV


ACKNOWLEDGMENTS
Many people supported my crazy journey towards finishing this thesis. Foremost among them is my advisor and mentor Dr. Jamie Hodgkins, who has offered continuous encouragement, guidance, and advice in this and so numerous other decisions. Secondly, Dr. Naomi Cleghom, who first invited me to South Africa and introduced me to the wonders of Paleolithic archaeology and zoological archaeology. Naomi also took the time to direct my interest into the first of many proposals and check my faunal identifications. Without them, this thesis would not be a reality.
Drs. Christopher Beekman and Anna Warrener, who have helped guide me through the thesis process and have been gracious enough to answer my numerous questions. The CU Denver Anthropology department, especially Connie Turner, Dr. Tammy Stone and Dr. Charles Musiba, for their assistance.
And to my fellow graduate students (especially Emily), for keeping me sane (ish). And finally, to Naomi Keller, for always listening.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION......................................................1
II. THE MARINE ADAPTATION AND IMPACTS ON TERRITORIALITY...............5
Behavioral Modernity and the Coastal Adaptation................5
Subsistence and Territoriality................................15
A case study: Knysna Eastern Heads cave 1.....................21
III. SITUATING KEH-1 WITHIN THE SOUTH AFRICAN MIDDLE AND
LATER STONE AGE..................................................26
MSA and LSA sites of South Africa.............................26
Eastern cape............................................27
Western cape............................................29
Environmental and Ecological Reconstruction...................31
Subsistence Strategies of the CFR.............................36
KEH-1.........................................................39
Stratigraphy............................................40
IV. METHODS OF DATA COLLECTION AND ANALYSIS .........................42
Taphonomy.....................................................43
Assessing Nutrient Processing.................................44
vi


Size Class and Taxon...............................................46
Additional Anthropogenic Modifications ............................47
Statistical Analysis...............................................48
V. RESULTS................................................................50
Fragment Size .....................................................50
Element and Animal class Representation............................48
Size class...................................................53
Number of Individual Specimens (NISP)........................54
Taxonomic Representation ..........................................57
Burning............................................................58
Long Bone Survival ................................................60
Fragmentation morphology...........................................61
Matrix cover.................................................65
Surface modification...............................................66
Statistical analysis...............................................61
VI. DISCUSSION.............................................................69
Subsistence Strategies of KEH-1....................................69
Nutritional Stress and a Distant Coastline ........................71
VII. CONCLUSION.............................................................75
REFERENCES....................................................................77
vii


APPENDIX
A. Number of Individual Specimens Present..............................90
B. Minimum Number of Elements..........................................94
C. Minimum Animal Units................................................96
viii


LIST OF TABLES
TABLE
1. The faunal specimens of KEH-1...........................................42
2. Tukey’s post hoc on length..............................................51
3. Tukey’ s post hoc on width..............................................51
4. Size class by aggregate.................................................54
5. Taxonomy................................................................58
6. Count of long bone shafts and epiphyses.................................61
7. Nutrient and Non-nutrient breaks........................................64
8. NISP and frequencies of nutritive/non-nutritive fractures for distal limb bones...64
9. Chi-square results......................................................65
10. Surface modification of long bones......................................67
11. Frequencies of Long Bone surface modification...........................68
IX


LIST OF FIGURES
FIGURE
2.1 KEH-1 in relation to significant MSA and LSA sites......................22
3.1 Map of Southern Africa .................................................27
3.2 Photo of the KEH-1 stratigraphy .........................................41
4.1 Non-nutritive fractures (left) and nutritive fractures (right) ..........45
5.1 Box plot showing the lengths and widths of faunal specimens in each
aggregate ......................................................................51
5.2 Graph of the frequency of long bone circumference .......................53
5.3 Graph of the frequency of size classes...................................54
5.4 Graph of the frequency of high and low survival elements, based on
Lam et al., 1988................................................................55
5.5 Graph of the frequency of high and low utility elements, based on Binford, 1980 56
5.6 Graph of the frequencies of maximum burning stages by aggregate..........59
5.7 Graph of the frequencies of maximum burning stages by aggregate..........61
5.8 Frequency of the non-nutritive fractures, compared to data from
Marean et al., 2000.............................................................63
5.9 Frequency of the nutritive fractures, compared to data from Marean et al., 2000 63
x


5.10 Graph of the frequency of matrix cover on faunal remains..............66
5.11 Frequencies of tooth and percussion marks, compared to data
from Marean et al., 2000.....................................................67
5.12 Count of surface modification by aggregate............................68
6.1 Coastal distance from KEH-1 and results of faunal analysis............68
xi


CHAPTER I
INTRODUCTION
The transition from the Middle to Later Stone Age in South Africa is poorly understood, due to the paucity of sites (Mitchell, 2008; Bousman and Brink, 2018). Researchers grappling with this shift have constructed a framework to describe the behavior of the early modern humans, but how these so-called modem behaviors are defined and when they appear archaeologically is contested (Binford, 1985; Tattersall, 1995; Klein, 2000; McBrearty and Brooks, 2000; Brumm and Moore, 2005; Nowell, 2010). By characterizing a suite of behaviors and linking them to proxies in material culture, archaeologists have attempted to draw out the factors associated with the transition to the Later Stone Age (LSA), and how this differs from earlier Middle Stone Age (MSA) behavior (Klein, 2000;
McBrearty and Brooks, 2000; Nowell, 2010). A Human Behavioral Ecology (HBE) framework utilizes an adaptative, environmental lens to describe changes in human behavior, and the resulting material culture represented in the archaeological record (Binford, 1980; Bird and O’Connell, 2006; Nettle et al., 2013). The manifestations of symbolic behavior are suggested to be linked with the development of territoriality and the necessity of signaling between an in- and out-group (Brumm and Moore, 2005; Nowell, 2010; Compton, 2011a; Marean, 2014). In South Africa, the development of these signals is argued to have begun ~160ka and may be connected to a preference for stable marine resources, i.e. a coastal adaptation (Marean, 2014, 2016). However, access to these dense and predictable marine resources is impacted by glacial cycles, which resulted in expansion and contraction of the coastline and the Paleo-Agulhas plain (Fisher et al., 2010; Compton, 2011a; Marean, 2014, 2016; Will et al., 2016). Sites located on the current coastline frequently have occupation
1


histories closely linked to these cycles, and the shifts between available marine and terrestrial subsistence (Butzer, 1973; Klein, 1976; Klein and Cruz-Uribe, 1987; Sealy, 2006; Bar-Matthews et al., 2010; Jacobs, 2010; Jerardino and Marean, 2010; Steele and Klein, 2013; Marean et al., 2014). Discussions of human behavior can thus be situated in the broader Paleo-climatic context, while proxies for territoriality may bring insight to the impacts of this alternating landscape on foragers.
A newly discovered site, Knysna Eastern Heads cave 1 (KEH-1), may offer a deeper understanding of human behavior during an interglacial and glacial transition. Excavations have revealed a reccuring and dense occupation sequence, ~46ka-19ka, revealing a site that abutted the ocean during interglacial cycles, and at times of glaciation sat overlooking the Paleo-Agulhas coastal plain. Researchers have hypothesized that the dearth of sites during the MSA-LSA transition was caused by a preference for nutrient heavy marine resources, drawing hunter-gathers out onto the Paleo-Agulhas plain (Mitchell, 2008; Faith, 2013). However, the occupation of KEH-1 during global glacial cycles when the ocean was distant, may have implications for this hypothesized preference for marine resources. One option is the possible development of territoriality surrounding these resources, forcing other groups away from the presumably desirable coast. However, paleo-ecological reconstructions of the Paleo-Agulhas Plain suggest a fertile grassy plain (Rector and Reed, 2010; Copeland et al., 2016). Conversely, the exposed coastal plain and the position of the cave (~23 meters above sea level), may have offered early humans access to herds of ungulates residing in the plain. Thus, an alternate theory, that early humans turned to marine foods when the paleo-Agulhas plain shrunk and terrestrial bovids vanished (Compton, 2011a; Faith, 2013), is a possibility. In order to determine if occupation of KEH-1 may have varied in desirability as the ecology
2


shifted, proxies for nutrient stresses were examined. In Pleistocene Europe, analysis of the nutrient processing of long bones for marrow demonstrated a significant difference in Neanderthal behavior between interglacial and glacial periods (Hodgkins et al., 2016). Increased marrow processing when the coastline and marine resources are readily available may indicate stresses due to loss of terrestrial biomass and increased population pressures. Conversely, higher frequencies of marrow processing when the coastline is distant may indicate an increased preference for marine foods, and territoriality around these resources. Therefore, analysis of the faunal remains of KEH-1 have the potential to reveal the subsistence strategies during these cycles.
In order to assess if early humans experience nutritional stress during rapid climatic events during MIS 3 and into MIS 2, the skeletal remains of animals that were transported to KEH-1 for butchering were analyzed. Long bones were of particular interest because they contain meat and marrow, and they preserve better over time than bones in the axial skeleton. For taphonomic reasons, given the high fragmentation of the KEH-1 faunal assemblage, the angle and outline of the long bone fragments were assessed, to determine the timing (pre- or post-depositional) as a proxy for bones which have been broken to retrieve nutrients. Additionally, surface modification (percussion, tooth and cut marks) of long bones allowed an assessment of the agents (individuals of various species with potential to alter or delete bones) involved. This, coupled with an understanding of the broader spatial and temporal shifts across Southern Africa, will contextualize the behavior and subsistence strategies of humans as they interact with a shifting landscape.
This thesis will first review the literature on Human Behavioral Ecology, and its implications for behavioral modernity and a marine adaptation. It will consider how these
3


previous researchers in southern Africa have discussed these ideas, and how the archaeological record supports these theories. The second chapter will discuss the South African MSA-LSA archaeological record, consider proxies for environmental reconstruction and early human subsistence, and contextualize the site of Knysna Eastern Heads cave 1. It will also review the methodology used in zooarchaeological analysis and examine previous research on taphonomy and anthropogenic behaviors. Following this, it will discuss the results of this analysis, and investigate the implications for human behavior at KEH-1.
4


CHAPTER II
THE MARINE ADAPTATION AND IMPACTS ON TERRITORIALITY Behavioral Modernity and the Coastal Adaptation
Human Behavioral Ecology (HBE) is the framework used in this analysis of KEH-1. HBE grew out of the processual framework, which brought a “scientific” approach to the field, advocating for studying cultures as a means of adaptation (Binford, 1962; Flannery, 1968). Proponents of HBE argue for a consideration of fitness in studies of human behavior (Bird and O’Connell, 2006; Nettle et al., 2013). Human organisms become adapted to certain environments, and it is these adaptations which provide differential reproductive successes (Bird and O’Connell, 2006). Without the variation in environments, different cultural behaviors (as seen in material culture) would not have arisen (Bird and O’Connell, 2006). Binford (1962) characterizes this as an energetic tradeoff, whereby certain behaviors (for example, tool use) are more adaptive in certain environments. Certain South African microtechnologies (the Still Bay and Howiesons Poort industries) are argued to appear and disappear with environmental shifts (Deacon, 1978; Scott and Neumann, 2018), and possibly the subsequent variation in available fauna (Clark, 2017). The subsistence strategy of the group is critical, given that energetic requirements have an impact on reproductive success, and many models are constructed around determining the availability of different foods, and how humans acquired them (Binford, 1980; Wiessner, 1982; Bird and O’Connell, 2006). More specifically, these models analyze how humans use various technologies (and how these technologies shift in response to environment), and how subsistence strategies impact energetic expenditures of the group (Bird and O’Connell, 2006).
5


Optimal foraging strategy is a frequently utilized HBE model. This model assumes that the individual will make prey consumption choices based on a ranked scale: if I found a low ranked but available rabbit, should I take the time to kill and process the animal, or search for a more desirable, higher ranked deer? (Bird and O’Connell, 2006). In order to be the most efficient (balance energy expended by energy gained), hunters would desire to maximize “profits” by reducing search and handling time (Binford, 1980; Bird and O’Connell, 2006). Shifts in environment or population increase may result in intensification of resources, and a broadening of this “diet breadth,” which often implies that humans are making a greater effort to obtain less desirable foods (Binford, 1980; Klein and Cruz-Uribe, 1983; Bird and O’Connell, 2006). Although optimal foraging theory has been subject of criticism, given that it applies a linear, teleological aspect to human decision making, and assumes that genetic and physical adaptations are geared to problem solving (Pierce and Ollason, 1987), it represents a scientific, environmentally derived approach to understanding modern humans. As I will explain below, many researchers of the South African MSA and LSA use the HBE framework to discuss the “marine adaptation,” which has been considered critical to early human behavior and modernity.
Archaeologists studying Early Anatomically Modem Humans (AMH) often try to trace the origins of our species to a “behaviorally modem” suite of characteristics. A discussion of the assumptions, problems, and ideas behind this phrase is crucial to understanding the literature on AMH in South Africa. Researchers in paleoanthropology struggle to overcome personal and field-wide biases, frequently inserting a racial, colonialist narrative into their reconstructions of early human evolution (Athreya and Ackermann, in press). The impacts of these biases can be traced throughout the history of
6


paleoanthropology, and still retain a position in the field today (Athreya and Ackermann, in press). Many of these inferences regarding behavioral modernity revolve around the supposed cognitive and cultural superiority of European populations (Athreya and Ackermann, in press).
Immediately following the onset of processual theory, many researchers linked the appearance of modern traits to a “human revolution” in Europe (Binford, 1985; Mellars and Stringer, 1989; Klein, 1994, 1995, 2000; Tattersall, 1995, but see Mcbrearty and Brooks, 2000). By arguing that these cultural changes (~50ka) were tied to genetic mutations (proposals made before recent advances in ancient genomic sequencing), they focused on a seemingly abrupt innovative transformation (McBrearty and Brooks, 2000; Nowell, 2010). Following subsequent excavations in Africa researchers began arguing for earlier, more gradual innovations that resulted in behavioral modernity (McBrearty and Brooks, 2000). Additionally, the earliest date for modern humans is now ~300ka (Hublin et al., 2017), creating a larger discrepancy between anatomical modernity and behavioral modernity. The relatively short period of time (~50ka) following the proposed AMH genetic changes has been criticized as insufficient for incorporation into the human genome and inconsistent with current knowledge about the migratory history of human populations (Renfrew, 1996). Interestingly, in 2016 geneticists surveying modem human genomes found no evidence of genes that coded for better cognitive abilities during the last lOOKa (Mallick et al., 2016). However, researchers are still split on the origins and implications of behavioral modernity (Brumm and Moore, 2005; Nowell, 2010).
Traits that researchers included in behavioral modernity are abstract thinking, planning, innovation (behavioral, cultural, or technological) and symbolism (McBrearty and
7


Brooks, 2000; Nowell, 2010). Thus, archaeological proxies of behavioral modernity incorporate archaeological and ethnographic studies of ecology, symbolism, technology, and social networking (McBrearty and Brooks, 2000; Nowell, 2010). Early populations are frequently described as lacking certain aspects: absence of art, symbolism, structures, complex burials, and instead consisted of “informal” artifacts like unstandardized tools (Klein, 2000). However, this concept has been closely tied to cognitive capacity, and has been criticized for proscribing assumptions of “modernity” onto early humans (Brumm and Moore, 2005; Nowell, 2010). A sudden temporally constrained “explosion” of these archaeological proxies for modernity might not relate to cognitive capacity per say, but instead to socio-cultural constructs and demographic densities which amplified or more frequently left symbolic or cultural traces in the record (Brumm and Moore, 2005).
Mcbrearty and Brooks also note that a perception of gradual changes to modern behavior may facilitate a closer inspection of the proximal causes of such shifts, especially when placed in proper temporal and spatial settings (2000).
In the genetic model, behavioral modernity does not occur in the south African record until ~50ka-40ka (Klein, 2000). It has been argued that before 50 ka-40 ka humans used simplistic subsistence strategies, for example, scavenging or hunting animals perceived as less dangerous (Klein, 2000). Klein (2000) has proposed that an advantageous mutation occurred in the human lineage around 50ka, which allowed humans to adapt to almost any environment behaviorally, while retaining their morphology (Klein, 2000). Klein has used this hypothesized unique cognitive mutation to explain the complex material culture and technology found in Europe and Africa following 40ka, a time when Homo sapiens were probably the only hominin in the area (Klein, 2000). Klein argues that if the species Homo
8


sapiens is defined by an advanced cognitive mutation, then the origin of modern Homo sapiens should be restricted to the end of the Pleistocene. Klein used the idea of a unique cognitive capacity in H. sapiens to explain how modern humans colonized the globe (Klein, 2000), although ignoring earlier migrations of Homo. Klein’s hypothesis is important to understand when studying the MSA/LSA transition because it has implications for subsistence. Namely, MSA populations have been considered incapable of hunting varied or dangerous prey, while LSA hunters were top of the food chain. These changes in hunting ability are the crux of the argument for a later modern behavioral shift (Klein, 1995, 2000; Klein et al., 2007).
However, other researchers debate the separation of anatomical and behavioral modernity. Analysis of numerous South African MSA sites (including Blombos, Border cave, Die Kelders, Klasies River Mouth, and Deipkloof) suggest that early modem humans were fully capable of complex hunting strategies, and targeted dangerous prey (Faith, 2008). Both the MSA and LSA have comparable assemblage evenness (suggesting similar capabilities) and MSA populations had a wider diet breadth than their successors (Faith,
2008). MSA hunters targeted so-called dangerous prey (buffalo and pig) as frequently as LSA hunters (Faith, 2008). Given this evidence, if behavioral modernity is predicated on a subsistence strategy that includes varied prey and complex decisions, then MSA populations should be considered modern (Faith, 2008). MSA populations have also been shown to consume prey which are smaller or more difficult to capture (Wadley, 2010; Thompson and Henshilwood, 2014a, 2014b). Frequent utilization of tortoises at several sites appears to suggest a focus on diverse prey prior to the LSA (Thompson and Henshilwood, 2014b). The
9


capture of small bovids has been argued to demonstrate a broad spectrum of hunting techniques, particularly the potential for snares (Wadley, 2010).
Some research has argued that a critical aspect of behavioral modernity in South Africa included an early coastal adaptation (Jerardino and Marean, 2010; Compton, 2011a; Marean, 2011, 2016; Will et al., 2016). The site of Pinnacle Point (PP) showcases this coastal adaptation, offering evidence of marine utilization as early as ~164ka (Jerardino and Marean, 2010). In addition to repeated consumption of shellfish, there appears to be evidence for the collection of aesthetic, non-dietary shells, suggesting an appreciation for beauty which may be linked to symbolism (Jerardino and Marean, 2010). Evidence from PP demonstrates that the site was frequently occupied while the coast was close (~12km) to the shoreline, (Fisher et al., 2010). This close proximity led researchers to tie human occupation to the availability of marine resources (Jerardino and Marean, 2010). It has been argued that the decision to settle on the coast would not have occurred until humans could utilize marine resources (otherwise, the group loses 50 percent of their foraging area to water), and that this required an abstract, complex thought process (Marean, 2011). It is argued that occupants needed to understand numerous variables, including tidal shifts, to safely and efficiently gather these resources (Jerardino and Marean, 2010; Marean, 2011).
Two models have been put forth to explain the necessity and the timing of a coastal adaptation. Both focus on the role of nutrition and cognitive capacity (Will et al., 2016). The first one, by Parkington et al. (2004) emphasizes the role of nutrition in encephalization (Will et al., 2016). The rich marine diet would have offered better resources for women and infants, allowing for an increase in cranial capacity (Will et al., 2016). Compton (2011) also notes that the easily digestible, protein rich, fatty acids and nutrients would have contributed
10


to speciation of AMHs, given our brains’ caloric needs and gracile morphology. In contrast, the second model posits that humans already had sufficient cranial capacity to connect tidal variation to lunar activity and efficient collection of marine resources, resulting in a relatively late adaptation ~164ka (Marean, 2011), given that the earliest humans are now dated to 300ka (Hublin et al., 2017). However, there is evidence of Neanderthal use of marine and aquatic resources throughout the middle and late Pleistocene (Stringer et al., 2008; Brown et al., 2011; Colonese et al., 2011; Hardy and Moncel, 2011; Haws et al., 2011), although it has been argued by Marean (2014), that this does not constitute a true marine adaptation, given the temporally and spatially scattered Neanderthal sites with marine subsistence. The extent to which this signal may be a result of taphonomic processes (including destructive ocean progressions during interglacial cycles) is debated (Brown et al., 2011; Haws et al., 2011; Marean, 2014; Will et al., 2019). A more recent review, which incorporated the data from excavations in north Africa, and compared them to southern African AMH and Neanderthals, came to a similar conclusion as Marean (Will et al., 2019). Although there appear to be differences in the degree that marine subsistence is utilized, the authors argue that the temporal and spatial distribution of this behavior suggests that marine resources were important for the genus Homo (Will et al., 2019). Consumption of aquatic resources (specifically fish and other small aquatic animals) by earlier hominins has been hypothesized, although the ephemeral nature of these remains makes it difficult to determine the extent or possibility of such behavior (Stewart, 2010, 2014).
However, in an earlier paper, Will et al. (2016) argues that neither of those models concerning cognition are sufficient to explain the causes of a coastal adaptation from a long term, evolutionary perspective. They extensively review the published data on marine
11


subsistence and conclude there was “continued, systematic and habitual consumption” for a period of more than 100,000 years in multiple areas of Africa (pg. 8). The authors note that the evidence points to a repeated use of coastal sites, only when the ocean was within 10km of the site, and were likely a result of scheduled visits, not opportunistic patterns (Will et al., 2016). Lithic technology does not appear to shift dramatically in response to a marine adaptation, although there may be shifts in how it was organized (Will et al., 2016). Will et al. (2016) explicitly tie adaptation and the consumption of marine foods together, arguing that it increased the survival potential of each child. Furthermore, by buffering the population when terrestrial foods were scarce, it reduced mortality rates, especially for at risk individuals (Will et al., 2016). This would have provided an advantage to coastal foragers, and influenced their social and cultural strategies (Will et al., 2016).
Mobility has been theorized to play a critical role in the marine adaptation. For example, Compton (2011) argues that ocean regression and transgressions isolated populations on the paleo-Agulhas plain (Compton, 2011a). Specifically, the steep Cape Folded Belt (a mountainous range edging the southern plain) would have prevented movement inland, especially during interglacials, when rising ocean levels blocked the lower elevation passes on the eastern and western ends of the Belt (Compton, 2011a). If this transpired, then during glacial periods, AMH may have followed herds onto the wetter plain, only to become trapped on the Plain due to floods during glacials (Compton, 2011a). This may have also brought them in contact with the coast, resulting in adaptation and consumption of the high-protein marine resources (Marean, 2010; Compton, 2011a). Marean (2011) has argued that random use of the coastal resources would be inefficient, and habitual settlements or monthly visits (in connection with lunar impacts on tides) would have allowed
12


AMH to gain the most energy from this adaptation. There also appears to be a link between proximity to the coast, with certain sites being abandoned as the ocean receded, only to be reinhabited when the shoreline again grew close (Will et al., 2016).
Glacial periods caused increased aridity on the African interior (Owen et al., 2018). It has been hypothesized that as oceans retreated, the newly emerged shelves would be a source of fresh water (Faure et al., 2002). On the Paleo-Agulhas Plain, this is the result of rivers running off the mountains and cutting through the Plain. The slope of the continental shelf would have drawn fresh groundwater in the form of an oasis, attracting hominins and mammals (Faure et al., 2002). Thus, these now-flooded shelves would have been attractive to early humans. An intensive marine subsistence strategy might have developed in AMH due to the environmental factors. If AMH are being driven to the floodplains due to repeated aridity, perhaps this environment created the right conditions for a coastal adaptation to develop.
It has been argued that a marine adaptation led to territoriality (Kelly, 1995; Sealy, 2006; Marean, 2016). Marine resources are dense and predictable, and thus desirable, and groups will engage in conflict to maintain their rights to a source (Marean, 2014, 2016). This does not appear to apply to seasonal resources because it is expect that foragers would move on after the season ends, however, South African coastal resources can be acquired year-round (Marean, 2014). Other aquatic resources, including lacustrine and riverine, would also be worth defending, and offer an advantage (Marean, 2014). Sealy (2006) argues for an “opportunistic sedentism” at the LSA sites of Nelson Bay Cave and Matjes River rock shelter. The resources offered by coastal proximity appear to result in a territorial organization, and year-round occupation, as determined by an isotopic analysis of early
13


human dental remains (Sealy, 2006). Although, the shifting coastline during the Pleistocene would have impacted these patterns (Will et al., 2016).
Significant behavioral and symbolic shifts are postulated to be linked to the marine adaptation. Early modern humans may have become more cooperative, extending trade networks (Marean, 2014). Additionally, these groups may “signal” their boundaries through symbolic behaviors (Marean, 2014). Groups trapped on a rapidly vanishing plain may have needed a way to differentiate or signal to others the territory (Compton, 2011a). Arnold (1992) has argued that environmental variability and subsistence stress led to increased specialization and appearance of elites. Disruption of marine resources during warmer periods when the coastline was near caused political and economic power to be concentrated, and a reduced access to resources guided individuals to produce surplus goods for trade to the mainland (Arnold, 1992). Compton (2011) also postulates that symbolic goods were used for trade in MSA population during resource scarcity during ocean regression, although he does not substantiate these claims.
However, there is a growth of symbolic items associated with the MSA, particularly in South Africa, including bone tools (Henshilwood et al., 2001), engraved ostrich eggshell (Texier et al., 2010, 2013), engraved ochre (Mackay and Welz, 2008; Henshilwood et al.,
2009), engraved bone (D’errico et al., 2001), burial with grave goods (d’Errico and Backwell, 2016), and use of ochre as a pigment (Barham, 2002; Wadley et al., 2004; Marean et al., 2007). These have been argued to demonstrate early modern human cognitive capacity. If these culturally complex behaviors occurred consistently and repeatedly in the record, then it appears that early modern humans had significant sophisticated mental capabilities (per (Brumm and Moore, 2005), and numerous researchers have argued to this effect (McBrearty
14


and Brooks, 2000; Faith, 2008; Compton, 2011a; Marean, 2011; Will et al., 2016). Whether this occurred prior to a marine adaptation or following swiftly after is still a topic of debate. However, symbolic proxies become a more frequent find in conjunction with the often densely occupied coastal cave sites. This suggests a shift in the social and cultural aspects of early modern humans, which fluctuated but remained relatively stable until the MSA-LSA transition. While researchers incorporate discussions of symbolism and its significance in the archaeological record, they nearly always do so from a processual viewpoint. The sociocultural shifts and symbolism are perceived as adaptive, offering early modem humans a way to cope with their environment, adapt to new subsistence strategies, and position or arrange themselves relative to other groups in an advantageous manner.
Subsistence and territoriality
If there are groups positioning themselves relative to desirable resources, this raises the possibility that there are other groups consuming fewer desirable resources. The economic defensibility model demands conflict when the resource is large and stable enough that it is worth the time and energy budget to establish a territory (Brown, 1964). Aggressiveness is perceived as an adaptive quality, given that it would secure access to resources (Brown, 1964). The establishment of territoriality as an adaptation has been depicted as initiating a violent suite of characteristics in Homo sapiens, which also coincided with a “hyperprosociality,” which led to cooperation with non-kin (per Marean 2016). However, the territoriality model also assumes that all groups desire the same resources. Furthermore, it assumes that males but not females evolve to be aggressive, or that evolution in driven by male priorities, as suggested by Marean (2016), when he discusses the possibility of male aggression revolving round females. Therefore, components of and
15


approaches to this model are likely biased by researchers. It has also been argued that adaptations cannot “choose” to be optimal, given a rapidly changing environment (Pierce and Ollason, 1987). However, if marine resources were more nutritious and if they were easier to acquire then terrestrial ones, then it seems reasonable to infer that humans would value and defend these resources. Although there may be other constructs or lens through which to view this decision, a reduced mortality rate would mean that the exploitation of marine resources would have favored populations targeting or sampling these resources by expanding their genetic pool. Hunting of large game animals at close range is hypothesized to contribute to the traumatic injuries on early Homo species (Berger and Trinkaus, 1995; Trinkaus, 2012; Beier et al., 2018), although by the late Pleistocene the risk is considered to be mitigated by MSA technology (Dusseldorp, 2010). While collecting shellfish has the potential to be dangerous, knowledge of the tidal movements would mitigate much of that risk (Marean, 2014), further increasing the benefits of this resource.
While significant research has been directed at understanding the benefits of a marine adaptation, relatively little work has been done on those NOT consuming these rich resources, although it is implied that they existed. Determining the impacts of territoriality in conjunction with a marine adaptation can provide insight into the potential impact on individuals. The processual viewpoint tends to describe things at a larger scale, generally focusing on interactions of groups from a broader perspective. Given the often-poor temporal resolution of sites, and general paucity of sites, describing a smaller scale is difficult. Other researchers have critiqued this viewpoint for failing to incorporate a more nuanced discussion. For example, Brumfiel (1992) argues that analyzing populations and behavior obscures gender. If gender roles are not explicitly defined, they tend to be assumed, which
16


relegates women to lesser status and obscures their agency (Brumfiel, 1992). Human skeletal remains are the most effective archaeological proxy for determining these effects (Ubelaker and Owsley, 2003; Sealy, 2006), however, the relative lack of remains dating to the MSA in South Africa, coupled with the incompleteness of the remains renders this difficult (Grine et al., 2017). Therefore, other proxies are necessary to examine territoriality, and will not likely provide the resolution necessary to discuss the impacts at the level of the individual.
Previous researchers have considered how territoriality impacts hunter-gatherers. In considering ideas of hunter-gatherer “complexity,” Price and Brown (1985) question the assumptions of researchers. They argue that simply characterizing groups as simple and complex does not allow for nuanced and contextualized studies, an observation which should be applied even to modern studies. Kusimba (2005) notes that ethnographic data of hunter-gatherers may not be applicable, given the social stratification of modern groups. Some of this ties back to ideas of behavioral complexity, and deciding when and how Homo sapiens shifted socially and culturally (Kusimba, 2005). Arnold (1992) has explored how environmental and social stress led to political opportunism and development of elites. The difficulties of food procurement on North American islands following loss of marine resources required managerial systems to obtain other subsistence (Arnold, 1992). A centralized power was more efficient, creating surplus and avoiding further stresses (Arnold, 1992). This provided an adaptive advantage, because it enabled the survival of this group.
In the LSA, territoriality might be tied to population expansion. At Byneskranskrop 1 and Die Kelders 1, an LSA (13,000- 250) and MSA/LSA site (~75-50ka, and 2000-1500), Klein and Cruz-Uribe analyzed the size of the tortoise bones. There is a change is terrestrial faunal material throughout the sequences, and a shift in proportion of marine fauna at DK1
17


(Klein and Cruz-Uribe, 1983). They note a significant reduction in tortoise size during the LSA, and conclude that this is most likely related to an increased human population (Klein and Cruz-Uribe, 1983). Subsequent studies of shellfish and tortoises at Ysterfontein appear to demonstrate the same population increase (Klein et al., 2004). Shellfish appear to undergo a reduction in size in several coastal sites during this time (Klein and Steele, 2013). The authors argue that this is due to changes in intensity of collection, because individuals would be more likely to grab larger tortoises before utilizing smaller ones (Klein et al., 2004), and over-intensification would result in smaller shellfish (Klein and Steele, 2013). If these proxies for population expansion are correct, then it suggests pressure on high protein resources. The desirable marine resources may be under particularly high stress, if territoriality has been established and less mobility is occurring, due to denser populations.
The nutritional value of tortoises is high enough that, given an optimal foraging theory framework, they should be taken when encountered (Thompson and Henshilwood, 2014a, 2014b). Tortoises, have a small body size, require little effort to capture and process, they are not dangerous, thus they are a unique example of optimal prey (Thompson and Henshilwood, 2014b). It has been argued that during the coastal regression at Bloombos and Pinnacle Point, tortoises would have become a more important resource to offset the loss of high protein shellfish (Thompson and Henshilwood, 2014b).They also represent a resource that can be collected by women and children, who require high caloric yields, especially during periods of pregnancy and lactation (Thompson and Henshilwood, 2014b). Unlike large mammals, which are linked to ideas of prestige and costly signaling, tortoises and small mammals might offer a supplementary, steady resource to offset the low capture rates of the largest mammals (Thompson and Henshilwood, 2014b; Armstrong, 2016). If so, these
18


resources could offer insight into the practices of individuals who are generally perceived as having a lower rank (due to potential inability to capture highly ranked prey) (Wadley, 2010; Thompson and Henshilwood, 2014b). If rank is tied to territoriality, then discussion of how rank might be enacted in the MSA and LSA should be a subject of further research, although proxies for this are difficult to substantiate.
Wadley (2010) argues that the use of snares is related to social demands. Snares would increase efficiency, given that they don’t require continual observation, and had higher catch rates than big game hunting (Wadley, 2010). Ethnographic data includes observation of frequent capture through snares of steenbok and grey duiker, animals which are ubiquitous in the archaeological record (Wadley, 2010). Wadley suggests delayed gratification, and the ability to conceptualize animal mobility patterns linked to capture technology. She also suggests proxies for snare technology in the MSA, including high proportions of difficult to capture, diverse prey, and no evidence of raptor consumption (Wadley, 2010). Even if small prey provided regular food, this resource is assumed to lack social status (Wadley, 2010). However, small prey capture has a much higher success rate, suggesting that it would offer stability for families (Wadley, 2010). Although rank might be more fluid and negotiable in foraging populations, this small prey/large prey dichotomy may suggest differential access to resources within the group. Unfortunately, determining which gender was hunting which animals and the consequent social implications may be little more than conjecture, given the taphonomic concerns of the Pleistocene archaeological record.
These socio-cultural and economic pressures may suggest the necessity for risk management. Wiessner (1982) argues for a consideration of risk management in terms of spatial layout. If the site is open, then individuals can observe who needs food (Wiessner,
19


1982). She describes how social ties would mitigate loss, coupled with territoriality, allowing groups to control resources (Wiessner, 1982). This idea is echoed by Porraz et al. (2013) in their analysis of Diepkloof, a rock shelter containing pulsed occupation sequences spanning 107 ka-14ka. Located in an environment in close proximity to the coast, it offered a variety of resources (Porraz et al., 2013). During the period associated with Howiesons Poort technology (~70ka), stacked hearths indicate an intense occupation, corresponding to symbolic signaling on engraved OES (Porraz et al., 2013). It has been argued that this symbolic and technological advancement originated with an increased population density, which would have necessitated territoriality and symbolic signaling (Porraz et al., 2013). Porraz et al. argue that “the formation of regional identities, strengthened using symbols, would have favored and increased cultural interactions between groups, providing a favorable context for the development and diffusion of innovations” (2013: 8). Social interactions can describe how networks were developed and established in southern Africa, allowing groups to develop social networks and access a variety of ideas, technology, and possibly subsistence resources, offsetting risk that may have occurred with environmental shifts.
Bousman (2005) argues that technology shifts coinciding with environmental change enabled occupants of Blydefontein, an LSA rock shelter, to cope with risk. Using optimal foraging theory, he describes strategies to reduce shortfalls, in terms of loss prevention (Bousman, 2005). Economic failure is difficult to observe in the archaeological record, and environmental proxies are potentially misleading, necessitating multiple lines of inquiry (Bousman, 2005). Various technologies incur differing collection, processing, and maintenance costs, and diverse strategies may be practiced by groups with varying forms of
20


mobility (Bousman, 2005). He suggests that LSA hunter-gathers demonstrated a more flexible strategy, shifting technology, mobility and subsistence practices in conjunction with environmental shifts (Bousman, 2005).
Populations under stress from reduced territories, environmental shifts (which can change the available prey profiles), and potentially deadly encounters with other groups might utilize a varying set of technology to offset risk. If they find themselves restricted by the territories of competing groups, they can choose a diverse set of actions, including direct confrontation/competition, cooperation, or migration. Parsing out which is practiced at different times by different groups will challenge archaeologists. The already sparse archaeological record in South Africa is complicated by the coastal recessions and transgressions, which may have drawn populations out onto the Agulhas plain. Potential sites would have been destroyed during the flooding of the plain during interglacial periods, complicating understanding of these early populations. If the development of territoriality is linked to a preference for marine resources, then researchers would expect to see an abandonment of the interior sites as populations move out onto the floodplains (Jerardino and Marean, 2010). Flooding of the Paleo-Agulhas Plain during interglacials may have caused resource stress, given the loss of land and the newfound proximity of foragers (Compton, 2011b; Faith, 2011). Alternatively, foragers living in the interior during the glacials may suggest that other groups have established themselves around dense and predictable marine resources.
A case study: Knysna Eastern Heads cave 1
Knysna Eastern Heads cave 1 (KEH-1) offers a potential lens onto territoriality and resource access during the MSA/LSA transition. Situated on the edge of the Knysna lagoon,
21


the cave sits about 23 meters above sea level, high enough to survive the current ocean transgression. Thus far, five stratigraphic levels have been identified spanning an environmentally transitional sequence dated between -46,000 and -19,000. During the Last Glacial Maximum, the site is intensely occupied even as adjacent sites are abandoned.
Cave of Hearths
Border Ccive
® Apollo 11
® Kathu Pan
Sehonhong®
Umhlotuzana
V^Putslaogte 1
\ Knysna Eastern Heads 1
l—-------—I Nelson Boy
Soomplaas ® 1 Cave-------
Pinnacle
Cape Floristic Region | Maximum Exposure of the Continental Shelf
O IOO 200 400 600 km
2.1 KEH-1 in relation to significant MSA and LSA sites (Cleghorn, pers. comm. Created by Erich Fisher). The maximum expansion of the Paleo-Agulhas Plain would be an estimated 70-75km in front of the site, based on modeling by Fisher et al. (2010).
Understanding the available resources and faunal ecology will mean drawing upon the specimens recovered from KEH-1. Researchers at nearby sites have used faunal material to answer a range of questions concerning hunting ability, local ecology, environmental proxies, subsistence shifts, and population pressures (Klein and Cruz-Uribe, 1983; Klein et al., 2004; Faith, 2008; Thompson, 2010; Thompson and Henshilwood, 2014b; Copeland et al., 2016). KEH-1 represents a little-known period of rapid environmental shifts. Analysis of
22


faunal remains can not only clarify the local environment that early humans were navigating, but also foster a deeper understanding of how they used and reacted to this environment. Drawing upon a framework of HBE, faunal analysis can illuminate how landscape impacted early modern human subsistence strategies, and the ways social networks, mobility, and territoriality were shaped.
If marine resources are as significant to early humans as has been assumed, then the intense site use over time despite ocean retreat is curious. Possible explanations for the recurrent occupation at KEH-1 during this time are 1) that use of marine foods was only one of several subsistence strategies employed by early humans in this region; 2) the intersection of terrestrial and riverine resources around Knysna rivaled coastal resources; and 3) competition and territoriality caused by dense marine resources on the coast forced groups into the interior. It may be that the location of KEH-1, overlooking the Agulhas plain, positioned next to potential riverine, lacustrine, and terrestrial resources, offered early humans a rich and valuable setting during all global climate changes. The Agulhas plain is a now-submerged continental shelf, which was the setting for rich terrestrial resources (Fisher et al., 2010; Compton, 2011a; Copeland et al., 2016). The cave, situated 23 meters above the Plain and current sea level, represented a defensible position, offering security to inhabitants. These factors made it a highly desirable site, and the location may have been preferred, as Hunter-gathers could track animal movement on the exposed Paleo-Agulhas Plain.
Conversely, during glacial periods when the coast was far away, if competition for coastal resources was high so that territories on the coastline are established and defended by rival groups, then the occupants of KEH-1 may not have had access to the rich marine resources. If so, occupants of KEH-1 might focus on obtaining protein rich resources in the
23


area around Knysna, to compensate for the dearth of marine foods. However, increased population densities are hypothesized to correlate with the contraction of the Plain, along with expansion of populations into presumably “less favorable” habitats (Faith, 2013). The loss of the rich terrestrial resources of the Paleo-Agulhas Plain to flooding may have resulted in stress, as populations lost ground and fought for territory. This thesis examines to what extent marine resources are a key component of the early human diet during the MSA-LSA transition in South Africa. If, per optimal foraging theory, they are a preferred resource (Sealy, 2006; Jerardino and Marean, 2010; Brown et al., 2011; Compton, 2011b; Marean, 2011, 2014; Will et al., 2016, 2019), then I expect to see indications of nutritional stress when the coastline is distant. If marine resource use is tied to population density and loss of the Paleo-Agulhas Plain, then I expect to see nutritional stress when the coastline is within the daily foraging radius. If the occupation of KEH-1 simply indicates multiple foraging strategies by early Hunter-Gatherers, then the indications of nutritional stress should be absent or not statistically different.
While marine resources offer significant protein and nutrients for foraging populations, there are other resources which can provide similar benefits. The nutritive components of long bone marrow are comparable to marine resources (Will et al., 2013). Populations under nutritional stress have been shown to process long bones more extensively for marrow (Hodgkins et al., 2016). Critically, this includes processing long bones with minimal resources, including the distal portions of the feet, which contain relatively insignificant amounts of marrow (Hodgkins et al., 2016). If early modern human populations at KEH-1 were stressed, then it is expected that they would process the terrestrial resources more strongly. Conversely, if there is relatively little processing of long bones, then it would
24


be argued that these groups were not stressed. This type of analysis could offer insight into the decision to remain at KEH-1 during ocean regression.
The answer to this question can offer a discussion on how early modern humans coped with environmental shifts. If the marine adaptation is tied to behavioral modernity, then the territoriality and competition which are hypothesized to follow can be tested. Although the flooded Agulhas plain likely represents a loss of much of the data concerning the MSA/LSA transition, KEH-1 offers a window onto this fascinating period. The potential for elucidating these impacts on early modern humans offers a unique perspective on MSA/LSA populations. Through this lens, we can begin to describe how territoriality and differential access to resources were shaped and discuss the potential effects on these peripheral groups.
25


CHAPTER III
SITUATING KEH-1 WITHIN THE SOUTH AFRICAN MIDDLE AND LATER
STONE AGE
MSA and LSA sites of South Africa
The Middle Stone Age/Upper Paleolithic was an important period, capturing both the globalization of EMH and disappearance of Homo neanderthalensis (Clark and Kandel, 2013). However, the MSA record in south Africa is sparse, especially the transition from the MSA to LSA (Faith, 2013). There is gap from 50 ka to 25ka, which occurs simultaneously at numerous sites (Klein, 1976). Sites containing an occupation sequence for this period occur sporadically, in the eastern part of southern Africa (Border Cave, Sehonghong, Sibudu, Rose Cottage cave, and Klasies River Mouth) to the western cape (Deipkloof, Elands Bay, Putslaagte, Varsh Rivier, Klein Kliphius) and the Cape Floral Region (CFR), which is included in the Western cape, but differs in rainfall patterns (Boomplaas, Buffelskloof, Nelson Bay Cave, and KEH-1). As KEH-1 is in the CFR, it represents a critically underrepresented and misunderstood period of prehistory. In order to understand the broader context of the MSA/LSA transition, this chapter will examine sites with occupations during the MSA/LSA transition, as well as discuss previous work on paleoenvironmental reconstruction, and summarize environmental, ecological, and archaeological discussions of the CFR. Finally, it will describe KEH-1.
26


3.1 Map taken from Mackay (2010), showing sites across Southern Africa in relation to rainfall zones (based on Chase and Meadows, 2007). Sites labelled by Mackay as follows: AXI = Apollo XI, BBC = Blombos, DRS = Diepkloof, EBC = Elands Bay Cave, HRS = Hollow Rock Shelter, KFR = Klipfonteinrand, KKH = Klein Kliphuis, KRM = Klasies River Main, NBC = Nelson Bay Cave, NT = Ntloana Tsoana, PC = Peers Cave, RCC = Rose Cottage Cave, SC = Sibudu Cave, SHH = Sehonghong, SHO = Shongweni, SIB = Sibede, UMH = Umhlatuzana. Sites mentioned in this text highlighted by the author. Sites mentioned but not depicted here are shown in figure 2.1.
Eastern cape
Border Cave (227 ka - 39ka) within the borders of Swaziland, demonstrates some of the oldest evidence for modem behavior proxies (d’Errico et al., 2012a; d’Errico and Backwell, 2016). In addition to the earliest symbolic burial, recovered bone tools are argued to be evidence of complex behavior, and the use of personal ornaments as reoccurring
27


symbolic motifs (d’Errico et al., 2012a). Rose Cottage Cave dates from 250ka to 20ka (Clark, 1997). The lithic industry contains a sequence similar to Sehonghong (~32ka -20ka), diverging from typical MSA or LSA technology (Clark, 1997). The Early Later Stone Age (ELSA) is poorly understood and lacks temporal contextualization, as it appears gradually across Southern Africa (Clark, 1997). Sehonghong contains a series of shallow hearths and a larger hearth “pit” (Mitchell, 1994), and two bone tools (Clark, 1997). Sibudu captures nearly 100,000 years of early modern human occupation (terminating ~30 ka), framing discussions of behavioral modernity and subsistence strategies (Clark and Plug, 2008). Evidence from this site suggests human accumulation of fauna from both closed and open habitats (Clark and Plug, 2008). Occupants may have ranged long distances from the cave to obtain prey, especially during the Howisen’s Poort (HP) age (Clark and Plug, 2008). Early modem humans at Sibudu focused on local ecology, which is argued to indicate “a higher degree of behavioral plasticity than that generally attributed to MSA populations” (Clark and Plug, 2008, pg. 11). These sites offer insight into the habits, technology, subsistence and culture of ELSA people in South-central southern Africa, however, they are over 200 kilometers from the coastline. Although it is possible that the inhabitants were part of a social network connecting them to the coastline, no evidence of this currently exists.
Klasies River Mouth (KRM) (125 ka - 30 ka), is a series of caves adjacent to the coastline (Klein, 1976). Early human remains have been discovered at the site, and have been argued by some to represent a hominin other than AMH (Lam et al., 1996), although the remains are mostly those of adults, unlike at other sites (Grine et al., 2017). The site displays several examples of behavioral modernity—carved ochre and hunting large ungulates (Milo, 1998; d’Errico et al., 2012b). analysis of lithic material from KRM and Rose Cottage
28


suggests an environmental driver for the disappearance of the HP during the glacial/interglacial transition ~60 ka (Villa et al., 2010). Although KRM is within 160 kilometers of KEH-1, it contains no occupations representing MIS2, when the ocean retreated.
Western Cape
The Western cape has yielded a handful of transitional MSA to LSA sites, which have aided in the environmental and ecological reconstruction, descriptions of hunting strategy, population estimates, and understanding technological shifts. Elands Bay contains three occupational “pulses,” ranging from 300 ka to 13 ka (uncalibrated) (Klein and Cruz-Uribe, 1987). Variability in recovered species between the stratigraphic levels correlates to the rising sea and vegetation shifts occurring during the beginning of the Holocene, as shrubland and forest replaced the grasslands of the late Pleistocene (Klein and Cruz-Uribe, 1987). The number of juvenile seal remains suggests a seasonal use of the site corresponding to seal mating cycles, and a reduction in tortoise size with increased human population densities (Klein and Cruz-Uribe, 1987). Diepkloof rock shelter has occupation sequences spanning the MSA (107ka) through post-HP (57-46ka), and the LSA (14ka) (Jacobs et al., 2008; Steele and Klein, 2013; Tribolo et al., 2013). Diepkloof provided AMH with a strategic setting, offering riverine, coastal, and terrestrial resources (Porraz et al., 2013). Seal remains occur infrequently in MIS4 (when the coast was much nearer), and more frequently during the LSA, despite the fact that the coast was at least 18 km distant (Steele and Klein, 2013). These factors have been used to suggest social networks, increased foraging radius, an importance of marine resources, and a possible use of the lake shore to reach the coast (Steele and Klein, 2013).
29


The occupation of inland rock shelter Putslaagte 8, occurred in pulses throughout MIS 4, 3, 2, (75 ka - 13 ka), with robust Holocene deposits (Mackay et al., 2015).
Occupation may not be limited by population expansion, but rather to mobility among hunter-gatherer groups using multiple resources as early as ~75 ka (Mackay et al., 2015). Varsh Rivier contains the youngest known occurrences of SB and HP industries (HP ~45 -41ka, SB ~59ka-45ka) (Steele et al., 2016). The LSA deposits contain marine shell despite a distance of more than 45 kilometers to the coastline (Steele et al., 2016). Klein kliphius rock shelter dates from 68.1ka to~20ka (Mackay, 2010). Although site use appears to cease or become intermittent around 55ka, it becomes more intense at 30ka (Mackay, 2010). Sites in the western cape are occupied in pulses, and both Diepkloof and Varsh Rivier are abandoned during the interglacial/glacial transition between MIS3-2. Despite significant distances to the coastline, marine resources continue to be transported to both sites.
In the CFR (the southern edge of the Western Cape), few published sites are dated to the MSA/LSA transition, while others are abandoned during this time. Boomplaas cave is an inland site with deposits spanning 65ka-15 ka, and an abundance of faunal remains (Faith, 2013). Analysis demonstrates that the remains were accumulated by carnivores until ~50ka, and then the agents of accumulation were mixed, with contributions from humans, carnivores and raptors until the LSA (~18ka), when humans appear to be the sole inhabitants (Faith, 2013). Faith (2013) argues that this signature indicates low-intensity occupations of the site until the LSA, when population densities are thought to have increased. Buffelskloof is approximately 40km to the west of Boomplaas, although excavations have not revealed any occupations earlier than 23ka (Opperman, 1978). Nelson Bay Cave (NBC) was occupied prior to 50ka, and then again from 18ka-5ka (Butzer, 1973). Although it represents one of the
30


closest excavated sites to KEH-1 (~30 km), there is a gap in the occupational sequence until ~20ka (Loftus et al., 2016), when both sites begin to line up. The site is notable for its unique technology, the Robberg (dated -20-12ka), and for its Holocene deposits, which contain human remains (Deacon, 1978; Sealy, 2006; Loftus et al., 2016). Analysis of the trophic levels from the human remains are consistent with habitual consumption of a marine-rich diet (Sealy, 2006).
The southern CFR is characterized by inadequate representation of transitional sites. The abandonment of NBC, only 30km east, makes the recurrent occupation of KEH-1 more fascinating. What kept humans at KEH-1, and sporadically, although consistently, at Boomplaas? NBC is similar in many respects to KEH-1: located on the current coastline, abutted by the Knysna forest, and offering a view of the exposed Paleo-Agulhas plain. Determining the potential of the caves will require analysis of the micro-environments and the resources available to AMH.
Environmental and ecological reconstruction
In order to understand how humans evolved, we must understand the environment in which they coped. Fisher et al., (2010) propose the term “paleoscape’ to encompass the environmental change which drastically alters the landscape of the southern coast. Paleo-environmental reconstruction is an important part of faunal analysis, as the animals occurring in an area can provide insight to the local environment. Different environments would have offered different challenges to EMH, and the different responses of groups led to different social-cultural behaviors, hunting decisions, prey return rates, and long-term skeletal element survival. Numerous researchers have used faunal assemblages to reconstruct these environments and understand the challenges faced by early modern humans.
31


Other Pleistocene faunal assemblages have been used to reconstruct environments (Klein, 1989; Rector and Reed, 2010; Faith, 2011, 2011; Thompson and Henshilwood, 2011; Copeland et al., 2016; Hodgkins et al., 2018). Hyena dens are particularly useful for this analysis, as hyenas consume a broader range of species and collect skulls and dentition (Rector and Reed, 2010). Cranial material, particularly dentition, is especially useful for comparative analysis (to determine the genus/species) (Rector and Reed, 2010), and isotopic analysis, which provides information on the mobility patterns of animals on the submerged continental shelf (Copeland et al., 2016; Hodgkins et al., 2018). Once species has been determined, the researchers can compare the modem behaviors of these animals to better understand the local Paleo-ecological zones. Ostrich eggshell has also been sampled isotopically, providing insight to paleo-environments (Lee-Thorp and Ecker, 2015; Hodgkins et al., 2018).
Smaller mammals (.75-4.5kg) and microfauna are often overlooked at archaeological sites, as larger animals are hypothesized to provide better insight into dangerous/risky game hunting, mobility, and paleoenvironmental reconstructions (Armstrong, 2016). However, at certain sites, including Die Kelders, where the assemblage is more than 85% small mammals, they are critical to understanding of accumulation processes and environmental trends (Armstrong, 2016). Smaller animals/microfauna are more frequently used to understand the accumulation processes, as etching may implicate non-human agents, and thus, differential spatial or temporal occupations (Marean et al., 2000; Thompson and Henshilwood, 2011, 2014a; Enloe, 2012; Faith, 2013). This does not preclude the potential for these species to contribute to environmental and ecological reconstructions, however, it does require that
32


more archaeologists become trained in the identification and analysis of smaller mammals (Armstrong, 2016).
Paleo-environmental reconstruction based on fauna is not without complications. A fundamental difficulty is that comparative analyses are closely tied to extant communities (Rector and Reed, 2010), requiring the researcher to draw assumptions about the similarity of behavior despite temporal variation. Certain ungulates utilize multiple biomes, further complicating the issue (Dominguez-Rodrigo and Musiba, 2010). Describing the environment and ecology is complicated by the fact that different predators will prey upon different animals, likely excluding a portion of the potential prey (Thompson and Henshilwood, 2011). The location of the site might impact the animals transported, as larger animals are more likely to be processed at the kill site (Schoville and Otarola-Castillo, 2014). Unfortunately, conclusions based on the skeletal elements will be impacted by the survivability of an element (Marean and Spencer, 1991; Lam et al., 1998; Cleghorn and Marean, 2004; Faith and Thompson, 2018). Thus, only a subset of the biocoenosis will become incorporated into the archaeological record.
In addition to human food preferences/prey choice, taphonomic issues may reduce the accuracy of reconstructions. Time averaging, or the variability in temporal deposition in the same stratigraphic layer, may result in an inaccurate recreation of the paleobiome, as multiple transitions might become collapsed into one layer (Behrensmeyer et al., 2000; Dominguez -Rodrigo and Musiba, 2010). Fauna accumulated outside of anthropogenic contexts (i.e. “natural” fossil deposits) tend to be derived from lacustrine/alluvial zones, which have a higher probability of preservation (Behrensmeyer et al., 2000; Kidwell and Holland, 2002; Dommguez-Rodrigo and Musiba, 2010). These either represent the fauna from this specific
33


environment, or migratory fauna from numerous biomes, and skew the data towards “mosaic” environments (Dommguez-Rodrigo and Musiba, 2010). Subsequently, researchers have raised concerns about the accuracy of paleo-environmental reconstruction (Dommguez-Rodrigo and Musiba, 2010). Although these should not be taken lightly, faunal analysis still represents one aspect of paleo-environmental reconstructions and should be combined with other methodologies to ensure accuracy.
Methods of Paleoscape reconstructions have also included isotopic analysis of speleothems (Bar-Matthews et al., 2010), and geospatial modeling to determine the extent of the exposed Paleo-Agulhas plain during the MSA (Fisher et al., 2010). Many of these are coupled with high-resolution temporal reconstructions, including optically stimulated luminescence (Jacobs, 2010) and cryptotephra (Smith et al., 2018). Paleo-botanical data, including pollen, has been used to determine the extent and type of arboreal versus grassland habitats in the CFR (Chase and Meadows, 2007). Currently, there are nine floral biomes recognized in South Africa, although their borders during the Pleistocene are disputed (Scott and Neumann, 2018). The fynbos biome encompasses much of the CFR, and is recognized as having the highest plant biodiversity in the world (Scott and Neumann, 2018). Additionally, the largest component of the forest biome in southern Africa rests in the area around Knysna (Scott and Neumann, 2018), although the extent of this forest during the Pleistocene is unknown.
These methodologies have provided previous researchers with a myriad of sound data on the effects of glacial/interglacial shifts on the environment and ecology of the Paleo-Agulhas plain. Bathymetric modeling has pinpointed the extent of exposure, allowing subsequent researchers to determine the bio-available landmass at intervals of 1.5 ka (Fisher
34


et al., 2010). Expanded grasslands are correlated to rainfall on the plain, while the area around the Pinnacle Point sites appears to be a transitional zone from the sparse shrublands from the northern mountains (Bar-Matthews et al., 2010; Rector and Reed, 2010). The paleo-Agulhas Plain offered a rich landscape, with a diverse and densely populated large animals utilizing the open, grassy area (Marean et al., 2014).
The increased frequency of the largest ungulates during interglacial periods may be linked to the environmental changes associated with the available species (Clark and Kandel, 2013). It has been proposed that similar ratios of buffalo in the MIS5 and MIS3 at Die Kelders, Klasies River Mouth, and Pinnacle Point suggest that open environments could be characteristic of interglacial periods (Clark and Kandel, 2013). Copeland et al. (2016) theorize that the paleo-Agulhas soils may have been considerably richer than those of the CFR, offering rich grass resources to larger ungulates during ocean regressions. Ungulates appear to prefer the paleo-Agulhas plain, only rarely venturing north (Copeland et al., 2016). Although a species capture rate is correlated to the encounter rate, if a species is highly represented in an assemblage, researchers consider other ways that human behavior might be affecting this accumulation. For example, eland are highly represented by Blombos, DK1, and KRM, although researchers believe that they have a low population density (Dusseldorp,
2010). This may suggest that MSA populations were well-informed of eland behavior and could choose their encounters (Dusseldorp, 2010). The variation between the expected and realized capture rates may indicate that the MSA populations responded to environmental shifts which influenced the availability of the species (Dusseldorp, 2010).
The reconstruction of Paleo-environments has allowed researchers to hypothesize correlations between subsistence strategies and technological innovation. However, although
35


the environment may have prompted the technological shifts, it does not control how the culture changes (Deacon, 1978). For example, during the early stages of Holocene occupation at NBC, there was a shift in environment (grassland to brush), causing hunters to take fewer grazers and more browsers, but also animals >5Okg (Deacon, 1978). This change is argued to correlate with shifts in micro-technology and a more diverse toolkit (Deacon, 1978). The Still Bay (~76-72ka) and Howiesons Poort industries (~65-59ka) appear to be associated with both glacial and interglacial pollen assemblages on both interior and coastal sites (Scott and Neumann, 2018). The variation of the fauna at Sibudu during the Still Bay might suggest a link between technological variability, available fauna, and environmental shifts (Clark, 2017). The decline in species during the Howiesons Poort might be attributed to reduced diet breadth, and thus possible intensification of reduced resources (Clark, 2017). Previous work on the CFR has created an environmental sequence covering the MSA and LSA. However, much work remains to be done in understanding the area around Knysna.
Subsistence strategies of the CFR
Subsistence strategies of MSA and LSA hominins have implications for human behavioral modernity, environmental adaptation, and social constructs. A key feature of the CFR is the intersection of marine and terrestrial resources, and the contributions of each to early human diets (Marean, 2011). Although archaeological discussions of subsistence are generally dominated by large mammals, it is argued that shellfish played a critical role in demographic expansion and adaptation in the CFR (Jerardino and Marean, 2010). However, it has been argued that the MSA does not have a singular subsistence strategy (Thompson and Henshilwood, 2011). This appears particularly true across the temporal span of the MSA and the environmental shifts which characterize MIS 5 to MIS 2. Although shellfish
36


exploitation is high, less evidence for marine mammals has been uncovered, and it appears that they were consumed opportunistically (Thompson and Henshilwood, 2011).
Understanding the transport decisions of early modern humans requires an understanding of the environmental, ecological, and cultural variables which impacted the availability and locations of resources (Dusseldorp and Langejans, 2013). The variety of marine resources may provide a proxy for the mobility patterns and transport decisions utilized by people (Dusseldorp and Langejans, 2013). However, determining whether shifts in marine resources, the quantity of these resources, and the prey age profiles reflects seasonal distribution, environmental factors, or intensification is another question (Dusseldorp and Langejans, 2013). Some animals, like whales, may be nearly invisible in the archaeological record, due to enormous, impossible to transport skeletal elements (Dusseldorp and Langejans, 2013), but see Jerardino and Marean (2010). Additional resources (seal and mussels) might have been processed where they were collected, vanishing from the record as the ocean regressed (Dusseldorp and Langejans, 2013).
However, researchers should focus on the spectrum of represented archaeological faunal material, to fully understand subsistence strategies of MSA populations (Thompson, 2010). Researchers have used the CFR to contribute to arguments of behavioral modernity not only in terms of marine subsistence, but also in context of the “large and dangerous prey” hypothesis (Klein, 1995, 2000; Faith, 2008; Dusseldorp, 2010). Dusseldorp (2010) and Faith (2008) argue that MSA hunters possessed the capacity for complex hunting strategies, targeting dangerous and difficult prey. This may suggest that large mammals held a place alongside marine foods or were targeted during ocean regressions.
37


Early modem humans also readily gathered tortoises, (Chersina angulata) which are available for collection and consumption year-round (Klein and Cmz-Uribe, 1983). Chersina is the most common tortoise recovered from Bloombos, and it appears that humans processed them with fire and hammerstones (Thompson and Henshilwood, 2014a). During the coastal regression at Blombos and Pinnacle Point, tortoises would have become a more important resource to offset the loss of high protein shellfish (Thompson and Henshilwood, 2014b).
Other frequently overlooked resources available to humans include small fauna, birds, ostrich eggshell, and botanical foods. It is hypothesized that early humans may have utilized the available small fauna to fill any gaps left by a receding coastline and “higher-ranked” coastal resources (Armstrong, 2016). Prevailing theories consider bird exploitation to be synonymous with the LSA, however, avian remains are noted throughout the MSA deposits at Sibudu (~77ka) (Val et al., 2016). The ubiquity of these remains in MSA deposits showcases the ability of early humans to utilize a diverse and complex resource base (Val et al., 2016). Ostrich eggshell is frequently recovered at MSA and LSA sites, suggesting that early humans were aware of the nutritional qualities, and utilized this resource for subsistence (Collins and Steele, 2017; Hodgkins et al., 2018). Consumption of geophytes (the CFR is noted for its geophyte diversity) is known from sites as early as 100 ka (Deacon, 1993). The discovery of these diverse resources in archaeological contexts suggests that early humans were not only aware of the myriad of resources but had the ability to utilize them. However, this does not always imply that they were reliant on any particular resource but may have utilized them more opportunistically.
Early modem humans in the CFR collected, hunted, and consumed a diverse subsistence base. Some of their food choices are arguably novel (marine subsistence), while
38


others may reflect strategies used more broadly across southern and eastern Africa. MSA and LSA populations do not appear to vary in their ability to procure diverse foods (Faith, 2008; Dusseldorp, 2010; Wadley, 2010; Thompson and Henshilwood, 2014b). Instead, the variation in prey choice appears to correlate with environmental shifts. As KEH-1 sits at the edge of an extinct ecosystem across extreme changes in sea level, it can offer critical insight into early human subsistence.
KEH-1: Location and stratigraphy
KEH-1 is located on the exterior of the Knysna heads, facing the Indian Ocean. It is approximately 23 meters above sea level, ensuring that the deposits were not washed out during recent high sea stands, although it would have been subjected to destruction during MIS5e (~128ka-l lOka) high seas stands (Fisher et al., 2010). The coastline in this area is part of the Table Mountain Super group, and quite rocky and riddled with caves. The closest site to KEH-1 is KEH-2, which sits directly east and slightly lower than KEH-1 but has lost much of its deposits to tidal action. This is an archaeologically rich area, and surveys have uncovered an ESA open air site within five kilometers, and several MSA/LSA sites embedded within the Featherbed caves on the western side of the Knysna Heads. Although no other archaeological excavations have been recorded at Knysna, several finds have been published. A painted seal scapula (age indet.) was recovered from one of the many caves around the Knysna heads and currently rests in the British museum (Sealy, 2006). Recently, fossilized human footprints estimated ~90 ka were discovered in a cave less than 2 kilometers to the west (Helm et al., 2018).
39


Stratigraphy
The deposits of KEH-1 at the mouth of the cave are cut by an erosional slope, allowing excavation to target the MSA/LSA transition. They are capped by 1.5-2 meters of Holocene shell midden. The cliff face above is sheer; below is a steep slope covered with thick fynbos, terminating in barren horizontal planes of quartzite which are partially submerged by tidal action. The cave is currently accessible by a narrow, steep path on the SW corner. Following a test excavation in 2014, the cave has been excavated in every subsequent year, removing more than 40,000 archaeological finds. The site is relatively undisturbed, except for the slumping deposits at the front of the cave. Archaeological material (out of context) has been recovered from the path below. Round depressions at the back of the cave (with a diameter > 1 meter) suggest that previous digging occurred, although no records have been uncovered. It is possible that this was unsanctioned excavation, resulting in the removal of indigenous burials (Cleghom, pers. comm). Smaller disturbances (burrows) have been recorded in several layers.
At KEH-1, “aggregate” refers to archaeological layers that, although they may have been accumulated over thousands of years, appear to have been deposited under similar environmental conditions. Therefore, the units of analysis here are not the individual units excavated, but rather the broadly similar deposits, in order to describe shifts in human behavior temporally (Karkanas et al., 2015). The lowest stratigraphic aggregate, the Dark Brown Spally (DBS) >46 ka-34 ka, contains sparse evidence of anthropogenic occupations, although charcoal is plentiful. The coastline is an unknown distance away but is not estimated within daily foraging radius (10-12km) (Cleghorn, pers. comm). Small amounts of fauna (N=101) have been recovered. The Dark Shelly (DS) ~32ka-29ka, is an anthropogenic
40


deposit with large accumulations of shell, and relatively dense fauna and lithics. The coastline is likely within ten kilometers of the cave, and it appears that occupants relied heavily on marine resources at this time (Cleghorn, pers. comm). Dates on the Dense Hearth Aggregate (DHA) occur around the start of the last glacial maximum (~26ka -22ka), when the Paleo-Agulhas Plain begins its final, most recent exposure. This aggregate contains the highest find density, and is unique in its large, well preserved hearths. These extend across the entire entrance of the cave and are densely stacked. To date, more than 50 hearths have been excavated. Large amounts of ochre have also been recovered. The Orange Brown Sandy (OBS) currently has one date, at ~19ka, at the height of the Last Glacial Maximum. Around this time, the maximum exposure of the Paleo-Agulhas plain is estimated at 70 kilometers (Cleghorn, pers. comm). Anthropogenic finds persist throughout this level, although high numbers of microfauna have been found, the accumulating agent of which is currently undetermined. Above this rests a thick sterile layer overlain by the shell midden, currently undated and mostly unexcavated.
3.2 Photo of the KEH-1 stratigraphy (Cleghorn, pers. comm).
41


CHAPTER IV
METHODS OF DATA COLLECTION AND ANALYSIS
KEH-1 was mapped in 2013, and control points were established to allow geodetic control of the cave (Cleghorn, pers. comm). Once the Universal Transverse Mercator grid was established, excavations were based on this system (Cleghorn, pers. comm). Excavations were restricted by stratigraphic unit and lot, contained within 50cm x 50cm excavation units, and each lot was assigned a unique identification number. Each unit was mapped with a total station, and finds were mapped upon removal, without any size cutoff, and assigned a find number. Upon removal, finds were washed, sorted, and stored in the Mosselbaai lab until analysis.
Aggregate mum total Analyzed (frag/tap) total % Analyzed
OBS Glacial 902 475 52%
DHA Transition 9303 1127 12%
DHA/DS Transition 291 263 90%
DS Interstadial 688 536 78%
DBS Glacial 101 32 32%
unassigned 98 1 1%
Total assemblage 11401 2434 21%
Table 1: Description of the number of faunal specimens analyzed for fragmentation
and taphonomy by stratigraphic aggregate
More than 11,000 faunal specimens have been recovered during the excavations (artifacts from the 2017/18 field seasons, which have yet to be sorted, are not included here). A majority (>9,000) of these are from the DHA (table 1). Only a small number of fauna have been recovered from the oldest aggregate, the DBS. For the scope of this project, we aimed to analyze a minimum of 10% of the fauna in each aggregate. Analysis occurred over the summers of 2017/2018, resulting in our analysis of 2,434 specimens, or 21% of the total
42


faunal assemblage. Fragments were initially chosen at random, (i.e., beginning with the earliest excavated units), until the minimum frequency of finds per aggregate were analyzed. Specimens were then analyzed sequentially, to begin working through the assemblage.
Taphonomy
Faunal specimens were assessed using zooarchaeological standards, to determine the extent and origin of fragmentation, degree of taphonomic damage, size and taxon of animals, and any anthropogenic modifications. Taphonomic effects, including rodent gnawing, sediment loading (trampling), gastric etching, acidic soil/rainwater, weathering, displacement, and density mediated attrition may delete or alter faunal remains (Behrensmeyer, 1978; Lyman, 1984; Gifford-Gonzales et al., 1985; Marean and Spencer, 1991; Capaldo and Blumenschine, 1994; Lyman and Lyman, 1994; Dibble et al., 1997; Lam et al., 1998; Behrensmeyer et al., 2000; Marean et al., 2000; Pickering et al., 2003; Dominguez-Rodrigo et al., 2009a; Enloe, 2012). In order to determine human behavior, archaeologists must first rule out natural accumulations, alterations, and effects that would skew the evidence. Archaeologists should not assume that the assemblage is a complete representation of everything that was deposited, much less that everything that was brought to the location by humans or non-human agents (Behrensmeyer et al., 2000). In order to assess the effects of taphonomy, specimens were examined to determine the completeness of surface, evidence of breaks, etching, and weathering. If the surface is eroded/removed, it suggests that an agent/process worked to remove it (for example, acidity in the soil/stomach, weathering), and that surface modifications related to nutrient processing (cut/tooth/percussion marks) may have been deleted (Marean et al., 2000). Fragments were measured for both length and width to compare overall specimen size between aggregates,
43


which could reflect size sorting by natural or anthropogenic processes, and potentially correlated with degree of fragmentation. When bones of similar sized animals undergo varying degrees of fragmentation, it may suggest either an intensive hominin processing strategy or more frequent exposure to destructive post-depositional processes, such as trampling.
When possible, bones were identified to element, using the comparative collection in the Mosselbaai Archaeology Project (MAP) lab. This allowed for the determination of the frequency of high survival versus low survival elements. Density mediated destruction can alter interpretations of archaeological assemblages (Brain, 1969; Lyman, 1984; Marean and Spencer, 1991; Lam et al., 1998; Marean et al., 2004). High survival, i.e. dense bones, include parts of the skull and long bone shafts, both of which are more likely to survive carnivore attrition and taphonomic processes (Cleghorn and Marean, 2004). Conversely, low survival elements, (ribs, vertebrae, long bone epiphyses, and pelvic bones) tend to be grease bearing, and consisting of spongy trabecular bone, both of which combined reduce the likelihood of survival (Cleghorn and Marean, 2004). Significant debates have arisen over the need to assess survivability, as the elements most likely to be destroyed are also the most identifiable, and thus, more likely to be retained and analyzed (Marean and Spencer, 1991; Marean and Kim, 1998; Pickering et al., 2003; Marean et al., 2004). Currently, many zooarchaeologists are trained to identify shaft fragments, which are not only more likely to survive, but are also more likely to retain anthropogenic surface modification.
Assessing Nutrient processing
Nutrient processing refers to the reduction of long bones to retrieve the marrow contents. Distinguishing between nutritive versus non-nutritive breaks (also known as green
44


versus dry fractures), requires analyzing the angle and outline of the ends of the shafts, based on methods developed by Villa and Mahieu (1991). Nutritive fractures include curved outlines and oblique angles, while non-nutritive fractures consist of straight outlines and right angles (fig. 4.1) (Villa and Mahieu, 1991). Processing of bones also results in changes in the circumference (Bunn, 1983). Non-nutritive breaks are the result of taphonomic effects, and can include trampling, burning, sediment loading, and excavation (Hodgkins et al., 2016). Frequently, this analysis is carried out on shaft fragments, for two reasons. First, as mentioned above, these are more likely to survive and have cut/percussion marks. Additionally, trabecular bone breaks in a different manner than cortical bone, rendering the criteria inapplicable (Villa and Mahieu, 1991). Multiple sets of experiments have been compiled by Marean et al. (2000), and provide a robust dataset with which to compare zooarchaeological data. Hodgkins et al. (2016) demonstrated the potential for using nutritive values across climatic events to understand how hominin subsistence strategies are altered. This method applies the data to broader paleo-environmental, paleo-ecological, and evolutionary trends.
4.1 Non-nutritive fractures (left) and nutritive fractures (right).
45


Archaeologists must also control for issues of equifinality (Lyman, 2004; Marean et al., 2004). Once the timing of the fractures has been established, the archaeologist must determine the agent that caused the fractures. Both carnivores and hominin processing will create green fractures on long bones (Marean and Spencer, 1991; Marean et al., 2000). By examining the bones under a microscope (analysis of surface modification), researchers can determine the presence of tooth, cut or percussion marks (Marean et al., 2000). Experimental work has demonstrated that tooth, cut and percussion marks can be reliably distinguished from each other (Brain, 1981; Potts and Shipman, 1981; Capaldo and Blumenschine, 1994; Blumenschine, 1995; Blumenschine et al., 1996; Capaldo, 1997; Marean et al., 2004; Pickering et al., 2004). Furthermore, students with some training are able to correctly analyze surface modification faunal remains with only incidence lighting and a low power magnification (Blumenschine et al., 1996). Thus, a sample of KEH-1 faunal remains were examined obliquely with a high intensity light source at 15-20x magnification.
Size class and Taxon
Faunal specimens were compared to the extensive collection at the MAP lab to identify size class and taxon. Size classes are based on those developed by (Brain, 1981), for ungulate species. These range from size 1 (2-20 kg), to size 5 (901-2000kg), although size 2 (21-113kg), size 3 (114-340kg) and size 4 (341-900kg) are frequently recovered in archaeological sites (Brain, 1981). Ethnographic and experimental work demonstrates the potential for animals size 1-3 to be carried to the site whole or almost whole (Bunn et al., 1988; O’Connell et al., 1990; Schoville and Otarola-Castillo, 2014). Larger animals tend to be processed in the field, or carried back to site if the group size is large (Schoville and Otarola-Castillo, 2014). Several high survival elements (the head and feet) are less likely to
46


be brought to the site if the distance is far (Schoville and Otarola-Castillo, 2014). However, high utility bones (including the vertebrae, ribs, humerus, and femur) are more likely to be transported (Binford, 1978; Schoville and Otarola-Castillo, 2014). The mix of high and low survival within this category offers a complicated look at hominin transport strategies, and has incurred much debate over early hominin hunting abilities (Blumenschine, 1986, 1989; Stiner and Kuhn, 1992; Marean and Kim, 1998; Pickering et al., 2003; Cleghom and Marean, 2004; Dommguez-Rodrigo et al., 2009b, 2014; Schoville and Otarola-Castillo, 2014). Elements were identified to species whenever possible, and subsequently a Minimum Number of Elements (MNE) was established by counting the number of fragments that overlapped on each bone (Marean et al., 2001). Then Minimum Animal Units (MAU) could be calculated. MAU is calculated on the MNE divided by the number of times the element occurs within the body (Binford, 1978). Thus, it removes the necessity of siding an element to determine how many individuals are represented (Binford, 1978).
Additional anthropogenic modifications
Finally, specimens were assessed for degree of burning. Bones were coded as calcined, carbonized, unaltered, or heat altered based on a visual inspection. As a majority (>9,000) were recovered from the DHA, it was expected that a number would show evidence of fire damage. However, evidence of burning on bone can also be used to track exposure to fire outside of hearth contexts. The placement and percentage of burnt surface may indicate cooking, debris disposal, or accidental/natural conflagration (Andrews and Cook, 1985;
Clark and Ligouis, 2010; Thompson and Henshilwood, 2014a).
47


Statistical analysis
To determine the if the observed patterns between aggregates are statistically significant, SPSS was used to run statistical analyses. To describe the variation in mean length and width, an ANOVA was applied. All 2434 specimens were used for this analysis. The stratigraphic aggregates were recoded (1= OBS, 2= DHA, 3= DHA/DS, 4=DS, AND 5= DBS) in order to run the ANOVA. The ANOVA compared the length while controlling for stratigraphic aggregate, to describe the dispersion around the mean length. To describe which variables were different, Tukey post-hoc test was then run. If there are significant variations between the different stratigraphic aggregates, then it is expected that more post-depositional damage or anthropogenic processing occurred in these levels, suggesting more intense site use. However, if there are no significant variations, then all levels were subjected to similar levels of processing and/or post depositional stress.
A chi-square analysis was used to assess whether there is any significant variation in the type of fracture (nutritive or non-nutritive) of long bones. Only long bones (N= 733) were used in this analysis. The fracture angles were recoded to reflect the number of nutritive breaks per aggregate (OBS= 172, DHA= 307, DHADS= 84, DS= 131, DBS= 7). A Cramer’s V and contingency coefficient post hoc test were run, to describe the extent of the association. Significant differences between the aggregates suggests that long bones were being processed for marrow differently across a shifting environment, triggered by global climatic changes. No significant differences would suggest that early humans/carnivores did not change their marrow processing strategies in response to the expansion and contraction of the Paleo-Agulhas Plain.
48


In order to determine if the number of tooth, percussion, and cut marks significantly vary between aggregates, an ANCOVA was run. This statistic has been used in previous analysis describing the extent of nutritive processing of long bones between climatic cycles (Hodgkins et al., 2016). ANCOVA determines the significance while controlling for multiple variables (in this case, surface area of the bone, calculated by Geometric mean). The sample size was restricted to long bones within the subsample examined for surface modification (N= 102). Given the absence of cut, percussion, and tooth marks in the DHA/DS and DBS aggregates, the analysis was only run on the OBS, DHA, and DS material. Significant differences between the aggregates would demonstrate a difference in human and carnivore behavior across time.
49


CHAPTER V
RESULTS Fragment size
The fauna of KEH-1 is heavily fragmented. There are few whole bones recorded in this analysis (.005%, or 12 of 2400). A majority (97%) have a length <40 mm and width <20 mm (fig. 5.1). The most heavily fragmented aggregates are the OBS, DHA, and DBS (mean length = 12.12, 14.04, and 15.19 mm respectively). The DS has a higher mean length (17.31 mm), while the transitional DHA/DS has a mean of 20.26 mm. While the OBS and DHA have low average length/widths, both also have high dispersals around the mean. An ANOVA demonstrates significance variance in length by level at the a = 0.05 level between several aggregates (F= 38.67, df = 4, p-value = .000). The OBS, DHA, DHA/DS, and DS aggregates are significantly different in length, however, the DBS does not significantly differ from any aggregate except the DHA/DS (table 2). ANOVA results on width (F= 19.31, df = 4, p-value = .000), demonstrate a significant variation between several aggregates (table 3). However, there is no significant difference in width between the OBS and the DHA, or the DHA/DS and the DS. In other words, it appears that aggregates with similar climatic trends (I.e. glacial and interstadial periods) have similar widths. Length appears to be more subjected to the individual aggregate. This suggests that there may be different factors controlling fragmentation temporally.
50


Aggregate Compared aggregate Mean Difference (mm) Standard Error P-value
OBS DHA 2.33 0.52 0
DHA/DS 8.38 0.74 0
DS 4.84 0.6 0
DHA DHA/DS 6.05 0.67 0
DS 2.51 0.5 0
DHA/DS DS 3.53 0.72 0
DBS 5.87 1.89 0.01
Table 2: Tukey’s post hoc test on ANOVA results for length. Only significantly different
results were included.
Aggregate Compared aggregate Mean Difference (mm) Standard Error P-value
OBS DHA/DS 3.11 1.07 0.03
DS 2.92 0.88 0.008
DBS 20.95 2.54 0
DHA DBS 19.86 2.5 0
DHA/DS DBS 17.84 2.6 0
DS DBS 18.03 2.53 0
Table 3: Tukey’s post hoc test on ANOVA results for width. Only significantly different results were included.
Length in mm â–  Width in mm
.a
110 100 90 80 70
•B 60
'$ 50
ft 40 o
>-J 30
20 10 0
<
OBS
6 T+ K
Fragment Size
DHA DHA/DS DS DBS
Stratigraphic Aggregate
51


5.1 Box plot showing the lengths and widths of faunal specimens in each aggregate. Boxes represent one standard deviation and whiskers represent two standard deviations, while individual points represent outliers. The analysis was run on all specimens (N= 2434). For individual aggregate counts refer to table 1 (previous chapter).
This pattern of high fragmentation persisted in the analysis of the circumference of long bone shafts and epiphyses (N= 733). Fewer than 10% of the long bone shafts in each aggregate had a complete (100%) circumference (fig. 5.2). While 70% or greater of each aggregate had less than 50% circumference (i.e. the shaft fragment would not wrap around more than 50% of a long bone), each aggregate had far fewer fragments with a circumference >50% of the shaft. Smaller fragments frequently occur in scenarios where marrow extraction is required (Bunn, 1983; Marean and Spencer, 1991). Previous analysis suggests that intensive human bone processing may result in higher frequencies of >%50 circumference (Bunn, 1983). Overall, both datasets indicate highly fragmented assemblage, across all aggregates. Determining the cause of the fragmentation requires a critical look at the fracture patterns and surface modification, to assess the potential taphonomic, anthropogenic or carnivore actions that occurred at KEH-1.
52


- N/A - <50%
>50% â–  100%
100%
STRATIGRAPHIC AGGREGATE
Circumference of Long Bone Shafts
5.2 Frequency of long bone circumference by aggregate. (N=733)
Element and Animal class representation
Size class
While most of the faunal remains were so badly fragmented as to be unidentifiable, we were able to identify 35% (N= 845) of the specimens to size class (table 4). Size two animals are consistently well-represented across the aggregates, consisting of 30-40% of the identified specimens (fig. 5.3). During the interstadial occupation (~29ka-32 ka), there is less of a reliance on size one (>15%) animals, while greater numbers of both size three and four (>50% combined) are transported to the cave. Both trends are reversed during the subsequent glacial period (~26ka-19 ka). As the frequency of size 4 mammals declines (from 17% to 14 % to 4%), the frequency which size 1 mammals are recovered increases. A notable increase in size one animals occurs during the OBS (from 13% to 38%), at the height of the last glacial maximum, possibly indicating a shift in local environment. This might also be due to
53


carnivore practices, as some of the microfauna in this aggregate have been identified as part of a genet midden (Cleghorn, pers. comm).
5.3 Graph of the frequency of specimens in each size class (N= 845).
Stratigraphic Aggregate Micromammal Size 1 Size 2 Size 3 Size 4 Non-ID Total
OBS 4 71 72 36 7 285 475
DHA 3 39 114 104 41 826 1127
DHA/DS 1 12 61 64 24 101 263
DS 23 54 69 29 361 536
DBS 7 10 6 2 7 32
Total 8 152 311 279 103 1580 2433
Table 4: Specimen counts of size class by aggregate
Number of Individual Specimens (NISP)
The frequency of recovered elements varies across stratigraphic aggregate. Around 39% (N=944) of the assemblage could be identified to skeletal element, and 34% of the assemblage (N=808) are from the high survival set, or elements most likely to survive the taphonomic processes (fig. 5.4). High survival elements are strongly represented across the
54


aggregates, composing 28-53% of the specimens (fig. 5.4). The frequency of low survival elements varied by aggregate. In most aggregates, low survival elements are less than 10% of the specimens, however, these elements make up 31% of the fauna in the DBS.
Unfortunately, the small sample size of the DBS precludes any further conclusions regarding element representation. The DHA has the lowest percentage of low survival elements (3%), possibly indicating differential destruction due to carnivore activity or other post-depositional processes like burning.
70% 767 335
STRATIGRAPHIC AGGREGATE
’ High survival * Low survival * Non-ID ■ No Data
Skeletal element representation by high and low survival from Lam et al. (1998)
5.4 Frequency of high and low survival skeletal elements based on Lam et al. (1988). “No data” refers to specimens for which this analysis was not completed (N= 2434).
There is no difference when comparing the overall frequency of high survival and high utility elements (elements attached to significant portions of meat, and thus more likely to be transported) across the assemblage (fig. 5.4, Appendix A). However, elements which are considered high utility and high survival do not always align (Binford, 1978; Lam et al., 1998; Cleghom and Marean, 2004). Elements included in both categories are: the humerus,
55


radius-ulna, tibia, and femur. In other words, longbones, with their hefty marrow content and robust attached muscles, are more likely to be transported to and recovered from archaeological sites. A complete list of KEH-1 faunal skeletal elements by size and their survival/utility status can be found in appendixes A and B. In each aggregate, longbone fragments (LBFs) form a substantial majority of the identifiable specimens (Appendix A). Every aggregate contains sizable amounts of high utility elements, mostly resulting from the large numbers of LBFs. Ribs are the second most frequently identified element, although they are considered a low survival element. Other low-survival, high-utility elements including scapula, vertebrae, and innominates were recovered (Appendix A), although they make up relatively small percentages of the identified fauna. Low utility bones are present in all aggregates (fig. 5.5, Appendix B), although in low frequencies (1%-18%). These consist of cranial elements and the distal portions of the limbs, neither of which offer significant portions of nutrition (Cleghorn and Marean, 2004).
60% 11
- High Utility Size 1/2 * High Utility Size 3/4 * High Utility size Non-ID
• Low Utility Size 1/2 Low Utility size 3/4 * Low Utility size Non-ID
Skeletal element representation by High and Low utility bones, from Binford (1980)
56


5.5 Frequency of high and low utility elements based on Binford (1980) (N= 2434).
Taxonomic representation
The highly fragmented nature of the assemblage also impacted the ability to taxonomically identify specimens. Approximately 60% of the specimens were only identifiable as mammal (N= 1472), while only 4% could be identified to family or genus. Diverse animal families have been recovered from KEH-1, including Bovidae, Equidae, Carnivora, Aves, Testudinidae, and Petromuridae (table 5). Out of these, bovids are the most frequently identified family group (more than 50% of the identified specimens). Two identifications were made to the genus level, Procavia capensis (in the OBS aggregate), and raphicerus sp. (in the DS aggregate). For MNI and MALI, analysis was restricted to bovids, and the specimens were separated by size class (1 and 2, and 3 and 4) (Appendixes B and C). The small fraction of identifiable specimens hindered any ability to adequately compare specimens (only size 3 and 4 bovids in the DS have a MALI higher than 1), however, there are changes in the MALI between sizes classes across aggregates. These follow the trends discussed in the paragraph above.
57


JNTFR- GLACIAL TRANS. TRANSITION ^AniAl GLACIAL Taxon OBS DHA DHA/DS DS DBS T°tal
Bovid 11 23 1 23 5 63
Equid 1 1
Ungulate 9 1 1 11
Raphicerus sp. 1 1
Carnivore 4 1 5
Procavia capensis 4 4
Aves 6 1 7
Tortoise 2 3 5
Rodent 4 1 5
Microfauna 6 6
Terrestrial Mammal 2 2
Mammal 344 583 258 266 21 1472
No Data 14 10 2 5 1 32
Non-ID 79 499 2 235 4 819
Total 475 1127 263 536 32 2433
Table 5: Num 3er of Indivic ual Specimens by taxonomic representation and by stratigraphic
aggregate.
Burning
The frequency of burned faunal remains varies between aggregate (fig. 5.6). Unsurprisingly, the highest frequencies of burned material were recovered from the DHA, which is comprised of more than fifty stacked hearth units. Around 55% of the faunal material from this aggregate is heat altered or burned. The other glacial aggregate, the OBS, has similarly high frequencies of heat altered bone. However, the specimens in the DBS, DS and DHA/DS have frequencies of burning lower than 40%. The reasons for this pattern are not entirely clear, although excavations suggest that the DBS (glacial) represents a low-intensity occupation (to date, one hearth has been located in this aggregate). Accumulations during the DS appear much more intense but lack the structured and conspicuous hearth units
58


that characterize the DHA. It appears that fire was being utilized during this time, but was either less intense, located elsewhere (possibly down the slope or in the back of the cave), or the evidence of hearths have been destroyed by other processes. It is interesting to hypothesize that fires were being lit on the beach, less than 10 km distant, but the evidence suggests that some level of burning occurred at KEH-1. At any rate, the DHA represents a new use of the cave from the previous interglacial period, and this is reflected in the heat alteration of the fauna. The burning process could have affected the survivability of the bones, possibly contributing to the higher degrees of fragmentation (particularly dry fragmentation) observed in the OBS and DHA.
70%
60%
W
Z
| 50%
HH
U
w 40%
in
W 30%
Z
OS
§ 20% £
10%
0%
249
165
168
503
335
29 29
~ MjT
333
OBS DHA
- Unbumt w Heat altered
DHA/DS
DS
18
3 3
DBS
Calcined â–  Carbonized - No data
Maximum Burning Stage
5.6 Frequencies of maximum burning stage by aggregate (N= 2434).
59


Long Bone survival
In order to determine the degree of nutrient processing, long bone fragments were analyzed for patterns of differential destruction (N= 733). Epiphyseal destruction atKEH-1 is highly variable across the stratigraphic aggregates, reflecting different degrees of post-depositional destruction (fig. 5.7, table 6). Both epiphysis and near epiphysis shafts (sensu Blumenschinne, 1995) include trabecular bone, making them highly susceptible to carnivore consumption or taphonomic destruction (Marean and Spencer, 1991; Marean et al., 2000; Pickering et al., 2004). Low epiphyseal survival relative to shafts in the OBS is consistent with experimental data that includes carnivore ravaging (Marean et al., 2000). Epiphyseal survival in the OBS, and to some extent in the DHA, are consistent with human marrow processing experiments (Marean et al., 2000). However, the other major stratigraphic aggregates (the DBS and DS) and the transitional DHA/DS have much higher rates of epiphyseal survival (fig. 5.7). These high frequencies (greater than 50%) have no experimental analogue, possibly suggesting that shaft fragments might be so fragmented as to be unidentifiable, skewing the data in favor of more identifiable epiphyseal ends. Non-ID fragments make up 30%-63% of the specimens from these aggregates (fig. 5.7). This may also imply that the faunal remains were not exposed to carnivore ravaging.
60


only to only
carnivore KEH-1
EXPERIMENTAL DATA
(Based on Marean et al., 2000) ’shaft ■ near epiphysis 'epiphysis
Long Bone Shaft vs. Epiphysis representation
5.7 Relative frequencies of longbone shaft, near-epiphysis, and epiphyseal fragments (N= 733). Some fragments were identified as longbones, but type was not distinguished.
Aggregate No Data Epiphysis Near Epiphysis Shaft Shaft Total
DBS 3 2 5 10
DHA 4 16 64 199 283
DHA/DS 1 42 47 47 137
DS 2 36 43 73 154
OBS 2 6 18 123 149
Total 9 103 174 447 733
Table 6: Count of Longbone shafts (thick cortical), Near Epiphyses (shafts with some
trabecular bone), and Epiphyses (longbone ends).
Fragmentation morphology
Overall, post-depositional breakage is quite high, generally falling outside the confidence intervals from the experimental data by Marean et al. (2000). However, while
61


most of the aggregate frequencies cluster between 10 and 20%, the DBS has much higher frequencies of transverse fracture outlines (fig. 5.8, table 7). This further suggests that this aggregate was subjected to higher levels of post-depositional destruction, possibly due to trampling and/or compression as the site was occupied for the next 30-40 thousand years. In contrast, nutritive fractures, suggesting marrow consumption by hominins or carnivores, are less frequent at Knysna than the experimental data (fig. 5.9). This may be a result of less intensive marrow processing, or possibly smaller carnivores with reduced jaw strength when compared to hyenas (Hodgkins and Marean, 2017). Nutritive fractures are most frequent in the glacial aggregates where only terrestrial fauna is available, and less frequent during the interstadial, presumably when marine foraging is the dominant subsistence pattern. Conversely, the DBS (glacial) has a lower frequency of green fractures, falling between the DS and DHA/DS. The breakage data suggests that post-depositional damage occurred in all aggregates, and that the long bones were processed to varying extents at different times.
62


35%
30% cn
H 25% O
^ 20% H
H 15% O
? 10%
£
5%
0%
0% 5% 10% 15% 20% % TRANSVERSE OUTLINE 25% 30%
• CARNIVORE ONLY • HUMAN TO CARNIVORE •HUMAN ONLY
•OBS eDHA •DHA/DS
*DS *DBS
Non-nutritive (dry) fracture shape: KEH-1 and experimental data from Marean et al.
(2000)



•
• m
| •
1 L
5.8 Frequency of non-nutritive fractures compared to data from Marean et al. 2000 (N=733).
100%
90%
m 80% W
^ 70%
60%
p 50% O'
3 40%
aa
O 30% ^ 20% 10% 0%
0% 10% 20% 30% 40% 50% 60% % CURVED OUTLINES 70% 80% 90% 100%
• CARNIVORE ONLY •HUMAN TO CARNIVORE •HUMAN ONLY
•OBS •DHA •DHA/DS
*DS ‘DBS
Nutritive (green) fracture shape: KEH-1 and experimental data from Marean et al. (2000)
ji—•—i
i—1


• 1 >

« 1 •



63


5.9 Frequency of the nutritive fractures compared to data from Marean et al. 2000 (N= 733).
— NISP Curved NISP Oblique NISP Transverse NISP Total Bone Ends
OBS 148 172 40 48 149 298
DHA 218 307 93 72 283 566
DHA/DS 55 84 30 27 137 274
DS 86 131 59 49 154 308
DBS 7 7 6 3 10 20
Table 7: Nutrient and Non-nutrient breaks
The distal portions of limbs offer significantly fewer marrow nutrients (Binford,
1980). In order to assess the extent of nutritional stress, Hodgkins et al. (2016) considered the number of distal limb bones which had been processed for marrow. Relatively few bones were identified as metapodial or phalanx in each aggregate (table 8). The DS has the highest portion of nutrient fractures, and none of the distal long bones in the OBS and DS preserve non-nutrient breaks. The DHA has a mix of non-nutrient and nutrient fractures, however the sample size is too small to draw any significant conclusions. The sample size of the DBS is even smaller (N = 3), and the only fractures are non-nutritive. While the sample size is small for all aggregates, it does show that some distal elements were being broken while fresh. To determine the agents responsible for the deposition of these and other long bone elements, we considered the surface modification.
Aggregate NISP Nutrient breaks Non-nutrient breaks Unbroken
OBS 7 57% 43%
DHA 6 60% 20% 20%
DS 8 62% 38%
DBS 3 66% 33%
Table 8: NISP and frequencies of nutritive/non-nutritive fractures for distal limb bones
64


The higher frequencies of green fractures when the coastline was distant (OBS=54% and DHA=55%) are statistically significantly different from frequencies when the coast was near (31% and 43%, X=2804, DF=16, P=0). Cramer’s V and a contingency coefficient post hoc suggest that there is a strong association between the variables (table 9). These results suggest that the long bones were processed or ravaged more intensely during glacial periods, despite the availability of diverse terrestrial resources.
DF P-value Cramer's V Contingency coefficient
2804 16 0 1 0.894
Table 9: Chi-square results.
Matrix cover
A sub-sample of 411 faunal specimens was examined for surface modification. The ability to record surface modification rests on the clarity of the surface. If the surface is obscured by matrix or destroyed by post-depositional processes, determining the agents involved becomes difficult. Within our sample, 54% of the specimens (N= 223) had no matrix cemented to their surface. The DHA has the largest percentage of specimens with some or all surface covered, and the OBS has the fewest (Fig. 5.10). Thus, analysis of the surface modification required removal of specimens which had their surfaces completely obscured.
65


70%
60% 'mm'
M jU70 5 40% W
43 8
h 30%
125 136,-5 17 11 15
O o? 20%
10% / S 1 7pl29 S’
n o/v ~L- , a / -1—
U yO 4 1 1
OBS DHA DS DBS
AGGREGATE
'100% '>50% '<50% 'Nomatrix
Frequency of matrix cover on faunal remains
5.10 Frequency of matrix cover on faunal remains (N= 411).
Surface modification
In order to consider the impact of human/carnivore interactions and the experimental data, all fragments with less than 20% visibility were removed from the analysis, and the comparative analysis was restricted to long bones within the surface subsample (N= 102). Surface modification frequencies (percussion and tooth marks) for stratigraphic aggregates fall well below the confidence intervals from the experimental data by Marean et al. (2000) (Fig. 5.11, table 10). Within our sample, percussion marks on long bones are present in the OBS, DHA, and DS (fig. 5.11). They are highest in the DS (13%). Tooth marks are absent in the DBS and DS, and infrequently present in the OBS and DHA, (6% and 8%, respectively). Tooth marks might result from human seasonal occupation of the cave, allowing carnivore access to the faunal remains. However, the low frequencies may also imply year-round occupation, which would reduce carnivore access to the remains (Hodgkins et al., 2016). Cut
66


marks vary across the assemblage, with the DS having a substantially higher frequency
(40%) than the OBS (3%), DHA (16%) and DBS (0%).
Aggregate / Mark OBS/PM OBS/TM DHA/PM DHA/TM DS/PM DS/TM
Long-Bone Fragment 1 2 1 1
Metapodial 1 1
Tibia 1 1
Total 2 3 1 1 2 0
Table 10: Surface modification on long bones. PM = percussion mark; TM = tooth
mark.
©Hominin 1 •Hominin 2 #Hominm-camivore 1 •Hominin-camivore 2 •Hominm-camivore 3 Carnivore 1
• Carnivore 2 ' Carnivore 3 •OBS
* DHA #DS •DBS
0% 20% 40% 60% 80% 100%
% OF TOOTH MARKS
Consumption: KEH-1 and experimental data from Marean et al (2000)
5.11 Frequencies of tooth and percussion marks compared to data from Marean et al. (2000) (N= 102).
In all instances, the ANCOVA found the number of marks to be non-significant. However, the small number of fragments with surface modifications (PM=5, TM=4, and
67


CM=11) in the sample may suggest that there are insufficient degree of freedom based on the number of variables in this analysis. A larger surface modification sample is needed to tackle the question of how the long bones are being broken.
Aggregate Tooth Percussion Cut N
OBS 6% 3% 3% 63
DHA 8% 8% 16% 12
DS 0% 13% 40% 22
DBS 0% 0% 0% 5
Table 11: frequencies of Lone Bone surface modification within the subsample
10 9 8 7 6 5 4 3 2 1 n ’PM ’CM *TM
ta y

3

r—
O
£ O u 3
2 2
—V 1 1
0 0 0 0
u OBS DHA DS DBS
AGGREGATE
Count of surface modification
Figure 5.12 Count of the specimens with surface modification by aggregate. One specimen in the DS had both a percussion mark and tooth mark (N= 12 in a sample of 102).
68


CHAPTER VI
DISCUSSION
Subsistence strategies of KEH-1
The analyses suggest that climatic shifts impacted human access to terrestrial and marine resources, and thus, how terrestrial resources were utilized. Initially, during the ~46ka -34 ka occupation, early humans hunted terrestrial mammals. A much more intense occupation occurred as the coastline drew near the cave ~34ka -29ka, although foragers intensely utilized the marine resources they continued to prey upon terrestrial mammals. When the coastline shifted, exposing the Paleo-Agulhas, site use intensified, as did use of terrestrial fauna. Foragers acquired more diverse species as the glacial progressed, and there was an increasing the number of smaller animals brought to the site. Additionally, foragers appear to have begun processing their prey more intensely during the Last Glacial Maximum.
69


Glacial-LGM Transition to LGM lnterstadial
~19ka ~22ka-26ka ~29 ka- ~32 ka
Transition to interstadial
-34 ka- >46 ka
G
GO
n
o
U
<10 km
Peak occupation intensity
Tapering occupation Increased occupation Intermittent occupation
intensity Unknown, but receding
Figure 6.1: Coastal distance from KEH-1 based on modeling by Fisher et al. (2010), and changes in the frequencies of the animal size classes, fracture patterns, and surface modification.
Our analysis of size class highlights several trends concerning transport decisions.
We collapsed sizes one/two, and three/four to describe transport decisions. High utility bones (which, in many cases, are also high survival bones) are more likely to be transported during the entire occupation sequence. However, the size of the animal impacted the frequency with which it was brought to the cave. During the DBS, when foragers focused on terrestrial fauna, more size one and two fauna was carried to the cave overall. As the coastline progressed towards the cave, the trend reversed, and the DS has greater frequencies of larger (size three and four) fauna, including low utility elements. During the transitional DHA/DS
70


and into the DHA as the coastline recedes, both size groups are transported frequently, although size three and four are still prioritized. When the coastline reaches nearly 70 kilometers distant, smaller animals were utilized more frequently.
Thus, throughout the Last Glacial Maximum, low utility elements of all sizes were transported at similar frequencies. It is interesting to note the spike in low utility elements of larger size mammals during the interstadial period. It could be that this is a result of “riders,” or elements transported along with the nutrient heavy portions. This is referred to as the “unconstrained transport strategy,” (Faith and Gordon, 2007), and may indicate either a relatively adjacent kill site or a large hunting party transporting all elements to the site (Schoville and Otarola-Castillo, 2014). Currently, the Knysna Heads (<2km west of KEH-1) sit at the intersection of multiple waterways, including an ocean passage through the rocky cliffs on which KEH-1 is perched, the lagoon, and the Knysna river. Phytolith analysis suggests that the vegetation during the interglacial ~34ka -29ka contained large numbers of C4 water-loving species, making it likely that the vegetation and landscape were similar to current conditions (Cleghorn et al., 2018). Possibly a landscape of this nature would have attracted animals and humans, thus making it easier (in terms of distance) to transport all portions of larger carcasses. It is also possible that early humans cleaned the site by discarding the larger pieces of fauna outside the cave. The steep slope leading to the cave would have made it easy to rapidly remove skeletal elements from the cave, especially considering the proximity of the hearths to the entrance.
Nutritional stress and a distant coastline
Were the foragers of KEH-1 nutritionally stressed? Analysis indicates that this is a possibility. Although the increase in smaller prey may be a result of the Genet midden
71


(Cleghorn, pers. comm), cutmarks on one bird and two carnivore elements (~19ka) suggest that humans obtained and consumed meat from diverse sources. Large ungulates are frequently discussed in the context of early human subsistence practices, smaller animals, including birds, are frequently overlooked (Val et al., 2016). It has been argued that populations will intensify their use of smaller animals when the larger ones have been over-exploited (Klein and Cruz-Uribe, 1983; Thompson and Henshilwood, 2014b; Armstrong, 2016). It is possible that an increased taxonomic diversity during the glacial period (~26ka -19ka) suggests that the foragers of KEH-1 were nutritionally stressed. However, it is also possible that this is an opportunistic exploitation of a diversified resource base, as the Paleo-Agulhas plain is exposed. Nearby sites suggest that earlier foragers were utilizing a broad resource base (Wadley, 2010; Thompson and Henshilwood, 2014b; Armstrong, 2016; Val et al., 2016), although in several cases, this has been linked to compensation for loss of marine resources. The apparently reduced capture rate of large mammals later in the sequence is noteworthy, as the Paleo-Agulhas is reconstructed as offering space for large, migratory mammals (Copeland et al., 2016). Thus, if the Last Glacial Maximum offers extensive bioavailable pasture south of the cave, it is curious as to why KEH-1 foragers do not intercept these animals as frequently during the DHA and OBS, as the proceeding interstadial period. The answer may lie in the territoriality of LSA hunter-gatherers. If the Paleo-Agulhas is carved into territories, and the large mammals are migrating E-W (as proposed by Copeland et al., 2016), then the foragers might only hunt them seasonally, and focus on smaller prey at other times. Additionally, it has been proposed that smaller animals (rabbit, tortoise, ect) offer a higher return in balanced micro-nutrients, which would be lacking in
72


larger game (Brown et al., 2011). Smaller game may have offered glacial foragers a more diverse array of nutrients and offset the loss of marine resources.
The fauna was fragmented through a combination of human action and post-depositional processes. Initially, the smaller fragment size (in the DBS) is related to taphonomic effects. However, the subsequent interglacial period has reduced post-depositional breakage, and overall lower frequencies of nutrient fractures. The reduction in length during the glacial period (~26ka -19 ka), results from statistically significantly higher frequencies of nutrient fractures. Some level of non-nutrient breakage is occurring, possibly due to the intensity of site use. Additionally, burned bones may be more susceptible to fragmentation (Stiner et al., 1995). This suggests that there may have been multiple factors influencing the fragmentation of long bones.
The surface modification data presents a more complex picture of the site. Tooth marks are absent from our sample of the DS, while present (albeit not to a large extent) in the DHA and OBS. However, anthropogenic modifications on bones are more frequent (at least double) in the DS. Moreover, some of these occur on the distal portions of the limbs. If early humans were nutritionally stressed, it is expected that they would utilize even the distal elements of the limb, which do not offer large amounts of protein (Hodgkins et al., 2016). It appears that both humans and carnivores are responsible for the nutrient fractures during the Last Glacial Maximum at KEH-1. This appears to be a shift from the interstadial (the DS), when carnivores had less access to the site, and humans appear to be responsible for the nutritive fractures.
Taken together, does the increased focus on smaller animals during the OBS, coupled with more intense nutritive processing of long bones, demonstrate nutritional stress during
73


the glacial period? The previous literature suggests that these are valid proxies for gauging human subsistence strategies (Thompson and Henshilwood, 2014b; Armstrong, 2016; Hodgkins et al., 2016). However, the frequency of anthropogenic surface modification in the interstadial exceeds the amount in the subsequent glacial periods. This specific proxy suggests that hominins were more stressed for nutrients during the interstadial (per Hodgkins et al., 2016). However, it should be noted that surface analysis has only been conducted on a small subsample of the long bone assemblage. An expanded analysis of the surface modification may alter this conclusion (for example, a subsample N= 12 for the DHA when the total number of faunal remains in the DHA >9,000, is not exactly representative). However, if a higher sample size demonstrates that the significantly higher results of nutrient-related fragmentation are more frequently a result of anthropogenic activity, then it would suggest that early humans are more stressed for marrow during the glacial period. Increasing the sample size will reveal higher frequencies of surface modification in all aggregates and permit a more complete understanding of the human subsistence.
The explanation that multiple subsistence strategies were utilized by foragers is certainly plausible. Lower levels of cut/percussion marks in the peak glacial period, interspersed with tooth marks, may suggest that humans were moving across the landscape, but returning consistently to KEH-1. This means that carnivores would have been able to access the site. The location, with diverse mammalian resources, provided substantial shelter and offer significant terrestrial resources for both humans and carnivores. While long bones were processed more frequently, it may be through a combined effort of humans and carnivores. If this is the case, then humans were still highly mobile, possibly traveling from
74


the coastline into the interior and making use of the entire offerings of the Paleo-Agulhas
plain.
75


CHAPTER VII
CONCLUSION
The Paleo-Agulhas plain provided early humans with alternating terrestrial and marine resources. KEH-1 offers a glimpse into human subsistence during a tumultuous climatic period, when the coastline and ecology underwent rapid change. KEH-1 is one of a handful of sites preserving occupational sequences throughout MIS2, providing a unique opportunity to understand early human foraging. Taphonomic analysis shows a significant difference of the fragmentation (size of fauna) between the aggregates. Aggregates with the highest fragmentation are also those with the highest frequency of fire modification. Bones which are high survival (dense and compact) are more frequently identified to element across all aggregates. These bones also contain large amounts of meat/marrow, making them highly desirable (Binford, 1978). Long bone fragments are likewise fragmentated, and during the later sequences (DHA and OBS) this is more frequently associated with fresh breaks, as observed in cases of marrow processing.
This pilot assessment of the fauna provides some compelling initial results. Over time, the fauna brought to the cave trend towards smaller bodied animals. Evidence suggests that the interstadial (late MIS3) offered substantial resources, with shellfish transported to the cave, while foragers continued to consume a variety of mammals of differing body sizes. The increased marrow processing represented in the glacial period is consistent with nutritional stress. Only a small subsample was assessed for surface modification, and the specimens with cut/percussion/tooth marks were so few as to be statistically insignificant. Further research is necessary to clarify the extent to which this represents hominin or carnivore
76


activity. This research can contribute to the reconstruction of hominin behavior on the paleo-
Agulhas plain.
77


REFERENCES
Andrews, P., Cook, J., 1985. Natural Modifications to Bones in a Temperate Setting. Man. 20, 675-691.
Armstrong, A., 2016. Small mammal utilization by Middle Stone Age humans at Die Kelders Cave 1 and Pinnacle Point Site 5-6, Western Cape Province, South Africa. Journal of Human Evolution. 101, 17-44.
Arnold, J., 1992. Complex Hunter-Gatherer-Fishers of Prehistoric California: Chiefs, Specialists, and Maritime Adaptations of the Channel Islands | American Antiquity | Cambridge Core [WWW Document], URL https://www.cambridge.org/core/journals/american-antiquity/article/div-classtitlecomplex-hunter-gatherer-fishers-of-prehistoric-california-chiefs-specialists-and-maritime-adaptations-of-the-channel-islandsdiv/93BE358F07E49A652557DAB7543C9B98 (accessed 4.10.18).
Athreya, S., Ackermann, R., in press. Colonialism and narratives of human origins in Asia and Africa. 10.31730/osf.io/jtkn2.
Barham, L., 2002. Systematic Pigment Use in the Middle Pleistocene of South-Central Africa Current Anthropology: Vol 43, No 1 [WWW Document], URL https://www-journals-uchicago-edu.aurarialibrary.idm.oclc.org/doi/full/10.1086/338292 (accessed 5.6.18)
Bar-Matthews, M., Marean, C.W., Jacobs, Z., Karkanas, P., Fisher, E.C., Herries, A.I., Brown, K.,
Williams, H.M., Bernatchez, J., Ayalon, A., 2010. A high resolution and continuous isotopic speleothem record of paleoclimate and paleoenvironment from 90 to 53 ka from Pinnacle Point on the south coast of South Africa. Quaternary Science Reviews. 29, 2131-2145.
Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. Paleobiology. 4, 150-162.
Behrensmeyer, A.K., Kidwell, S.M., Gastaldo, R.A., 2000. Taphonomy and paleobiology. Paleobiology. 26, 103-147.
Beier, J., Anthes, N., Wahl, J., Harvati, K., 2018. Similar cranial trauma prevalence among Neanderthals and Upper Palaeolithic modern humans. Nature. 563, 686.
Berger, T.D., Trinkaus, E., 1995. Patterns of trauma among the Neandertals. Journal of Archaeological Science. 22, 841-852.
Binford, L.R., 1962. Archaeology as Anthropology. American Antiquity. 28, 217-225.
Binford, L.R., 1978. Nunamiut: Ethnoarchaeology. New York: Academic Press.
78


Binford, L.R., 1980. Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity. 45, 4-20.
Binford, L.R., 1985. Human ancestors: changing views of their behavior. Journal of Anthropological Archaeology. 4, 292-327.
Bird, D., O'Connell, J., 2006. Behavioral Ecology and Archaeology on JSTOR [WWW Document], URL http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/41053227?pq-origsite=summon (accessed 9.14.17).
Blumenschine, R.J., 1986. Early hominid scavenging opportunities: implications of carcass availability in the Serengeti and Ngorongoro ecosystems. British Archaeological Reports.
Blumenschine, R.J., 1989. A landscape taphonomic model of the scale of prehistoric scavenging opportunities. Journal of Human Evolution. 18, 345-371.
Blumenschine, R.J., 1995. Percussion marks, tooth marks, and experimental determinations of the timing of hominid and carnivore access to long bones at FLK Zinjanthropus, Olduvai Gorge, Tanzania. Journal of Human Evolution. 29, 21-51.
Blumenschine, R.J., Marean, C.W., Capaldo, S.D., 1996. Blind tests of inter-analyst correspondence
and accuracy in the identification of cut marks, percussion marks, and carnivore tooth marks on bone surfaces. Journal of Archaeological Science. 23, 493-507.
Bousman, C.B., 2005. Coping with risk: Later stone age technological strategies at Blydefontein Rock Shelter, South Africa. Journal of Anthropological Archaeology. 24, 193-226.
Bousman, C.B., Brink, J.S., 2018. The emergence, spread, and termination of the Early Later Stone Age event in South Africa and southern Namibia. Quaternary International, In Honour of James Brink. 495, 116-135.
Brain, C.K., 1969. The contribution of Namib Desert Hottentots to an understanding of
australopithecine bone accumulations. Scientific Papers of the Namib Desert Research Station. 1969, 13-22.
Brain, C.K., 1981. The hunter or the hunted. An Introduction to African Cave Taphonomy, The University of Chicago press, Chicago et Londres.
Brown, J.L., 1964. The evolution of diversity in avian territorial systems. The Wilson Bulletin. 160-169.
Brown, K., Fa, D.A., Finlayson, G., Finlayson, C., 2011. Small game and marine resource exploitation by Neanderthals: the evidence from Gibraltar. In: Trekking the Shore. Springer, pp. 247-272.
Brumfiel, E.M., 1992. Distinguished Lecture in Archeology: Breaking and Entering the Ecosystem-Gender, Class, and Faction Steal the Show. American Anthropologist. 94, 551-567.
79


Brumm, A., Moore, M.W., 2005. Symbolic revolutions and the Australian archaeological record. Cambridge Archaeological Journal. 15, 157-175.
Bunn, H.T., 1983. Comparative analysis of modern bone assemblages from a San hunter-gatherer
camp in the Kalahari Desert, Botsuwana, and from a spotted hyena den near Nairobi, Kenya. Animal and Archaeology, 1. Hunters and Their Prey; BAR International Series. 163, 141-148.
Bunn, H.T., Bartram, L.E., Kroll, E.M., 1988. Variability in bone assemblage formation from Hadza hunting, scavenging, and carcass processing. Journal of Anthropological Archaeology. 7, 412-457.
Butzer, K.W., 1973. Geology of Nelson Bay Cave, Robberg, South Africa. The South African Archaeological Bulletin. 28, 97-110.
Capaldo, S.D., 1997. Experimental determinations of carcass processing by Plio-Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai Gorge, Tanzania. Journal of Human Evolution. 33, 555-597.
Capaldo, S.D., Blumenschine, R.J., 1994. A quantitative diagnosis of notches made by hammerstone percussion and carnivore gnawing on bovid long bones. American Antiquity. 59, 724-748.
Chase, B.M., Meadows, M.E., 2007. Late Quaternary dynamics of southern Africa's winter rainfall zone. Earth-Science Reviews. 84, 103-138.
Clark, A., 1997. The MSA/LSA Transition in Southern Africa: New Technological Evidence from Rose Cottage Cave on JSTOR [WWW Document], URL
http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/3889076?casa_token=TkfUt6ILJXYA AAAA:mMeocsx4xrZiFb5RDnU81SidGfWnx7jkXoQRuub8XjlhlLXh_ED-kthRODqjOi5-uXgJM7EWzF7PCIAXyoUw9HiF5EQoQ6tXHLk4WBQOilxZo3g_wQ (accessed 3.19.18).
Clark, J.L., 2017. The Howieson's Poort fauna from Sibudu Cave: Documenting continuity and change within Middle Stone Age industries. Journal of Human Evolution. 107, 49-70.
Clark, J.L., Kandel, A.W., 2013. The Evolutionary Implications of Variation in Human Hunting Strategies and Diet Breadth during the Middle Stone Age of Southern Africa | Current Anthropology: Vol 54, No S8 [WWW Document], URL https://www-journals-uchicago-edu.aurarialibrary.idm.oclc.org/doi/abs/10.1086/673386 (accessed 4.1.18).
Clark, J.L., Ligouis, B., 2010. Burned bone in the Howieson's Poort and post-Howieson's Poort Middle Stone Age deposits at Sibudu (South Africa): behavioral and taphonomic implications.
Journal of Archaeological Science. 37, 2650-2661.
Clark, J.L., Plug, I., 2008. Animal exploitation strategies during the South African Middle Stone Age: Howiesons Poort and post-Howiesons Poort fauna from Sibudu Cave. Journal of Human Evolution. 54, 886-898.
Cleghorn, N., Keller, H., McGrath, J.R., Shelton, C.M., Watson, S., Schoville, B.J., Peart, D., Bickerton, N., Esteban, I., 2018. Assessing the development of societal complexity at the Middle to
80


Later Stone Age transition in the context of the Economic Defensibilty Model: Evidence from Knysna, coastal South Africa.
Cleghorn, N., Marean, C.W., 2004. Distinguishing selective transport and in situ attrition: a critical review of analytical approaches. Journal of Taphonomy. 2, 43-67.
Collins, B., Steele, T.E., 2017. An often overlooked resource: Ostrich (Struthio spp.) eggshell in the archaeological record. Journal of Archaeological Science: Reports. 13, 121-131.
Colonese, A.C., Mannino, M.A., Bar-Yosef Mayer, D.E., Fa, D.A., Finlayson, J.C., Lubell, D., Stiner, M.C., 2011. Marine mollusc exploitation in Mediterranean prehistory: An overview. Quaternary International, Shell Midden Research: An Interdisciplinary Agenda for the Quaternary and Social Sciences. 239, 86-103.
Compton, J.S., 2011a. Pleistocene sea-level fluctuations and human evolution on the southern coastal plain of South Africa. Quaternary Science Reviews. 30, 506-527.
Compton, J.S., 2011b. Pleistocene sea-level fluctuations and human evolution on the southern coastal plain of South Africa. Quaternary Science Reviews. 30, 506-527.
Copeland, S.R., Cawthra, H.C., Fisher, E.C., Lee-Thorp, J.A., Cowling, R.M., le Roux, P.J., Hodgkins, J., Marean, C.W., 2016. Strontium isotope investigation of ungulate movement patterns on the Pleistocene Paleo-Agulhas Plain of the Greater Cape Floristic Region, South Africa. Quaternary Science Reviews. 141, 65-84.
d'Errico, F., Backwell, L., 2016. Earliest evidence of personal ornaments associated with burial: The Conus shells from Border Cave. Journal of Human Evolution. 93, 91-108.
d'Errico, F., Backwell, L., Villa, P., Degano, I., Lucejko, J.J., Bamford, M.K., Higham, T.F.G., Colombini, M.P., Beaumont, P.B., 2012a. Early evidence of San material culture represented by organic artifacts from Border Cave, South Africa. Proceedings of the National Academy of Sciences. 109, 13214-13219.
d'Errico, F., Moreno, R.G., Rifkin, R.F., 2012b. Technological, elemental and colorimetric analysis of an engraved ochre fragment from the Middle Stone Age levels of Klasies River Cave 1, South Africa. Journal of Archaeological Science. 39, 942-952.
Deacon, H.J., 1993. Planting an Idea: An Archaeology of Stone Age Gatherers in South Africa. The South African Archaeological Bulletin. 48, 86-93.
Deacon, J., 1978. Changing patterns in the late Pleistocene/early Holocene prehistory of Southern Africa as seen from the Nelson Bay Cave Stone artifact sequence. Quaternary Research. 10, 84-111.
D'errico, F., Henshilwood, C., Nilssen, P., 2001. An engraved bone fragment from c. 70,000-year-old Middle Stone Age levels at Blombos Cave, South Africa: implications for the origin of symbolism and language. Antiquity. 75, 309-318.
81


Dibble, H.L., Chase, P.G., McPherron, S.P., Tuffreau, A., 1997. Testing the reality of a "living floor" with archaeological data. American Antiquity. 62, 629-651.
Dominguez-Rodrigo, M., Bunn, H.T., Mabulla, A.Z.P., Baquedano, E., Uribelarrea, D., Perez-Gonzalez, A., Gidna, A., Yravedra, J., Diez-Martin, F., Egeland, C.P., Barba, R., Arriaza, M.C., Organista,
E., Anson, M., 2014. On meat eating and human evolution: Ataphonomic analysis of BK4b (Upper Bed II, Olduvai Gorge, Tanzania), and its bearing on hominin megafaunal consumption. Quaternary International, The Evolution of Hominin Behavior during the Oldowan-Acheulian Transition: Recent Evidence from Olduvai Gorge and Peninj (Tanzania). 322-323, 129-152.
Dominguez-Rodrigo, M., de Juana, S., Galan, A.B., Rodriguez, M., 2009a. Anew protocol to
differentiate trampling marks from butchery cut marks. Journal of Archaeological Science.
36, 2643-2654.
Dominguez-Rodrigo, M., Mabulla, A., Bunn, H.T., Barba, R., Diez-Martin, F., Egeland, C.P., Espilez, E., Egeland, A., Yravedra, J., Sanchez, P., 2009b. Unraveling hominin behavior at another anthropogenic site from Olduvai Gorge (Tanzania): new archaeological and taphonomic research at BK, Upper Bed II. Journal of Human Evolution. 57, 260-283.
Dominguez-Rodrigo, M., Musiba, C.M., 2010. How accurate are paleoecological reconstructions of early paleontological and archaeological sites? Evolutionary biology. 37, 128-140.
Dusseldorp, G.L., 2010. Prey Choice During the South African Middle Stone Age: Avoiding Dangerous Prey or Maximising Returns? African Archaeological Review. 27, 107-133.
Dusseldorp, G.L., Langejans, G.H.J., 2013. Carry that weight: coastal foraging and transport of
marine resources during the South African Middle Stone Age. Southern African Humanities. 25, 105-135.
Enloe, J.G., 2012. Middle Palaeolithic Cave Taphonomy: Discerning Humans from Hyenas at Arcy-sur-Cure, France. International Journal of Osteoarchaeology. 22, 591-602.
Faith, J.T., 2008. Eland, buffalo, and wild pigs: were Middle Stone Age humans ineffective hunters? Journal of Human Evolution. 55, 24-36.
Faith, J.T., 2011. Ungulate community richness, grazer extinctions, and human subsistence behavior in southern Africa's Cape Floral Region - ScienceDirect [WWW Document], URL http://www.sciencedirect.com.aurarialibrary.idm.oclc.org/science/article/pii/S00310182110 02276 (accessed 12.7.17).
Faith, J.T., 2013. Taphonomic and paleoecological change in the large mammal sequence from Boomplaas Cave, western Cape, South Africa. Journal of Human Evolution. 65, 715-730.
Faith, J.T., Gordon, A.D., 2007. Skeletal element abundances in archaeofaunal assemblages:
economic utility, sample size, and assessment of carcass transport strategies. Journal of Archaeological Science. 34, 872-882.
82


Faith, J.T., Thompson, J.C., 2018. Low-survival skeletal elements track attrition, not carcass transport behavior in Quaternary large mammal assemblages. In: Zooarchaeology in Practice.
Springer, pp. 109-126.
Faure, H., Walter, R.C., Grant, D.R., 2002. The coastal oasis: ice age springs on emerged continental shelves. Global and Planetary Change. 33, 47-56.
Fisher, E.C., Bar-Matthews, M., Jerardino, A., Marean, C.W., 2010. Middle and Late Pleistocene
paleoscape modeling along the southern coast of South Africa. Quaternary Science Reviews. 29, 1382-1398.
Flannery, K.V., 1968. Archaeological systems theory and early Mesoamerica. Anthropological archaeology in the Americas. 67, 87.
Gifford-Gonzales, D., Damrosch, David, Damrosch, Debra, Pryor, J., Thunen, R., 1985. The Third
Dimension in Site Structure: An Experiment in Trampling and Vertical Dispersal | American Antiquity | Cambridge Core [WWW Document], URL
https://www.cambridge.org/core/journals/american-antiquity/article/div-classtitlethe-third-dimension-in-site-structure-an-experiment-in-trampling-and-vertical-dispersaldiv/0F9657D5DAA7DAlF1946B0E5FD3A9C09# (accessed 9.7.17).
Grine, F.E., Wurz, S., Marean, C.W., 2017. The middle stone age human fossil record from Klasies River main site. Journal of human evolution. 103, 53-78.
Hardy, B.L., Moncel, M.-H., 2011. Neanderthal Use of Fish, Mammals, Birds, Starchy Plants and Wood 125-250,000 Years Ago. PLOS ONE. 6, e23768.
Haws, J.A., Funk, C.L., Benedetti, M.M., Bicho, N.F., Daniels, J.M., Minckley, T.A., Denniston, R.F.,
Jeraj, M., Gibaja, J.F., Hockett, B.S., 2011. Paleolithic landscapes and seascapes of the west coast of Portugal. In: Trekking the Shore. Springer, pp. 203-246.
Helm, C.W., McCrea, R.T., Cawthra, H.C., Lockley, M.G., Cowling, R.M., Marean, C.W., Thesen,
G.H.H., Pigeon, T.S., Hattingh, S., 2018. A New Pleistocene Hominin Tracksite from the Cape South Coast, South Africa. Scientific Reports. 8, 3772.
Henshilwood, C.S., d'Errico, F., Watts, I., 2009. Engraved ochres from the Middle Stone Age levels at Blombos Cave, South Africa. Journal of Human Evolution. 57, 27-47.
Henshilwood, C.S., D'errico, F., Marean, C.W., Milo, R.G., Yates, R., 2001. An early bone tool industry from the Middle Stone Age at Blombos Cave, South Africa: implications for the origins of modern human behaviour, symbolism and language. Journal of Human Evolution. 41, 631-678.
Hodgkins, J., Marean, C.W., 2017. New information from old excavations: A comparative analysis of Paleolithic zooarchaeological assemblages from the Zagros Mountains. In: Biglari F, Mashkour M, Shidrang S (Eds.) The Pleistocene Archaeology of the Iranian Plateau, Iraq and the Caucasus, National Museum of Iran Paleolithic Studies Series. National Museum of Iran, Tehran.
83


Hodgkins, J., Marean, C.W., Turq, A., Sandgathe, D., McPherron, S.J.P., Dibble, H., 2016. Climate-mediated shifts in Neandertal subsistence behaviors at Pech de I'Aze IV and Roc de Marsal (Dordogne Valley, France). Journal of Human Evolution. 96, 1-18.
Hodgkins, J., Roux, P. le, Marean, C.W., Penkman, K., Crisp, M., Fisher, E., Lee-Thorp, J., 2018. The
role of ostrich in shaping the landscape use patterns of humans and hyenas on the southern coast of South Africa during the late Pleistocene. Multispecies Archaeology. 333-346.
Hublin, J.-J., Ben-Ncer, A., Bailey, S.E., Freidline, S.E., Neubauer, S., Skinner, M.M., Bergmann, I., Le
Cabec, A., Benazzi, S., Harvati, K., 2017. New fossils from Jebel Irhoud, Morocco and the pan-African origin of Homo sapiens. Nature. 546, 289.
Jacobs, Z., 2010. An OSL chronology for the sedimentary deposits from Pinnacle Point Cave 13B—a punctuated presence. Journal of Human Evolution. 59, 289-305.
Jacobs, Z., Roberts, R., Galbraith, R., Deacon, H.J., Grun, R., Mackay, A., mitchell, P., Vogelsang, R., Wadley, L., 2008. Ages for the Middle Stone Age of Southern Africa: Implications for Human Behavior and Dispersal | Science [WWW Document], URL http://science.sciencemag.org/content/322/5902/733 (accessed 5.21.18).
Jerardino, A., Marean, C.W., 2010. Shellfish gathering, marine paleoecology and modern human behavior: perspectives from cave PP13B, Pinnacle Point, South Africa. Journal of Human Evolution. 59, 412-424.
Karkanas, P., Brown, K.S., Fisher, E.C., Jacobs, Z., Marean, C.W., 2015. Interpreting human behavior from depositional rates and combustion features through the study of sedimentary microfacies at site Pinnacle Point 5-6, South Africa. Journal of Human Evolution. 85, 1-21.
Kelly, R.L., 1995. The foraging spectrum. Washington, DC: Smithsonian Institution Press.
Kidwell, S.M., Holland, S.M., 2002. The quality of the fossil record: implications for evolutionary analyses. Annual Review of Ecology and Systematics. 33, 561-588.
Klein, R., 1976. The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape Province, South Africa on JSTOR [WWW Document], URL
http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/3887730 (accessed 5.27.18).
Klein, R.G., 1989. Why does skeletal part representation differ between smaller and larger bovids at Klasies River Mouth and other archeological sites? Journal of Archaeological Science. 16, 363-381.
Klein, R.G., 1994. Southern Africa before the iron age. Integrative paths to the past: Paleoanthropological advances in honor of F. Clark Howell. 471-519.
Klein, R.G., 1995. Anatomy, behavior, and modern human origins. Journal of World Prehistory. 9, 167-198.
84


Klein, R.G., 2000. Archeology and the evolution of human behavior. Evolutionary Anthropology: Issues, News, and Reviews. 9, 17-36.
Klein, R.G., Avery, G., Cruz-Uribe, K., Halkett, D., Parkington, J.E., Steele, T., Volman, T.P., Yates, R., 2004. The Ysterfontein 1 Middle Stone Age site, South Africa, and early human exploitation of coastal resources. Proceedings of the National Academy of Sciences. 101, 5708-5715.
Klein, R.G., Avery, G., Cruz-Uribe, K., Steele, T.E., 2007. The mammalian fauna associated with an archaic hominin skullcap and later Acheulean artifacts at Elandsfontein, Western Cape Province, South Africa. Journal of Human Evolution. 52, 164-186.
Klein, R.G., Cruz-Uribe, K., 1983. Stone Age Population Numbers and Average Tortoise Size at
Byneskranskop Cave 1 and Die Kelders Cave 1, Southern Cape Province, South Africa. The South African Archaeological Bulletin. 38, 26-30.
Klein, R.G., Cruz-Uribe, K., 1987. Large mammal and tortoise bones from Eland's Bay Cave and
nearby sites, Western Cape Province, South Africa. Papers in the prehistory of the western Cape, South Africa. 1, 132-163.
Klein, R.G., Steele, T.E., 2013. Archaeological shellfish size and later human evolution in Africa. Proceedings of the National Academy of Sciences. 110, 10910-10915.
Kusimba, S.B., 2005. What Is a Hunter-Gatherer? Variation in the Archaeological Record of Eastern and Southern Africa. Journal of Archaeological Research. 13, 337-366.
Lam, Y.M., Chen, X., Marean, C.W., Frey, C.J., 1998. Bone Density and Long Bone Representation in Archaeological Faunas: Comparing Results from CT and Photon Densitometry. Journal of Archaeological Science. 25, 559-570.
Lam, Y.M., Pearson, O.M., Smith, C.M., 1996. Chin morphology and sexual dimorphism in the fossil hominid mandible sample from Klasies River Mouth. American Journal of Physical Anthropology: The Official Publication of the American Association of Physical Anthropologists. 100, 545-557.
Lee-Thorp, J.A., Ecker, M., 2015. Holocene Environmental Change at Wonderwerk Cave, South
Africa: Insights from Stable Light Isotopes in Ostrich Eggshell. African Archaeological Review. 32, 793-811.
Loftus, E., Sealy, J., Lee-Thorp, J., 2016. New radiocarbon dates and Bayesian models for Nelson Bay Cave and Byneskranskop 1: Implications for the South African later stone age sequence. Radiocarbon. 58, 365-381.
Lyman, R.L., 1984. Bone density and differential survivorship of fossil classes. Journal of Anthropological archaeology. 3, 259-299.
Lyman, R.L., 2004. The Concept of Equifinality in Taphonomy. Journal of Taphonomy. 2, 15-26.
Lyman, R.L., Lyman, C., 1994. Vertebrate Taphonomy. Cambridge University Press.
85


Mackay, A., 2010. The Late Pleistocene Archaeology of Klein Kliphuis Rock Shelter, Western Cape, South Africa: 2006 excavations. The South African Archaeological Bulletin. 65, 132-147.
Mackay, A., Jacobs, Z., Steele, T.E., 2015. Pleistocene Archaeology and Chronology of Putslaagte 8 (PL8) Rockshelter, Western Cape, South Africa. Journal of African Archaeology. 13, 71-98.
Mackay, A., Welz, A., 2008. Engraved ochre from a Middle Stone Age context at Klein Kliphuis in the Western Cape of South Africa. Journal of Archaeological Science. 35, 1521-1532.
Mallick, S., Li, H., Lipson, M., Mathieson, I., Gymrek, M., Racimo, F., Zhao, M., Chennagiri, N.,
Nordenfelt, S., Tandon, A., 2016. The Simons genome diversity project: 300 genomes from 142 diverse populations. Nature. 538, 201.
Marean, C.W., 2010. Pinnacle Point Cave 13B (Western Cape Province, South Africa) in context: The Cape Floral kingdom, shellfish, and modern human origins. Journal of Human Evolution. 59, 425-443.
Marean, C.W., 2011. Coastal South Africa and the Coevolution of the Modern Human Lineage and the Coastal Adaptation. In: Trekking the Shore, Interdisciplinary Contributions to Archaeology. Springer, New York, NY, pp. 421-440.
Marean, C.W., 2014. The origins and significance of coastal resource use in Africa and Western
Eurasia. Journal of Human Evolution, The Role of Freshwater and Marine Resources in the Evolution of the Human Diet, Brain and Behavior. 77, 17-40.
Marean, C.W., 2016. The transition to foraging for dense and predictable resources and its impact on the evolution of modern humans. Phil. Trans. R. Soc. B. 371, 20150239.
Marean, C.W., Abe, Y., Frey, C.J., Randall, R.C., 2000. Zooarchaeological and taphonomic analysis of the Die Kelders Cave 1 Layers 10 and 11 Middle Stone Age larger mammal fauna. Journal of Human Evolution. 38, 197-233.
Marean, C.W., Abe, Y., Nilssen, P.J., Stone, E.C., 2001. Estimating the minimum number of skeletal elements (MNE) in zooarchaeology: a review and a new image-analysis GIS approach. American Antiquity. 66, 333-348.
Marean, C.W., Bar-Matthews, M., Bernatchez, J., Fisher, E.C., Goldburg, P., Herries, A.I.R., Jacobs, Z., Jerardino, A., Karkanas, P., Minichillo, T., Nilssen, P., Thompson, E., Watts, I., Williams, H., 2007. Early human use of marine resources and pigment in South Africa during the Middle Pleistocene | Nature [WWW Document], URL https://www.nature.com/articles/nature06204 (accessed 5.6.18).
Marean, C.W., Cawthra, H.C., Cowling, R.M., Esler, K.J., Fisher, E., Milewski, A., Potts, A.J., Singels, E., Vynck, J.D., 2014. Stone Age people in a changing South African Greater Cape Floristic Region. In: Allsopp, N., Colville, J.F., Verboom, G.A. (Eds.), Fynbos. Oxford University Press, pp. 164-199.
86


Marean, C.W., Dominguez-Rodrigo, M., Pickering, T.R., 2004. Skeletal element equifinality in
zooarchaeology begins with method: the evolution and status of the "shaft critique." Journal of Taphonomy. 2, 69-98.
Marean, C.W., Kim, S.Y., 1998. Mousterian Large-Mammal Remains from Kobeh Cave Behavioral Implications for Neanderthals and Early Modern Humans. Current Anthropology. 39, S79-S114.
Marean, C.W., Spencer, L.M., 1991. Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity. 56, 645-658.
McBrearty, S., Brooks, A.S., 2000. The revolution that wasn't: a new interpretation of the origin of modern human behavior. Journal of Human Evolution. 39, 453-563.
Mellars, P., Stringer, C., 1989. The human revolution. Edinburgh University Press Edinburgh.
Milo, R.G., 1998. Evidence for Hominid Predation at Klasies River Mouth, South Africa, and its
Implications for the Behaviour of Early Modern Humans. Journal of Archaeological Science. 25, 99-133.
Mitchell, J.P., 1994. Understanding the MSA/LSA transition: the pre-20 000 BP assemblages from new excavations at Sehonghong:rock shelter, Lesotho. Southern African Field Archaeology.
3, 15-25.
Mitchell, P., 2008. Developing the archaeology of marine isotope stage 3. Goodwin Series. 52-65.
Nettle, D., Gibson, M.A., Lawson, D.W., Sear, R., 2013. Human behavioral ecology: current research and future prospects. Behavioral Ecology. 24, 1031-1040.
Nowell, A., 2010. Defining Behavioral Modernity in the Context of Neandertal and Anatomically Modern Human Populations | Annual Review of Anthropology [WWW Document], URL https://www.annualreviews.org/doi/abs/10.1146/annurev.anthro.012809.105113 (accessed 4.9.18).
O'Connell, J.F., Hawkes, K., Jones, N.B., 1990. Reanalysis of large mammal body part transport among the Hadza. Journal of Archaeological Science. 17, 301-316.
Opperman, H., 1978. Excavations in the Buffelskloof Rock Shelter near Calitzdorp, Southern Cape. The South African Archaeological Bulletin. 33, 18-38.
Owen, R.B., Muiruri, V.M., Lowenstein, T.K., Renaut, R.W., Rabideaux, N., Luo, S., Deino, A.L., Sier, M.J., Dupont-Nivet, G., McNulty, E.P., 2018. Progressive aridification in East Africa over the last half million years and implications for human evolution. Proceedings of the National Academy of Sciences. 201801357.
Parkington, J., Poggenpoel, C., Halkett, D., Hart, T., 2004. Initial observations on the Middle Stone Age coastal settlement in the western Cape, South Africa. Settlement dynamics of the middle paleolithic and Middle Stone Age. 2, 5-51.
87


Pickering, T.R., Dominguez-Rodrigo, M., Egeland, C.P., Brain, C.K., 2004. New data and ideas on the foraging behaviour of Early Stone Age hominids at Swartkrans Cave, South Africa : research article. South African Journal of Science. 100, 215-219.
Pickering, T.R., Marean, C.W., Dominguez-Rodrigo, M., 2003. Importance of limb bone shaft
fragments in zooarchaeology: a response to "On in situ attrition and vertebrate body part profiles" (2002), by M.C. Stiner. Journal of Archaeological Science. 30, 1469-1482.
Pierce, G.J., Ollason, J.G., 1987. Eight Reasons Why Optimal Foraging Theory Is a Complete Waste of Time. Oikos. 49, 111-118.
Porraz, G., Parkington, J.E., Rigaud, J.-P., Miller, C.E., Poggenpoel, C., Tribolo, C., Archer, W.,
Cartwright, C.R., Charrie-Duhaut, A., Dayet, L., Igreja, M., Mercier, N., Schmidt, P., Verna, C., Texier, P.-J., 2013. The MSA sequence of Diepkloof and the history of southern African Late Pleistocene populations. Journal of Archaeological Science, The Middle Stone Age at Diepkloof Rock Shelter, Western Cape, South Africa. 40, 3542-3552.
Potts, R., Shipman, P., 1981. Cutmarks made by stone tools on bones from Olduvai Gorge, Tanzania. Nature. 291, 577.
Rector, A.L., Reed, K.E., 2010. Middle and late Pleistocene faunas of Pinnacle Point and their paleoecological implications. Journal of Human Evolution. 59, 340-357.
Renfrew, C., 1996. The sapient behaviour paradox: how to test for potential. Modelling the early human mind. 11-14.
Schoville, B.J., Otarola-Castillo, E., 2014. A model of hunter-gatherer skeletal element transport: The effect of prey body size, carriers, and distance. Journal of human evolution. 73, 1-14.
Scott, L., Neumann, F.H., 2018. Pollen-interpreted palaeoenvironments associated with the Middle and Late Pleistocene peopling of Southern Africa. Quaternary International. 495, 169-184.
Sealy, J., 2006. Diet, Mobility, and Settlement Pattern among Holocene Hunter-Gatherers in Southernmost Africa. Current Anthropology. 47, 569-595.
Smith, E.I., Jacobs, Z., Johnsen, R., Ren, M., Fisher, E.C., Oestmo, S., Wilkins, J., Harris, J.A., Karkanas, P., Fitch, S., Ciravolo, A., Keenan, D., Cleghorn, N., Lane, C.S., Matthews, T., Marean, C.W., 2018. Humans thrived in South Africa through the Toba eruption about 74,000 years ago. Nature.
Steele, T., Mackay, A., Fitzsimmons, K., 2016. Varsche Rivier 003: A Middle and Later Stone Age Site with Still Bay and Howiesons Poort Assemblages in Southern Namaqualand, South Africa. 64.
Steele, T.E., Klein, R.G., 2013. The Middle and Later Stone Age faunal remains from Diepkloof Rock Shelter, Western Cape, South Africa. Journal of Archaeological Science, The Middle Stone Age at Diepkloof Rock Shelter, Western Cape, South Africa. 40, 3453-3462.
88


Full Text

PAGE 1

THE SUBSISTENCE STRATEGIES OF MIDDLE AND LATER STONE AGE FORAGERS DURING INTER STADIAL /GLACIAL TRANSITIONS AT KNYSNA, SOUTH AFRICA b y HANNAH MAY K ELLER B.A., University of Texas at Arlington A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment Of the requirements for the degree of Master of Arts Anthropology Program 2019

PAGE 2

i © 2019 HANNAH MAY K ELLER ALL RIGHTS RESERVED

PAGE 3

ii This thesis for the Master of Arts Anthropology degree by Hannah May Keller Has been approved for the Anthropology Program By Jamie M. Hodgkins, Chair Christopher Beekman Naomi Cleghorn Anna Warrener Date: May 18 th , 2019

PAGE 4

iii Keller, Hannah May . (M.A., Anthropology Program) The Subsistence strategies of Middle and Later Stone Age foragers during Inter stadial /Glacial transitions in Knysna, South Africa Thesis directed by Assistant Professor Jamie M. Hodgkins ABSTRACT Knysna Eastern Heads Cave 1 (KEH 1) demonstrates varying human occupation over the course of multiple ocean transgressions and regressions in the late Middle Stone Age and early Later Stone Age (46,000 to 18,000 Cal BP). The position of the sites on the sh allow South African coastal shelf offered occupants potential access to riverine, coastal, and terrestrial resources , as high and low sea stands drastically altered the location of these resources. It has been hypothesized that the dearth of sites dated to MIS 2 on the south coast, a time when the ocean was 75 90 km distant, is due to a preference for dense and high protein marine resources, drawing early modern humans out onto the now submerged Paleo Agulhas plain. Possible explanations for the recurring occupation at KEH 1 during this time are 1) that use of marine foods was only one of several subsistence strategies employed by early humans; 2) the intersection of terrestrial and riverine resources around Knysna rival ed coastal resources ; and 3) competi tion and territoriality caused by dense marine resources on the coast forced groups into the interior. Z ooarchaeological analysis of faunal remains deposited during the MIS 3 2 transition at KEH 1 examined variation in the intensity of long bone marrow pro cessing as a proxy for changing nutritional stress. Approximately 2400 faunal specimens were analyzed across the sequence. Although the degree of post depositional fragmentation varies throughout the sequence, the high rate of long bone

PAGE 5

iv fragmentation (84% with an incomplete circumference) at KEH 1 is most likely caused by human processing, with minimal carnivore activity. The higher frequencies of green fractures (54% ) when the coastline is distant are statistically significant ly different P =0) from the frequencies of green bone fractures when the coast is near ( 31 and 43%) . These results suggest that the long bones were processed more intensely during glacial periods, despite the availability of diverse terrestrial resources. The form and contract of this publication are approved. I recommend its publication. Approved: Jamie M. Hodgkins

PAGE 6

v ACKNOWL E DGMENTS M any p eople supported my crazy journey to wards finishing this thesis. F oremost among them is my advisor and mentor Dr. Jamie Hodgkins, who has offered continuous encouragement , guidance , and advi c e in this and so numerous other decisions . S econdly, Dr. Naomi Cle g horn, who first invited me to South Africa and introduced me to the wonders of Paleolithic archaeology and zoological archaeology . N aomi also took the time to d irect my interest into the first of m any proposals and check my faunal identifications. W ithout them, this thesis would not be a reality. Dr s. Christopher Beekman and Anna Warrener, who have helped guide me through the thesis pro cess and have been gracious enough to answer my numerous questions . T he CU Denver A nthropology department, especially C o nnie Turner, Dr. Tammy Stone and Dr. Charles Musiba, for their assistance . A nd to my fellow graduate students (especially Emily), for keeping me s ane (ish) . And finally, to Naomi Keller , for always listening .

PAGE 7

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ... 1 II. THE MARINE ADAPTATION AND IMPACTS ON TERRITORIALITY .......... 5 Behavioral Modernity and the Coastal Adaptation ................................ ............ 5 Subsistence and Territoriality ................................ ................................ .......... 1 5 A case study: Knysna Eastern Heads cave 1 ................................ ................... 2 1 III. SITUATING KEH 1 WITHIN THE SOUTH AFRICAN MIDDLE AND LATER STONE AGE ................................ ................................ ............................ 2 6 MSA and LSA sites of South Africa ................................ ............................... 2 6 Eastern cape ................................ ................................ ......................... 2 7 Western cape ................................ ................................ ....................... 2 9 Environmental and Ecological Reconstruction ................................ ............... 31 Subsistence Strategies of the CFR ................................ ................................ ... 3 6 KEH 1 ................................ ................................ ................................ .............. 3 9 Stratigraphy ................................ ................................ .......................... 40 IV. METHODS OF DATA COLLECTION AND ANALYSIS ................................ 4 2 Tap honomy ................................ ................................ ................................ ...... 4 3 Assessing Nutrient Processing ................................ ................................ ........ 4 4

PAGE 8

vii Size Class and Taxon ................................ ................................ ....................... 4 6 Additional Anthropogenic Modifications ................................ ....................... 4 7 Statistical A nalysis ................................ ................................ .......................... 4 8 V. R ESULTS ................................ ................................ ................................ .............. 50 Fragment Size ................................ ................................ ................................ . 50 Element and Animal class Representation ................................ ....................... 48 Size class ................................ ................................ ............................. 53 Number of Individual Specimens (NISP) ................................ ............ 54 Taxonomic Representation ................................ ................................ ............. 5 7 Burning ................................ ................................ ................................ ............ 5 8 Long Bone Survival ................................ ................................ ........................ 60 Fragmentation morphology ................................ ................................ .............. 61 Matrix cover ................................ ................................ ......................... 65 Surface modification ................................ ................................ ........................ 6 6 Statistical analysis ................................ ................................ ........................... 61 VI. DISCUSSION ................................ ................................ ................................ ........ 69 Subsistence Strategies of KEH 1 ................................ ................................ ..... 6 9 Nutritional Stress and a Distant Coastline ................................ ...................... 71 VII. CONCLUSION ................................ ................................ ................................ ...... 75 REFERENCES ................................ ................................ ................................ .................... 77

PAGE 9

viii APPENDIX A. Number of In dividual S pecimens Present ................................ .............................. 90 B. Minimum Num ber of Elements ................................ ................................ ............. 9 4 C. M i nimum Animal Units ................................ ................................ ........................ 9 6

PAGE 10

ix LIST OF TABLES TABLE 1. The faunal specimens of KEH 1 ................................ ................................ ............ 4 2 2. Tukey s post hoc on length ................................ ................................ .................... 51 3. Tukey s post hoc on width ................................ ................................ ..................... 51 4. S ize class by aggregate ................................ ................................ .......................... 54 5. Taxonomy ................................ ................................ ................................ .............. 5 8 6. Count of long bone shafts and epiphyses ................................ ............................... 61 7. Nutrient and Non nutrient breaks ................................ ................................ .......... 64 8. NISP and frequencies of nutritive/non nutritive fractures for distal limb bones ... 6 4 9. Chi square results ................................ ................................ ................................ ... 6 5 10. S urface modification of long bones ................................ ................................ ....... 6 7 11. Frequencies of Long Bone surface modification ................................ ................... 6 8

PAGE 11

x LI ST OF FIGURES FIGURE 2.1 KEH 1 in relation to significant MSA and LSA sites ................................ ..... 2 2 3.1 Map of Southern Africa ................................ ................................ .................. 2 7 3.2 Photo of the KEH 1 stratigraphy ................................ ................................ .... 41 4.1 Non nutritive fractures (left) and nutritive fractures (right) ........................... 4 5 5.1 Box plot showing the lengths and widths of faunal specimens in each aggregate ................................ ................................ ................................ ..................... 51 5.2 Graph of the frequency of long bone circumference ................................ ...... 53 5.3 Graph of the frequency of size classes ................................ ............................ 54 5.4 Graph of the frequency of high and low survival elements, based on ................................ ................................ ................................ ....... 5 5 5.5 Graph of the frequency of high and low utility elements, b ased on Binford, 1980 ................................ ................................ ................................ ................................ ...... 5 6 5.6 Graph of the frequencies of maximum burning stages by aggregate .............. 5 9 5.7 Graph of the frequencies of maximum burning stages by aggregate .............. 61 5.8 Frequency of the non nutritive fractures, compared to data from Marean et al., 200 0 ................................ ................................ ................................ ....... 63 5.9 Frequency of the nutritive fractures, compared to data from Marean et al., 2000 ................................ ................................ ................................ ................................ ...... 63

PAGE 12

xi 5.10 Graph of the frequency of matrix cover on faunal remains ............................. 6 6 5.11 Frequencies of tooth and percussion marks, compared to data from Marean et al., 2000 ................................ ................................ .............................. 6 7 5.12 Count of surface modification by aggregate ................................ ................... 6 8 6.1 Co astal distance from KEH 1 a nd results of fa unal analysis .......................... 6 8

PAGE 13

1 CHAPTER I INTRODUCTION The transition from the Middle to Later S tone A ge in South Africa is poorly understood, due to the paucity of sites (Mitchell, 2008; Bousman and Brink, 2018) . Researchers grappling with this shift have constructed a framework to describe the behavior of the early modern humans, but how these so called modern behaviors are defined and when they appear archaeologically is contested (Binford, 1985; Tattersall, 1995; Klein, 2000; McBrearty and Brooks, 2000; Brumm and Moore, 2005; Nowell, 2010) . By characterizing a suite of behaviors and linking them to proxies in material culture, archaeologists have attempted to draw out the factors associat ed with the transition to the Later Stone Age (LSA), and how this differs from earlier Middle Stone Age (MSA) behavior (Klein, 2000; McBrearty and Brooks, 2000; Nowell, 2010) . A Human Behavioral Ecology (HBE) framework utilizes an adaptati ve, environment al lens to describe changes in human behavior , and the resulting material culture represented in the archaeological record (Binford, 1980; . The manifestations of sym bolic behavior are suggested to be linked with the development of territoriality and the necessity of signaling between an in and out group (Brumm and Moore, 2005; Nowell, 2010; Compton, 2011a; Marean, 2014) . In S outh Africa, the development of these signals is argued to have begun ~160 ka and may be connected to a preference for stable marine resources, i.e. a coasta l adaptation (Marean, 2014, 2016) . However, access to these dense and predic t able marine resources is impacted by glacial cycles, which result ed in expansion and contraction of the coastline and the Paleo Agulhas plain (Fisher et al., 2010; Compton, 2011a; Marean, 2014, 2016; Will et al., 2016) . Sites located on the current coastline frequently have occupation

PAGE 14

2 histories closely linked to these cycles, and the shifts between available marine and terrestrial subsistence (Butzer, 1973; Klein, 1976; Klein and Cruz Uribe, 1987; Sealy, 2006; Bar Matthews et al., 2010; Jacobs, 2010; Jerardino and Marean, 2010; Steele and Klein, 2013; Marean et al., 2014) . Discussions of hu man behavior can thus be situated in the broader Paleo climatic context, while proxie s for territoriality may bring insight to the impacts of this alternating landscape on foragers . A newly discovered site, Knysna Eastern Heads cave 1 (KEH 1), may offer a deeper understanding of human behavior during an interglacial and glacial transition . Excavations have revealed a reccuring and dense occupation sequence, ~46 ka 19 ka , revealing a site that abutted the ocean during interglacial cycles , and at t imes of glaciation sat over look ing the Paleo Agulhas coastal plain. Researchers have hypothesized that the dearth of sites during the MSA LSA transition was caused by a preference for nutrient heavy marine resources, drawing hunter gathers out onto the Pal eo Agulhas plain (Mitchell, 2008; Faith, 2013) . However, the occupation of KEH 1 during global glacial cycles wh en the ocean was distant , may have implications for this hypothesized preference for marine resource s . One option is the possible development of territoriality surrounding these resources, forcing other groups away from the presumably desirable coast. However, p aleo ecological reconstructions of the Paleo Agulhas Plain suggest a fertile grassy plain (Rector and Reed, 2010; Copeland et al., 2016) . Conversely, the exposed coastal plain and the position of the cave (~23 meters above sea level), may have offered early humans access to herds of ungulates residing in the plain . Thus, an a lternate theory, that early humans turned to marine foods when the paleo Agulhas plain shrunk and terrestrial bovids vanished (Compton, 2011a; Faith, 2013) , is a possibility. In order to determine if occupation of KEH 1 may have varied in desirability as the ecology

PAGE 15

3 shifted, proxies fo r nutrient stresses were examined. In Pleistocene Europe, analysis of the nutrient processing of long bones for marrow demonstrated a significant difference in Neanderthal behavior between interglacial and glacial periods (Hodgkins et al., 2016 ) . Increased marrow processing when the coastline and marine resources are readily available may indicate stresses due to loss of terrestrial biomass and increased population pressures. Conversely, higher frequencies of marrow processing when the coastl ine is distant may indicate an increased preference for marine foods, and territoriality around these resources. T herefore, a nalysis of the faunal remains of KEH 1 have the potential to reveal the subsistence strategies during these cycles . In order to assess if early humans experience nutritional stress during rapid climatic events during MIS 3 and into MIS 2 , the skeletal remains of animals that were transported to KEH 1 for butchering were analyzed. L ong bones were of particular interest because they contain meat and marrow, and the y preserve better over time than bones in the axial skeleton . For taphonomic reasons , given the high fragmentation of the KEH 1 faunal assemblage, the angle and outline of the long bone fragments were assessed, to determine the timing (pre or post depositional) as a proxy for bones which have been broken to retrieve nutrients. Additionally, surface modification (percussion, tooth and cut marks) of long bones allowed an assessment of the agents (individuals of various species with potential to alter or delete bones) involved. This, coupled with an understanding of the broader spatial and temporal shifts across Southern Africa , will contextualize the behavior and subsistence strategies of humans as they interact with a shifting landscape. This thesis will first review the literature on Human Behavioral Ecology, and its implications for behavioral modernity and a marine adaptation. It will consider how these

PAGE 16

4 previous researchers in southern Africa have discussed these ideas, and how the archaeological record supports these theories. The second chapter will discuss the South African MSA LSA archaeological record, consider proxies for environmental reconstruction and early human subsistence, and contextualize the site of Knysna Eastern Heads cave 1. It will also review the methodology used in zooarchaeological analysis and examine previous research on taphonomy and anthropog enic behaviors. Following this, it will di scuss the results of this analysis, and investigate the implications for human behavior at KEH 1.

PAGE 17

5 CHAPTER II THE MARINE ADAPTATION AND IMPACTS ON TERRITORIALITY Behavioral Modernity and the Coastal Adaptation Hu man B ehavioral E cology (HBE) is the framework used in this analysis of KEH 1. field, advocating for studying cultures as a means of adaptation (Binford, 1962; Flannery, 1968) . Proponen ts of HBE argue for a consideration of fitness in studies of human behavior . Human organisms become adapted to certain environments, and it is these adaptations which provide differential reproductive successes . Witho ut the variation in environments, different cultural behaviors (as seen in material culture) would not have arisen . Binford (1962) characterizes this as an energetic tradeoff, whereby certain behaviors (for example, tool use) are more adaptive in certain environments . Certain South African micro technologies (the Still Bay and How iesons Poort industries) are argued to appear and disappear with environmental shifts (Deacon, 1978; Scott and Neumann, 2018) , and possibly the subsequent variation in available fauna (Clark, 2017) . The subsistence strategy of the group is critical, given that energetic requirements have an impact on reproductive success, and many models are constructed around determining the availability of different foods, and how human s acqui red them (Binford, 198 . More specifically, these models analyze how humans use various technologies (and how these technologies shift in response to environment), and how subsistence strategies impact energetic expenditures of the group .

PAGE 18

6 Optimal foraging strategy is a frequently utilized HBE model. This model assumes that the individual will make prey consumption choices based on a ranked scale: if I found a low ranked but availab le rabbit, should I take the time to kill and process the animal, or search for a more desirable, higher ranked deer? . In order to be the most efficient (balance energy expended by energy gained), hunters would desire to (Binford, 1980; Bird and . Shifts in environment or population increase may result in intensification of resources , which often implies that humans are making a greater effort to obtain less desirable foods (Binford, 1980; Klein and Cruz Uribe, . Altho ugh o ptimal foraging theory has been subject of criticism, given that it applies a linear, teleological aspect to human decision making, and assumes that genetic and physical adaptation s are geared to problem solving (Pierce and Ollason, 1987) , it represents a scientific, environmentally derived approach to understanding modern humans . As I will explain below, many researchers of the South African MSA and critical to early human behavior and modernity. Archaeologists studying Early Anatomically Modern Humans (AMH) often try to trace the origins of our species discussion of the assumptions, problems, and ideas behind this phrase is crucial to understanding the literature on AMH in South Africa. Researchers in paleoanthropology struggle to overcome personal and field wide biases, frequently inserting a racial, colonial ist narrative into their reconstructions of early human evolution (Athreya and Ackermann, in press) . The impacts of these biases can be traced throughout the history of

PAGE 19

7 paleoanthropology, and still retain a position in the field today (Athreya and Ackermann, in press) . Many of these inferences regarding behavioral modernity revolve around the supposed cognitive and cultural su periority of European populations (Athreya and Ackermann, in press) . Immediately following the onset of processual theory, many researchers lin ked the (Binford, 1985; Mellars and Stringer, 1989; Klein, 1994, 1995, 2000; Tattersall, 1995 , but see Mcbrearty and Brooks, 2000) . By a rguing that these cultural changes (~50 ka ) were tied to genetic mutations (proposals made before recent a dvances in ancient genomic sequencing), they focused on a seemingly abrupt innovative transformation (McBrearty and Brooks, 2000; Nowell, 2010) . Following subsequent excavations in Africa researchers began arguing for earlier, more gradual innovations that resulted in behavioral modernity (McBrearty and Brooks, 2000) . Additionally, the earliest date for modern humans is now ~300 ka (Hublin et al., 2017) , creating a larger discrepancy between anatomical modernity and behavioral modernity . T he relatively short period of time (~ 50 ka ) following the proposed AMH genetic changes has been criticized as insufficient for incorporation into the human genome and inconsistent with curr ent knowledge about the migratory history of human populations (Renfrew, 1996) . Intere stingly, in 2016 geneticists surveying modern human genomes found no evidence of genes that coded for better cognitive abilities during the last 100Ka (Malli ck et al., 2016) . However, researchers are still split on the origins and implications of behavioral modern ity (Brumm and Moore, 2005; Nowell, 2010) . Traits that researchers included in behav ioral modernity are abstract thinking, planning, innovation (behavioral, cultural, or technological) and symbolism (McBrearty and

PAGE 20

8 Brooks, 2000; Nowell, 2010) . Thus, a rchaeological proxies of behavioral modernity incorporate archaeological and ethnographic studies of ecology, symbolism, technology, and social networking (McBrearty and Brooks, 2000; Nowell, 2010) . Ear ly populations are frequently described as lacking certain aspects : absence of art, symbolism, structures, complex burials, and instead consisted of like unstandardized tools (Klein, 2000) . However, this concept has been closely tied to cognitive capacity, and has (Brumm and Moore, 2005; Nowell, 2010) . A sudden temporally constrained these a rchaeological p roxies for m odernity might not relate to cognitive capacity per say, but instead to soci o cultural constructs and demographic densities which amplified or more frequently left symbolic or cultural traces in the record (Brumm and Moore, 2005) . Mcbrearty and Brooks also note that a perception of gradual changes to modern behavior may faci litate a closer inspection of the proximal causes of such shifts, especially when placed in proper temporal and spatial settings (2000). In the genetic model, behavioral modernity does not occur in the south African record until ~50 ka 40 ka (Klein, 2000) . It has been argued that before 50 ka 40 ka humans used simplistic subsistence strategies , for example, scavenging or hunti ng animals perceived as less dangerous (Klein, 2000) . Klein (2000) has proposed that an advantageous mutation occurred in the human linea ge around 50 ka , which allowed humans to adapt to almost any environment behaviorally, while retaining their morphology (Klein, 2000) . Kle in has used this hypothesized unique cognitive mutation to explain the complex material culture and technology found in Europe and Africa following 40 ka , a time when Homo sapiens were probably the only hominin in the area (Klein, 2000) . Klein argues that i f the species Homo

PAGE 21

9 sapiens is defined by an advanced cognitive mutation , then the origin of modern Homo sapiens should be restricted to the end of the Pleistocene . Klein used the idea of a unique cognitive capacity in H. sapiens to explain how modern human s colonize d the globe (Klein, 2000) , although ignoring earlier migrations of Homo . understand when studying the MSA/LSA tran sition because it has implications for subsistence . Namely, MSA populations have been considered incapable of hunting varied or dangerous prey, while LSA hunters were top of the food chain. These change s in hunting ability are the crux of the argument for a later modern behavioral shift (Klein, 1995, 2000; Klein et al., 2007) . However, o ther researchers debate the separation of anatomical and behavioral modernity. Analysis of numerous S outh African MSA sites (including Blombos, Border cave, Die Kelders, Klasies River Mouth, and Deipkloof ) suggest that early modern humans were fully capable of complex hunting strategies, and targeted dangerous prey (Faith, 2008) . Both the MSA and LSA have comparable assemblage evenness (suggesting similar capabilities) and MSA populations had a wider diet breadth than their successors (Faith, 2008) . MSA hunters targeted so called dang erous prey (buffalo and pig) as frequently as LSA hunters (Faith, 2008) . Given this evidence, i f behavioral modernity is predicated on a subsistence strategy that includes varied prey and complex decisions, then MSA populations should be considered modern (Faith, 2008) . MSA populations have also been shown to consume prey which are s maller or more difficult to capture (Wad ley, 2010; Thompson and Henshilwood, 2014a, 2014b) . Frequent utilization of tortoises at several sites appears to suggest a focus on diverse prey prior to the LSA (Thompson and Henshilwood, 2014b) . The

PAGE 22

10 capture of small bovids has been argued to demonstrate a broad spectrum of hunting techniques, particularly the p otential for snares (Wadley, 2010) . Some research has argued that a critical aspect of behavioral modernity in S outh Africa included an early coastal adaptation (Jerardino and Marean, 2010; Compton, 2011a; Marean, 2011, 2016; Will et al., 2016) . The site of Pinnacle Point (PP) showcases this coastal adaptation, offering evidence of marine utilization as early as ~164 ka (Jerardino and Marean, 2010) . In addition to repeated consumption of shellfish, there appears to be evidence for the collection of aesthetic, non dietary shells, suggesting a n appreciation for beauty which may be linked to symbolism (Jerardino and Marean, 2010) . Evidence from PP demonstrates that the site was frequently occupied while the coast was close (~12km) to the shoreline, (Fisher et al., 2010) . This close proximity led researchers to tie human occupation to the availability of marine resources (Jerardino and Marean, 2010) . It has been argued that the decision to settle on the coast would not have occurred until humans could utilize marine resources (otherwise, the group loses 50 percent of their foraging area to water ), and that this required an abstract, complex thought process (Marean, 2011) . It is argued that o ccupants needed to understand numerous variables, including tidal shifts, to safely and efficiently gather these resources (Jerardino and Marean, 2010; Marean, 2011) . Two models have been put forth to explain the nec essity and the timing of a coastal adaptation. Both focus on the role of nutrition and cognitive capacity (Will et al., 2016) . The first one, by Parkington et al. (2004) emphasizes the role of nutrition in encephalization (Will et al., 2016) . The rich marine diet would have offered better resources for women and infants, allowing for an increase in cranial capacity (Will et al., 2016) . Compton (2011) also notes that the easily digestible, protein rich, fatty acids and nutrients would have contributed

PAGE 23

11 to spec the second model posits that humans already had sufficient cranial capacity to connect tidal variation to lunar activity and efficient collection of marine resources, resu lting in a relatively late adaptation ~164 ka (Marean, 2011) , given that the earliest humans are now dated to 300 ka (Hublin et al., 2017) . However, there is evidence of Neanderthal use of marine and aquatic resources throughout the middle and late Pleistocene (Stringer et al., 2008; Brown et al., 2011; Colonese et al., 2011; Hardy and Monc el, 2011; Haws et al., 2011) , although it has been argued by Marean (2014), that this does not constitute a true marine adaptation, given the temporally and spatially scattered Neanderthal sites with marine subsistence. The extent to which this signal may be a result of taphonomic processes (including destructive ocean progressions during interglacial cycles) is debated (Brown et al., 2011; Haws et al., 2011; Marean, 2014; Will et al., 2019) . A more recent review, which incorporated the data from excavations in north Africa, and compared them to southern African AMH and Neanderthals, came to a similar conclusion as Marean (Will et al., 2019) . Although there appear to be differences in the degree that marine subsistence is utilized, the authors argue that the temporal and spatial distribution of this behavior suggests that marine resources were important for the genus Homo (Will et al., 2019) . Consumption of aquatic resources (specifically fish and other small aquatic animals) by earlier hominins has been hypothesized, although the ephemeral nature of these remains makes it difficult to determine the extent or possibility of such behavior (Stewart, 2010, 2014) . However, in an earlier paper , Will et al. (2016) arg ue s that neither of those models concerning cognition are sufficient to explain the causes of a coastal adaptation from a long term, evolutionary perspective. They extensively review the published data on marine

PAGE 24

12 period of more than 100,000 years in multiple areas of Africa (pg. 8). The authors note that the evidence points to a repeated use of coastal sites, only when the ocean was within 10km of the site, and were likely a result of scheduled visits, not opportunistic patterns (Will et al., 2016) . Lithic technology does not appear to shift dramatically in response to a marine adapt ation, although there may be shifts in how it was organized (Will et al., 2016) . Will et al. (2016) explicitly tie adaptation and the consumption of marine foods together, arguing that it increased the survival potential of each child. Furthermore, by buffering the population when terrestrial foods were scarce, it reduced mortality rates, especially for at risk individuals (Will et al., 2016) . This would have provided an advantage to coastal fo ragers, and influenced their social and cultural strategies (Will et al., 2016) . Mobility has been theorized to play a critical role in the marine adaptation. For example, Compton (2011) argues that ocean regression and tr ansgressions isolated populations on the paleo Agulhas plain (Compton, 2011a) . Specifically, t he steep Cape Folded Belt (a mountainous range edging the southern plain) would have prevented movement inland, especially during interglacials, when rising ocean levels blo cked the lower elevation passes on the eastern and western ends of the Belt (Compton, 2011a) . If this transpired, then d uring glacial periods, AMH may have followed herds onto the wetter plain, only to become trapped on the Plain due to floods during glacials (Compton, 2011a) . This may have also brought them in contact with the coast, resulting in adaptation and consumption of the high protein marine resources (Marean, 2010; Compton, 2011a) . Marean (2011) has argued tha t random use of the coastal resources would be inefficient, and habitual settlements or monthly visits (in connection with lunar impacts on tides) would have allowed

PAGE 25

13 AMH to gain the most energy from this adaptation. There also appears to be a link between proximity to the coast, with certain sites being abandoned as the ocean receded, only to be re inhabited when the shoreline again grew close (Will et al., 2016) . Glacial periods caused increased aridity on the African inter ior (Owen et al., 2018) . It has been hypothesized that as oceans retreated, the newly emerged shelves would be a source of fresh water (F aure et al., 2002). On the Paleo Agulhas Plain, this is the result of rivers running off the mountains and cutting through the Plain. The slope of the continental shelf would have drawn fresh groundwater in the form of an oasis, attracting hominins and mam mals (Faure et al., 2002) . Thus, these now flooded shelves would have been attractive to early humans. An intensive marine subsistence strategy might have developed i n AMH due to the environmental factors. If AMH are being driven to the floodplains due to repeated aridity, perhaps this environment created the right conditions for a coastal adaptation to develop. It has been argued that a marine adaptation led to terri toriality (Kelly, 1995; Sealy, 2006; Marean, 2016) . Marine resources are dense and predictable, and thus desirable, and groups will engage in conflict to maintain their rights to a source (Marean, 2014, 2016) . This does not appear to apply to seasonal resources because it is expect that foragers would move on after the season ends , however, S outh African coastal resources can be acquired year round (Marean, 2014) . Other aquatic resources, including lacustrine and riverine, would also be worth defending, and offer an advantage (Marean, 2014) . Sealy (2006) argues for an shelter. The resources offered by coastal proximi ty appear to result in a territorial organization, and year round occupation , as determined by an isotopic analysis of early

PAGE 26

14 human dental remains (Sealy, 2006) . Although, the shifting coastline during the Pleistocene would have impacted these patterns (Will et al., 2016) . Significant behavioral and symbolic shifts are postulated to be linked to the marine adaptation. Early modern humans may have become more cooperative, extending trade networks (Marean, 2014) symbolic behaviors (Marean, 2014) . Groups trapped on a rapidly vanishing plain may have needed a way to di fferentiate or signal to others the territory (Compton, 2011a) . Arnold (1992) has argued that environmental variability and subsistence stress led to increased specialization and appearance of elites. Disruption of marine resources during warmer periods when the coastline was near caused political and economic power to be concentrated, and a reduced access to resources guided individuals to produce surplus goods for trade to the mainland (Arnold, 1992) . Compton (2011) also postulates that symbolic goods were used for trade in MSA population during resou rce scarcity during ocean regression , although he does not substantiate these claims . However, t here is a growth of symbolic items associated with the MSA, particularly in South Africa, including bone tools (Henshilwood et al., 2001) , engraved ostrich eggshell (Texier et al., 2010, 2013) , engraved ochre (Mackay and Welz, 2008; Henshilwood et al., 2009) , engraved bone , burial with grave goods Backwell, 2016) , and use of ochre as a pigment (Barham, 2002; Wadley et al., 2004; Marean et al., 2007) . These have been argued to demonstrate early modern human cognitive capacity. If these culturally complex behaviors occurred consistently and repeatedly in the r ecord, then it appears that early modern humans had significant sophisticated mental capabilities (per (Brumm and Moore, 2005) , and numerous researchers have argued to this effect (McBrearty

PAGE 27

15 and Brooks, 2000; Faith, 2008; Compton, 2011a; Marean, 2011; Will et al., 2016) . Whether this occurred prior to a marine adaptation or following swiftly after is still a topic of debate. Howev er, symbolic proxies become a more frequent find in conjunction with the often densely occupied coastal cave sites. This suggests a shift in the social and cultural aspects of early modern humans, which fluctuated but remained relatively stable until the M SA LSA transition. While researchers incorporate discussions of symbolism and its significance in the archaeological record, they nearly always do so from a processual viewpoint. The socio cultural shifts and symbolism are perceived as adaptive, offering e arly modern humans a way to cope with their environment, adapt to new subsistence strategies, and position or arrange themselves relative to other groups in an advantageous manner. Subsistence and territoriality If there are groups position ing them sel ves relative to desirable resources , this raises the possibility that there are other groups consuming fewer desirable resources. The economic defensibility model demands conflict when the resource is large and stable enough that it is worth the time and e nergy budget to esta blish a territory (Brown, 1964) . Aggressiveness is perceive d as an adaptive quality, given that it would secure access to resources (Brown, 1964) . The establishment of territoriality as an adaptation has been de picted as initiating a violent suite of characteristics in Homo sapiens , which also coincided kin (per Marean 2016) . However, th e territoriality model also assumes that all groups desire the same resources . F urthermore, it assumes that males but not females evolve to be aggressive, or that evolution in driven by male priorities , as suggested by Marean (2016) , when he discusses the possibility of male aggression revolving round females . Therefore, components of and

PAGE 28

16 approaches to this model are likely biased by researchers. It has also been argued that adaptations canno o (Pierce and Ollason, 1987) . However, if marine resources were more nutritious and if they were easier to acquire then terrestrial ones , then it seems reasonable to infer that humans would value and defend these resources. Although there may be other constructs or lens through which to view this decision, a reduced mortali ty rate would mean that the exploitation of marine resources would have favored populations targeting or sampling these resources by expanding their genetic pool . Hunting of large game animals at close range is hypothesized to contribute to the traumatic i njuries on early Homo species (Berger and Trin kaus, 1995; Trinkaus, 2012; Beier et al., 2018) , although by the late Pleistocene the risk is considered to be mitigated by MSA technology (Dusseldorp, 2010 ) . While collecting shellfish has the potential to be dangerous, knowledge of the tidal movements would mitigate much of that risk (Marean, 2014) , further increasing the benefits of this resource . While significant research has been directed at understanding the benefits of a marine adaptation, relativel y little work has been done on those NOT consuming these rich resources, although it is implied that they existed. Determining the impacts of territoriality in conjunction with a marine adaptation can provide insight into the potential impact on individual s . T he processual viewpoint tends to describe things at a larger scale, generally focusing on interactions of groups from a broader perspective. Given the often poor temporal resolution of sites, and general paucity of sites, describing a smaller scale is difficult. Other researchers have critiqued this viewpoint for failing to incorporate a more nuanced discussion. For example, Brumfiel (1992) argues that analyzing populations and behavior obscures gender. If gender roles are not explicitly defined, they tend to be assumed, which

PAGE 29

17 relegates women to lesser status and obscures their agency (Brumfiel, 1992) . Human skeletal remains are the most effective archaeological proxy for determining these effects (Ubelaker and Ows ley, 2003; Sealy, 2006) , however, the relative lack of remains dating to the MSA in South Africa, coupled with the incompleteness of the remains renders this difficult (Grine et al., 2017) . Therefore, other proxies are necessary to examine territoriality , and will not likely provide the resolution necessary to discuss the impacts at the level of the individual . Previous res earchers have considered how territoriality impacts hunter gatherers. In considering ideas of hunter (1985) question the assumptions of researchers. They argue that simply characterizing groups as simple and complex d oes not allow for nuanced and contextualized studies , an observation which should be applied even to modern studies . Kusimba (2005) notes that ethnographic data of hunter gatherers may not be applicable, given the social stratification of modern groups. Some of this ties back to ideas of behavioral complexity, and deciding when and how Homo sapiens shifted socially and culturally (Kusimba, 2005) . Arnold (1992) has explored how environmental and social stress led to political opportunism and development of elites. The difficulties of food procurement on North American islands following loss of marine resource s required managerial systems to obtain other subsistence (Arnold, 1992) . A centralized power was more efficient, creating surplus and avoiding further stresses (Arnold, 1992) . This provided an adaptive advantage, because it enabled the survival of this group. In the LSA, territoriality might be tied to p opulation expansion. At Byneskranskrop 1 and Die Kelders 1, an LSA (13,000 250) and MSA/LSA site (~75 50 ka , and 2000 1500), Klein and Cruz Uribe analyzed the size of the tortoise bones. There is a change is terrestrial faunal material throughout the seque nces, and a shift in proportion of marine fauna at DK1

PAGE 30

18 (Klein and Cruz Uribe, 1983) . They note a significant reduction in tortoise size during th e LSA, and conclude that this is most likely related to an increased human population (Klein and Cruz Uribe, 1983) . Subsequent studies of shellfi sh and tortoises at Ysterfontein appear to demonstrate the same population increase (Klein et al., 2004) . Shellfish appear to undergo a reduct ion in size in several coastal sites during this time ( Klein and Steele, 2013) . The authors argue that this is due to changes in intensity of collection, because individuals would be more likely to grab larger tortoises before utilizing smaller ones (Klein et al., 2004) , and over intensification would result in smaller shellfish (Klein and Steele , 2013) . If these proxies for population expansion are correct, then it suggests pressure on high protein resources. The desirable marine resources may be under particularly high stress, if territoriality has been established and less mobility is occurr ing, due to denser populations. The nutritional value of tortoises is high enough that, given an optimal foraging theory framework, they should be taken when encountered (Thompson and Henshilwood, 2014a, 2014b) . Tortoises, have a small body size, require little effort to capture and process, they are not dangerous, thus they are a unique example of optimal prey (Thompson and Henshilwood, 2014b) . It has been argued that during the coastal regression at Bloombos and Pinnacle Point, tortoises would have become a more important resource to offset the loss of high protein shellfish (Thompson and Henshilwood, 2014b) .They also represent a resource that can be collected by women and children, who require high caloric yields, especially during periods of pregnancy and lactation (Thompson and Henshilwood, 2014b) . Unlike large mammals, which are linked to ideas of prestige and costly signaling, tortoises and small mammals might offer a supplementary, steady resource to offset the low capture rates of the largest mammals (Thompson and Henshilwood, 2014b; Armstrong, 2016) . If so, these

PAGE 31

19 resources could offer insight i nto the practices of individuals who are generally perceived as having a lower rank (due to potential inability to capture highly ranked prey) (Wadley, 2010; Thompson and Henshilwood, 2014b) . If rank is tied to territoriality , then discussion of how rank might be enacted in th e MSA and LSA should be a subject of further research , although proxies for this are difficult to substantiate . Wadley (2010) argues that the use of snares is related to social demands. Snares inual observation, and had higher catch rates than big game hunting (Wadley, 2010) . Ethnographic data includes observ ation of frequent capture through snares of steenbok and grey duiker, animals which are ubiquitous in the archaeological record (Wadley, 2010) . Wadley suggest s delayed gratification, and the ability to conceptualize animal mobility patterns linked to capture technolog y. She also suggests proxies for snare technology in the MSA, including high proportions of difficult to capture, diverse prey, and no evidence of raptor consumption (Wadley, 2010) . Even if small prey provided regular food, this r esource is assumed to lack social status (Wadley, 2010) . However, small prey capture has a much higher success rate, suggesting that it would offer stability for families (Wadley, 2010) . Although r ank mig ht be more fluid and negotiable in foraging populations, this small prey/large prey dichotomy may suggest differential access to resources within the group. Unfortunately, determining which gender was hunting which animals and the consequent social implic a tions may be little more than conjecture, given the taphonomic concerns of the Pleistocene archaeological record. These socio cultural and economic pressures may suggest the necessity for risk management. Wiessner (1982) ar gues for a consideration of risk management in terms of spatial layout. If the site is open, then individuals can observe who needs food (Wiessner,

PAGE 32

20 1982) . She describes how social ties would mitigate loss, coupled with territoriality, allowing groups to control resource s (Wiessner, 1982) . This idea is echoed by Porraz et al. (2013) in their analysis of Diepkloof, a rock shelter containing pulsed occupation sequences spanning 107 ka 14 ka . Located in an environment in close proximity to the coast, it offered a variety of resources (Porraz et al., 2013) . During the period associated with Howi esons Poort technology (~70 ka ), stacked hearths indicate an intense occupation, corresponding to symbolic signaling on engraved OES (Porraz et al., 2013) . It has been argued that this symbolic and technological advancement originated with an increased population density, which would have necessitated territoriality and symbolic signaling (Porraz et al., 2013) . Porraz et a would have favored and increased cultural interactions between groups, providing a favorable context for the development and diffusion of innovations (2013: 8). Social intera ctions can describe how networks were developed and established in southern Africa, allowing groups to develop social networks and access a variety of ideas, technology, and possibly subsistence resources , offsetting risk that may have occurred with enviro nmental shifts. Bousman (2005) argues that technology shifts coinciding with environmental change enabled occupants of Blydefontein, an LSA rock shelter , to cope with risk. Using optimal foraging theory, he describes strategies to reduce shortfalls, in terms of loss prevention (Bousman, 2005) . Econ omic failure is difficult to observe in the archaeological record, and environmental proxies are potentially misleading, necessitating multiple lines of inquiry ( Bousman, 2005) . Various technologies incur differing collection, processing, and maintenance costs, and diverse strategies may be practiced by groups with varying forms of

PAGE 33

21 mobility (Bousman, 2005) . He suggests that LSA hunter gathers demonstrated a more flexible strategy, shifting technology, mobility and subsistence practices in conjunction with environmental shifts (Bousman, 2005) . Populations under stress from reduced territories, environmental shifts (which can change the available prey profiles), and potentially deadly encounters with other groups might utili ze a varying set of technology to offset risk. If they find themselves restricted by the territories of competing groups, they can choose a diverse set of actions, including direct confrontation/competition, cooperation, or migration. Parsing out which is practiced at different times by different groups will challenge archaeologists. The already sparse archaeological record in South Africa is complicated by the coastal recessions and transgressions , which may have drawn populations out onto the Agulhas plai n. Potential sites would have been destroyed during the flooding of the plain during interglacial periods, complicating understanding of these early populations. I f the development of territoriality is linked to a preference for marine resources, then rese archers would expect to see an aband on ment of the interior sites as populations move out onto the floodplains (Jerardino and M arean, 2010) . Flooding of the Paleo Agulhas Plain during interglacials may have caused resource stress, given the loss of land and the newfound proximity of foragers (Compton, 2011b; Faith, 2011) . Alternatively, foragers living in the interior during the glacials may suggest that other groups have established themselves around dense and predictable marine resources. A case study: Knysna Eastern Heads cave 1 Knysna Eastern Heads cave 1 (KEH 1) offe rs a potential lens onto territoriality and resource access during the MSA/LSA transition. Situated on the edge of the Knysna lagoon,

PAGE 34

22 the cave sits about 23 meters above sea level, high enough to survive the current ocean transgression. Thus far, five stra tigraphic levels have been identified spanning an environmentally transitional sequence dated between ~46,000 and ~19,000 . During the Last Glacial Maximum, the site is intensely occupied even as adjacent sites are abandoned. 2.1 KEH 1 in relation to significant MSA and LSA sites (Cleghorn, pers. comm. Created by Erich Fisher ). The maximum expansion of the Paleo Agulhas Plai n would be an estimated 70 75km in front of the site, based on modeling by Fisher et al. (2010). Understanding the available resources and faunal ecology will mean drawing upon the specimens recovered from KEH 1. Researchers at nearby sites have used faun al material to answer a range of questions concerning hunting ability, local ecology, environmental proxies, subsistence shifts, and population pressures (Klein and Cruz Uribe, 1983; Klein et al., 2004; Faith, 2008; Thompson, 2010; Thompson and Henshilwood, 2014b; Copeland et al., 2016) . KEH 1 represents a little known period of rapid environmental shifts. Analysis of

PAGE 35

23 faunal remains can not only clarify the local environment that early humans were navigating, but also foster a deeper understanding of how they used and reacted to this environment. Drawing upon a framework of HBE, faunal analysis can illuminate how landsc ape impacted early modern human subsistence strategies, and the ways social networks, mobility, and territoriality were shaped. If marine resources are as significant to early humans as has been assumed, then the intense site use over time despite ocean r etreat is curious. Possible explanations for the re current occupation at KEH 1 during this time are 1) that use of marine foods was only one of several subsistence strategies employed by early humans in this region ; 2) the intersection of terrestrial and r iverine resources around Knysna rivaled coastal resources; and 3) competition and territoriality caused by dense marine resources on the coast forced groups into the interior. It may be that the location of KEH 1, overlooking the Agulhas plain, positioned next to potential riverine, lacustrine, and terrestrial resources, offered early humans a rich and valuable setting during all global climate changes . The Agulhas plain is a now submerged continental shelf, which was the setting for rich terrestrial resour ces (Fisher et al., 2010; Compton, 2011a; Copeland et al., 2016) . The cave , situated 23 meters above the Plain and current sea level, represented a defensible position, offering security to inhabitants. These factors made it a highly desirable site, and the location may have been prefer red , as Hunter gathers could track animal movement on the exposed Paleo Agulhas Plain. Conversely, during glacial periods when the coast was far away, if competition for coastal resources wa s high so that territories on the coastline are established and defended by rival groups, then the occupants of KEH 1 may not have had access to the rich marine resources. If so, occupants of KEH 1 might focus on obtaining protein rich resources in the

PAGE 36

24 are a around Knysna, to compensate for the dearth of marine foods. However, i ncreased population densities are hypothesized to correlate with the contraction of the Plain, along with expansion of populations into presumably (Fa ith, 2013) . The loss of the rich terrestrial resources of the Paleo Agulhas Plain to flooding may have resulted in stress, as populations lost ground and fought for territory . This thesis examines to what extent marine resources are a key component of the early human diet during the MSA LSA transition in South Africa. If, per optimal foraging theory, they are a preferred resource (Sealy, 2006; Jerardino and Marean, 2010; Brown et al., 2011; Compton, 2011b; Marean, 2011, 2014; Will et al., 2016, 2019) , then I expect to see indications of nutritional stress when the coastline is distant. If marine resource use is tied to population density and loss of the Paleo Agulhas Plain, then I expect to see nutritional stress when the coastline is within the daily foraging radius. If the occupation of KEH 1 simply indicates multiple foraging strategies by early Hunter Gatherers, then th e indica tions of nutritional stress should be absent or not statistically different. While marine resources offer significant protein and nutrients for foraging populations, there are other resources which can provide similar benefits. The nutritive components of long bone marrow are comparable to marine resources (Will e t al., 2013) . Populations under nutritional stress have been show n to process long bones more extensively for marrow (Hodgkins et al., 2016) . Critically, this includes processing long bones with min imal resources, including the distal portions of the feet, which contain relatively insignificant amounts of marrow (Hodgkins et al., 2016) . If early modern human populations at KEH 1 were stressed, then it is expected that they would process the terrestrial resources more strongly. Conversely, if there is relatively little processing of long bones, then it would

PAGE 37

25 be argued that these groups were not stressed. This type of analysis could offer insight into the decision to remain at KEH 1 during ocean regression. The answer to this question can offer a discussion on how early modern hum ans coped with environmental shifts . If the marine adaptation is tied to behavioral modernity, then the territoriality and competition which are hypothesized to follow can be tested. Although the flooded Agulhas plain likely represents a loss of much of th e data concerning the MSA/LSA transition, KEH 1 offers a window onto this fascinating period. The potential for elucidating the se impacts on early modern humans offers a unique perspective on MSA/LSA populations. Through this lens, we can begin to describe how territoriality and differential access to resources were shaped and discuss the potential effects on these peripheral groups.

PAGE 38

26 CHAPTER III SITUATING KEH 1 WITHIN THE SOUTH AFRICAN MIDDLE AND LATER STONE AGE MSA and LSA sites of South Africa The Middle Stone Age/Upper Paleolithic was an important period, capturing both the globalization of EMH and disappearance of Homo neanderthalensis (Clark and Kandel, 2013) . However, the MSA record in south Africa is sparse, especially the transition from the MSA to LSA (Faith, 2013) . There is gap from 50 k a to 2 5ka, which occurs simultaneously at numerous sites (Klein, 1976) . Sites containing an occupation seque nce for this period occur sporadically, in the eastern part of southern Africa (Border Cave, Sehonghong, Sibudu, Rose Cottage cave, and Klasies R iver M outh) to the western cape (Deipkloof, Elands Bay, Putslaagte, Varsh Rivier, Klein Kliphius ) and the Cape F loral R egion (CFR) , w hich is included in the Western cape, but differ s in rainfall patterns (Boomplaas , Buffelskloof, Nelson Bay Cave, and KEH 1 ) . As KEH 1 is in the CFR, it represents a critically underrepresented and misunderstood period of prehistory. In order to understand the broader context of the MSA/LSA transition, this chapter will ex amine sites with occupations during the MSA/LSA transition , as well as discuss previous work on p aleoenvironmental reconstruction, and summarize environmental, ecological, and archaeological discussions of the CFR. Finally, it will describe KEH 1.

PAGE 39

27 3.1 Map taken from Mackay (201 0 ), showing sites across Southern Africa in relation to rainfall zones (based on Chase and Meadows , 2007) . Sites labelled by Mackay as follows: AXI = Apollo XI, BBC = B l ombos, DRS = Diepkloof , EBC = Elands Bay Cave , HRS = Hollow R ock Shelter, KFR = Klipfonteinrand, KKH = Klein Kliphuis , KRM = Klasies River Main , NBC = Nelson Bay Cave , NT = Ntloana Tsoana, PC = Peers Cave, RCC = Rose Cottage Cave , SC = Sibudu Cave , SHH = Sehonghong , SHO = Shongweni, SIB = Sibede, UMH = Umhlatuzana. Sites mentioned in this text highlighted by the author. Sites mentioned but not depicted here are shown in figure 2.1. Eastern cape Border C ave ( 227 ka 39ka) within the borders of Swaziland, demonstrates some of the oldest evidence for modern behavior proxies Backwell, 2016) . In addition to the earliest symbolic burial, recovered bone tools are argued to be evidence of complex behavior, and the use of personal ornaments as reoccurring

PAGE 40

28 symbolic motifs . Rose Cottage Cave dates from 250ka to 20ka (Clark, 1997) . The lithic industry contains a sequence similar to Sehonghong (~32 ka 20ka), diverging from typical MSA or LSA technology (Clark, 1997) . The Early Later Stone Age (ELSA) is poorly understood an d lacks temporal contextualization, as it appears gradually across Southern Africa (Clark, 1997) . Sehonghong contains a serie s of shallow hearths and a (Mitchell, 1994) , and two bone tools (Clark, 1997) . Sibudu captures nearly 100,000 years of early modern human occupation ( terminating ~30 ka), framing discussions of behavioral modernity and subsistence strategies (Clark and Plug, 2008) . Evidence from this site suggests human accumulation of fauna from both closed and open habitats (Clark and Plug, 2008) . Occupants may have ranged long distances from the cave to obtain prey, (HP) age (Clark and Plug, 2008) . Early modern humans at S ibudu focus ed on local ecology , which is argued to (Clark and Plug, 2008, pg. 11) . These sites offer insight into the habits, technology, subsistence and culture of ELSA people in South central southern Africa , however, they are over 200 kilometers from the coastline. Although it is possible that the inhabitants were part of a social network connecting them to the coastline, no evidence of t his currently exists. Klasies River Mouth ( KRM) (125 ka 30 ka ), is a series of caves adjacent to the coastline (Klein, 1976) . E arly human remains have been discovered at the site, and have been argued by some to represent a hominin other than AMH (Lam et al., 1996) , although the remains are mostly those of adults, unlike at other sites (Grine et al., 2017) . The site displays several examples of behavioral modernity carved ochre and hunting large ungulates (Milo, . analysis of lithic material from KRM and Rose Cottage

PAGE 41

29 suggests an environmental driver for the disappearance of the HP during the glacial/interglacial transition ~60 ka (Villa et al., 2010) . Although KRM is within 160 kilometers of KEH 1, it contains no occupations representing MIS2, when the ocean retreat ed . Western C ape The W estern cape has yielded a handful of transitional MSA to LSA sites, which have aided in the environmental and ecological reconstruction, descriptions of hunting strategy, population estimates, and understanding technological shifts. Elands Bay contains thr from 300 ka to 13 ka ( uncalibrated) (Klein and Cruz Uribe, 1987) . Variability in recovered species between the stratigraphic levels correlates to the rising sea and vegetation shifts occurring during the beginning of the Holocene, as shrubland and forest replaced the grasslands of the late Pleistocene (Klein and Cruz Uribe, 1987) . The number of juvenile seal remains suggests a seasonal use of the site corresponding to seal mating cycles, and a reduction in tortoise size with increased human population densities (Klein and Cruz Uribe, 1987) . Diepkloof rock shelter has occupation sequences spanning th e MSA (107 ka ) through post HP (57 46 ka ), and the LSA (14 ka ) (Jacobs et al., 2008; Steele and Klein, 2013; Tr ibolo et al., 2013) . Diepkloof provided AMH with a strategic setting, offering riverine, coastal, and terrestrial resources (Porraz et al., 2013) . Seal remains occur infrequently in MIS4 (when the coast was much nearer), and more frequently during the LSA, despite the fact that the coa st wa s at least 18 km distant (Steele and Klein, 2013) . These factors have b een used to suggest social networks, increased foraging radius, an importance of marine resources, and a possible use of the lake shore to reach the coast (Steele and Klein, 2013) .

PAGE 42

30 The occupation of inland rock shelter Putslaagte 8, occurred in pulses throughout MIS 4, 3 , 2, (75 ka 13 ka), with robust Holocene deposits (Mackay et al., 2015) . Occupation may not be limited b y population expansion, but rather to mobility among hunter gatherer groups using multiple resources as early as ~75 ka (Mackay et al., 2015) . Varsh Rivier contains t he youngest known occur rences of SB and HP industries (HP ~45 41 ka , SB ~59 ka 45 ka ) (Steele et al., 2016) . The LSA deposits contain marine shell despite a distance of more than 45 kilometers to the coastline (Steele et al., 2016) . Klein kliphius rock shelter dates from 68.1 ka to~20 ka (Mackay, 2010) . Although site use appears to cease or beco me intermittent around 55 ka , it becomes more intense at 30 ka (Mackay, 2010) . Sites in the western cape are occupied in pulses, and both Diepkloof and Varsh Rivier are abandoned during the interglacial/glacial transition between MIS3 2. De spite significant distances to the coastline, marine resources continue to be transported to both sites. In the CFR (the southern edge of the W estern C ape ) , few published site s are dated to the MSA/LSA transition, while others are abandoned during this time . Boomplaas cave is an inland site with deposits spanning 65 ka 15 ka , and an abundance of faunal remains (Faith, 2013) . Analysis demonstrates that the remains were accumulated by carnivores until ~50 ka , and then the agents of accumulation were mixed, with contributions from humans, carnivores and raptors until the LSA (~18 ka ), when humans appear to be the sole inhabitants (Faith, 2013) . Faith (2013) argues that this signature indicates low intensity occupations of the site until the LSA, when population densities are thought to have increased. Buffelskloof is approximately 40km to the west of Boompla as, alt hough e xcavations have not revealed any occupations earlier than 23ka (Opperman, 1978) . Nelson Bay Cave (NBC) was occupied prior to 50 ka , and then again from 18 ka 5 ka (Butzer, 1973) . Although it represents one of the

PAGE 43

31 closest excavated sites to KEH 1 (~30 km), there is a gap in the occupational sequence un til ~20ka (Loftus et al ., 2016) , when both sites begin to line up . The site is notable for its unique technology, the Robberg (dated ~20 12 ka ), and for its Holocene deposits, which contain human remains (Deacon, 1978; Sealy, 2006; Loftus et al., 2016) . Analysis of the trop h ic levels from the human remains are consistent with habitual consumption of a marine rich diet (Sealy, 2006) . T he southern CFR is characterized by inadequate representation of transitional sites. The abandonment of NBC, only 30km east, makes the recurrent occupation of KEH 1 more fascinating. What kept humans at KEH 1, and sporadically, although consistently, at Boomplaas? NBC is similar in many respects to KEH 1: located on the current coastline, abutted by the Knysna forest, and offering a view of the exposed Paleo Agulhas plain. Determining the potential of the caves will require analysis of the micro environments and the resour ces available to AMH . Environmental and ecological reconstruction In order to understand how humans evolved, we must understand the environment in which they coped. Fisher et al., ( 2010) environmental change which drastically alters the landscape of the souther n coast . Paleo environmental reconstruction is an important part of faunal analysis, as the animals occurring in an area can provide insight to the local environment. Different environments would have offered different challenges to EMH, and the different responses of groups led to different social cultural behaviors, hunting decisions, prey return rates, and long term skeletal element survival. Numerous researchers have used faunal assemblages to reconstruct these environments and understand the challenges faced by early modern humans.

PAGE 44

32 Other Pleistocene faunal assemblages have been used to reconstruct environments (Klein, 1989; Rector and Reed, 2010; Faith, 2011, 2011; Thompson and Henshilwood, 2011; Copeland et al., 2016; Hodgkins et al., 2018) . Hyena dens are particularly useful for this analysis, as hyenas consume a broader range of species and collect skulls and dentition (Rector and Reed, 2010) . Cranial materi al, particularly dentition, is especially useful for comparative analysis (to determine the genus/species) (Rector and Reed, 2010) , and isotopic analysis, which provides information on the mobility patterns of animals on the submerged continental shelf (Copeland et al., 2016; Hodgkins et al., 2018) . Once species has been determined, the researchers can compare the modern behaviors of these animals to better understand the local Paleo ecological zones. Ostrich eggshell has also been sampled isotopically, providing insight to paleo environments (Lee Thor p and Ecker, 2015; Hodgkins et al., 2018) . S maller mammals (.75 4.5kg) and microfauna are often overlooked at archaeological sites, as larger animals are hypothesized to provide better insight into dangerous/risky game hunting, mobility, and paleoenvir onmental reconstructions (Armstrong, 2016) . However, at certain sites, including Die Kelders, where the assemblage is more than 85% small mammals, they are critical to understanding of accumulation processes and environmental trends (Armstrong, 2016) . Smaller animals/microfauna are more frequently used to understand the accumulation processes, as etching may implicate non human agents, and thus, differential spatial or temporal occupations (Marean et al., 2000; Thompson and Henshilwood, 2011, 2014a; Enloe, 2012; Faith, 2013) . This does not preclude the potential for these species to contribute to environmental and ecological reconstructions, however, it does require that

PAGE 45

33 more archaeologists become trained in the identification and analysis of smaller mammals (Armstrong, 2016) . Paleo environmental reconstruction based on fauna is not without complications. A fundamental difficulty is that comparative analys e s are closely tied to extant communities (Rector and Reed, 2010) , requiring the researcher to dra w assumptions about the similarity of behavior despite temporal variation. Certain ungulates utilize multiple biomes, further complicating the issue (Domíngu ez Rodrigo and Musiba, 2010) . Describing the environment and ecology is complicated by the fact that different predators will prey upon different animals, likely excluding a portion of the potential prey (Thompson and Henshilwood, 2011) . The location of the site mi ght impact the animals transported, as larger animals are more likely to be processed at the kill site (Schoville and Otárola Castillo, 2014) . Unfortunately, conclusions based on the skeletal elements will be impacted by the survivability of an element (Marean and Spencer, 1991; Lam et al., 1998; Cleghorn and Marean, 2004; Faith and Thompson, 2018) . Thus, only a subset of the biocoenosis will become incorporated into the archaeologi cal record. In addition to human food preferences/prey choice, taphonomic issues may reduce the accuracy of reconstructions. Time averaging, or the variability in temporal deposition in the same stratigraphic layer, may result in an inaccurate recreation of the paleobiome, as multiple transitions might become collapsed into one layer (Behrensmeyer et al., 2000; Domí nguez Rodrigo and Musiba, 2010) . Fauna accumulated outside of anthropogenic contexts (i.e. higher probability of preservation (Behrensmeyer et al., 2000; K idwell and Holland, 2002; Domínguez Rodrigo and Musiba, 2010) . These either represent the fauna from this specific

PAGE 46

34 environment, or migratory fauna from numerous biomes, and skew the data towards (Domínguez Rodrigo and Musiba, 2010) . Subsequently, researchers have raised concerns about the accuracy of paleo environmental reconstruction (Domínguez Rodrigo and Musiba, 2010) . Although these should not be taken lightly, faunal analysis still represents one aspect of paleo environmental reconstructions and should be combined with other met hodologies to ensure accuracy. Methods of Paleoscape reconstruction s have also included isotopic analysis of speleothems (Bar Matthews et al., 2010) , and geospatial modeling to determine the extent of the exposed Paleo Agulhas plain during the MSA (Fisher et al., 2010) . Many of these are coupled with high resolution temporal reconstructions, including optically stimulated luminescence (Jacobs, 2010) and cryptotephra (Smith et al., 2018) . Paleo botanical data, including pollen, has been used to determine the extent and type of arboreal versus grassland habitats in the CFR (Chase and Meadows, 2007) . Currently, there are nine floral biomes recognized in South Africa, although their borders during the Pleistocene are disputed (Scott and Neumann, 2018) . The fynbos biome encompasses much of the CFR, and is recognized as having the highest plant biodiversity in the world (Scott and Neumann, 2018) . Additionally, the largest component of the forest biome in southern Africa rests in the area around Knysna (Scott and Neumann, 2018) , although the extent of this forest during the Pleistocene is unknown. These methodologies have provided previous researchers with a myriad of sound data on the effects of glacial/interglacial shifts on the envi ronment and ecology of the Paleo Agulhas plain. Bathymetric modeling has pinpointed the extent of exposure, allowing subsequent researchers to determine the bio available landmass at intervals of 1.5 ka (Fisher

PAGE 47

35 et al., 2010) . Expanded grasslands are correlated to rainfall on the plain, while the area around the Pinnacle Point sites appears to be a transitional zone from the sparse shrublands from the northern mountains (B ar Matthews et al., 2010; Rector and Reed, 2010) . The paleo A gulhas P lain offered a rich landscape, with a diverse and densely populated large animals utilizing the open, grassy area (Marean et al., 2014) . The increased frequency of the largest ungulates during interglacial periods may be linked to the environmental changes associated with the available species (Clark and Kandel, 2013) . It has been proposed that similar ratio s of buffalo in the MIS5 and MIS3 at Die Kelders, Klasies River Mouth, and Pinnacle Point suggest that open environmen ts could be characteristic of interglacial periods (Clark and Kandel, 2013) . Copeland et al. ( 2016) theorize that the paleo Agulhas soils may have been considerably richer than those of the CFR, offering rich grass resources to larger ungulates during ocean regressions. Ungulates appear to prefer the paleo Agulhas plain, only rarely venturing north (Copeland et al., 2016) . A lthough a species capture rate is correlated to the encounter rate, if a species is hig hly represented in an assemblage, researchers consider other ways that human behavior might be affecting this accumulation . For example, eland are highly represented by Blombos, DK1, and KRM, although researchers believe that they have a low population density (Dusseldorp, 2010) . This may s uggest that MSA populations were well informed of eland behavior and could choose their encounters (Dusseldorp, 2010) . The variation between the expected and realized capture rates may indicate that the MSA populations responded to environmental shifts which influenced the availability of the species (Dusseldorp, 2010) . The reconstruction of Paleo environments has allowed researchers to hypothesize correlations between subsistence strategies and techn ological innovation. However, although

PAGE 48

36 the environment may have prompted the technological shifts, it do es not control how the culture changes (Deacon, 1978) . For example, during the early stages of Holocene occupation at NBC, there was a shift in environment (grassland to brush), causing hunters to take fewer grazers and more browsers, but also animals >50kg (Deacon, 1978) . This change is argued to correlate with shifts in micro technology and a more diverse toolkit (Deacon, 1978) . The Still Bay (~76 72 ka ) and Howie sons Poort industries (~65 59 ka ) appear to be associated with both glacial and interglacial pollen assemblages on both interior and coastal sites (Scott and Neumann, 2018) . The variation of the fauna at Sibudu during the Still Bay might suggest a link between technological variability, available fauna, and environmental shifts (Clark, 2017) . The decline in species during the Howiesons Poort might be attributed to reduced diet breadth, and thus possible intensification of reduced resources (Clark, 2017) . Previous work on the CFR has created an environmental se quence covering the MSA and LSA. However, much work remains to be don e in understanding the area around Knysna. Subsistence strategies of the CFR S ubsistence strategies of MSA and LSA hominins have implications for human behavioral modernity, environmental adaptation, and social constructs. A key feature of the CFR is the intersection of marine and terrestrial resources, and the contributions of each to early human diets (Marean, 2011) . Although archaeological discussions of subsistence are generally dom inated by large mammals, it is argued that shellfish played a critical role in demographic expansion and adaptation in the CFR (Jerardino and Marean, 2010) . However, it has been argued that the MSA does not have a singular subsistence strategy (Thompson and Henshilwood, 2011) . This appears particularly true across the temporal span of the MSA and the environmental shifts which characterize MIS 5 to MIS 2. Although shel lfish

PAGE 49

37 exploitation is high, less evidence for marine mammals has been uncovered, and it appears that they were consumed opportunistically (Thompson and Henshilwood, 2011) . Understanding the transport decisions of early modern humans requires an understanding of the en vironmental, ecological, and cultural variables which impacted the availability and locations of resources (Dusseldorp and Langejans, 2013) . The variety of marine resources may provide a proxy for the mobility patterns and transport decisions utilized by people (Dusseldorp and Langejans, 2013) . However, determining whether shifts in marine resources, the quantity of these resources, and the pr ey age profiles reflects seasonal distribution, environmental factors, or intensification is another question (Dusseldorp and Langejans, 2013) . Some animals, like whales, may be nearly invisible in the archaeological record, due to enormous, impossible to transport skeletal elements (Dusseldorp and Langejans, 2013) , but see Jerardino and Marean ( 2010) . Additional resources (seal and mussels) might have been processed where they were collected, vanishing from the record as the ocean re gressed (Dusseldorp and Langejans, 2013) . However, researchers should focus on the spectrum of represented archaeological faunal material, to fully understand subsistence strategies of MSA populations (Thompson, 2010) . Researchers have used the CFR to contribute to arguments of behavioral modernity hypothesis (Klein, 1995, 2000; Faith, 2008; Dusseldorp, 2010) . Dusseldorp (2010) and Faith (2008) argue that MSA hunters possessed the capacity for complex hunting strategies, targeting dangerous and difficult prey. This may suggest that large mammals held a plac e alongside marine foods or were targeted during ocean regressions.

PAGE 50

38 Early modern humans also readily gathered tortoises, ( Chersina angulata ) which are available for collection and consumption year round (Klein and Cruz Uribe, 1983) . Chersina is the most common tortoise recovered from Bloombos, and it appears that humans processed them with fire and hammerstones (Thompson an d Henshilwood, 2014a) . During the coastal regression at Blombos and Pinnacle Point, tortoises would have become a more important resource to offset the loss of high protein shellfish (Thompson and Henshilwood, 2014b) . Other frequently overlooked resources available to humans include small fauna, birds, ostrich eggshell, and botanical foods. It is hypothesized that early humans may have utilized the available small fauna to fill any gaps left by a receding coastline and coastal resources (Armstrong, 20 16) . Prevailing theories consider bird exploitation to be synonymous with the LSA, however, avian remains are noted throughout the MSA deposits at Sibudu (~77 ka ) (Val et al., 2016) . The ubiquity of these remains in MSA deposits showcases the ability of early humans to utiliz e a diverse and complex resource base (Val et al., 2016) . Ostrich eggshell is frequently recovered at MSA and LSA sites, suggesting that early humans were aware of the nutritional qualities, and utilized this resource for subsistence (Collins and Steele, 2017; Hodgkins et al., 2018) . Consumption of geophytes (the CFR is noted for its geophyte dive rsity) is known from sites as early as 100 ka (Deacon, 1993) . The discovery of these diverse resources in archaeological contexts suggests that early humans were not on ly aware of the myriad of resources but had the ability to utilize them. However, this does not always imply that they were reliant on any particular resource but may have utilized them more opportunistically. Early modern humans in the CFR collected, hu nted, and consumed a diverse subsistence base. Some of their food choices are arguably novel (marine subsistence), while

PAGE 51

39 others may reflect strategies used more broadly across southern and eastern Africa. MSA and LSA populations do not appear to vary in th eir ability to procure diverse foods (Faith, 2 008; Dusseldorp, 2010; Wadley, 2010; Thompson and Henshilwood, 2014b) . Instead, the variation in prey choice appears to correlate with environmental shifts. As KEH 1 sits at the edge of an extinct ecosystem across extreme changes in sea level , it can of fer critical insight into early human subsistence. KEH 1 : Location and stratigraphy KEH 1 is located on the exterior of the Knysna heads, facing the Indian O cean. It is approximately 23 meters above sea level, ensuring that the deposits were not washed out during recent high sea stands , although it would have been subjected to destruction during MIS5e (~ 128 ka 110ka) high seas stands (Fisher et al., 2010) . The coastline in this area is part of the Table Mountain Super group, and quite rocky and riddled with caves. The closest site to KEH 1 is KEH 2, which sits directly east and slightly lower than KEH 1 but has lost much of its deposits to tidal action. T his is an archaeologically rich area, and surveys have uncovered an ESA open air site within five kilometers, and several MSA/LSA sites embedded within the Featherbed caves on the western side of the Knysna Heads . Although no other archaeological excavations have been recorded at Knysna, several finds have been published. A painted seal scapula (age indet.) was recovered from one of the many caves around the Knysna heads and currently rests in the British museum (Sealy, 2006) . Recently, fossilized human footprints estimated ~90 ka were discovered in a cave less than 2 kilometers to the west (Helm et al., 2018) .

PAGE 52

40 Stratigraphy The deposits of KEH 1 at the mouth of the cave are cut by an erosional slope , allowing excavation to target the MSA/LSA transition. They are capped by 1.5 2 meters of Holocene shell midden. The cliff face above is sheer; bel ow is a steep slope covered with thick fynbos, terminating in barren horizontal planes of quartzite which are partially submerged by tidal action. The cave is currently accessible by a narrow, steep path on the SW corner. Following a test excavation in 201 4, the cave has been excavated in every subsequent year , removing more than 40,000 archaeological finds. The site is relatively undisturbed, except for the slumping deposits at the front of the cave. Archaeological material (out of context) has been recove red from the path below. Round depressions at the back of the cave (with a diameter > 1 meter) suggest that previous digging occurred, although no records have been uncovered. It is possible that this was unsanctioned excavation, resulting in the removal o f indigenous burials (Cleghorn, pers. comm). Smaller disturbances (burrows) have been recorded in several layers . At KEH 1, archaeological layers that , although they may have been accumulated over thousands of years, appear to have been deposited under similar environmental conditions. Therefore, the units of analysis here are not the individual units excavated, but rather the broadly similar deposits, in order to describe shifts in human behavior temporally (Karkanas et al., 2015) . The lowe st stratigraphic aggregate, the Dark Brown Spally (DBS) >46 ka 34 ka , contains sparse evidence of anthropogenic occupations, although charcoal is plentiful. The coastline is an unknown distance away but is not estimated within daily foraging radius (10 12km) (Cleghorn, pers. comm) . Small amounts of fauna (N= 101) have been recovered. The Dark Shelly (DS) ~32 ka 29 ka , is an anthropogenic

PAGE 53

41 deposit with large accumulations of shell, and relatively dense fauna and lithics. The coastline is likely within ten kilometers of the cave, and it appears that occupants relied heavily on marine resources at this time (Cleghorn, pers. comm) . Dates on the Dense Hearth Aggre gate (DHA) occur around the start of the last glacial maximum (~26 ka 22 ka ), when the Paleo Agulhas P lain begins its final , most recent exposure. This aggregate contains the highest find density, and is unique in its large, well preserved hearths. These exten d across the entire entrance of the cave and are densely stacked. To date, more than 50 hearths have been excavated. Large amounts of ochre have also been recovered. The Orange Brown Sandy (OBS) currently has one date , at ~19 ka , at the height of the L ast G lacial M aximum. Around this time, the maximum exposure of the Paleo Agulhas plain is estimated at 70 kilometers (Cleghorn, pers. comm) . Anthropogenic finds persist throughout this level, although high numbers of microfauna have been found, the accumulating agent of which is currently undetermined. Above this rests a thick sterile layer overlain by the shell midden, currently undated and mostly unexcavated. 3. 2 Photo of the KEH 1 stratigraphy (Cleghorn, pers. comm) .

PAGE 54

42 CHAPTER IV METHODS OF DATA COLLECTION AND ANALYSIS KEH 1 was mapped in 201 3 , and control points were established to allow geodetic control of the cave (Cleghorn, pers. comm). Once the Universal Transverse Mercator grid was established, excavations were based on this system (Cleghorn, pers. comm). Excavations were restricted by stratigraphic unit and lot , contained withi n 50cm x 50cm excavation units , and each lot was assigned a unique identification number. Each un it was mapped with a total station, and finds were mapped upon removal, without any size cutoff, and assigned a find number. Upon removal, finds were washed, sorted, and stored in the Mosselbaai lab until analysis. Aggregate Climate cycle total Analyzed (frag/tap) total % Analyzed OBS Glacial 902 475 52% DHA Transition 9303 1127 12% DHA/DS Transition 291 263 90% DS Interstadial 688 536 78% DBS Glacial 101 32 32% unassigned 98 1 1% Total assemblage 11401 2434 21% Table 1: Description of the number of faunal specimens analyzed for fragmentation and taphonomy by stratigraphic aggregate More than 11,000 faunal specimens have been recovered during the excavations (artifacts from the 2017/18 field seasons, which have yet to be s orted, are not included here). A majority (>9,000) of these are from the DHA (table 1 ). Only a small number of fauna ha ve been recovered from the oldest aggregate, the DBS. For the scope of this project, we aimed to analyze a minimum of 10% of the fauna in each aggregate. Analysis occurred over the summers of 2017/2018, resulting in our analysis of 2,434 specimens, or 21% of the total

PAGE 55

43 f aunal assemblage. Fragments were initially chosen at random, (i.e., beginning with the earliest excavated units), until the minimum frequency of finds per aggregate were analyzed. Specimens were then analyzed sequentially, to begin working through the assemblage. Taphonomy Faunal specimens were assessed using zooarchaeological standards, to determine the extent and origin of fragmentation, degree of taphonomic damage, size and taxon of animals, and any anthropogenic modifications. Taphonomic effects, including rodent gnawing, sediment loadi ng (trampling), gastric etching, acidic soil/rainwater, weathering, displacement, and density mediated attrition may delete or alter faunal remains (Behrensmeyer, 1978; Lyman, 1984; Gifford Gonzales et al., 1985; Marean and Spencer, 1991; Cap aldo and Blumenschine, 1994; Lyman and Lyman, 1994; Dibble et al., 1997; Lam et al., 1998; Behrensmeyer et al., 2000; Marean et al., 2000; Pickering et al., 2003; Domínguez Rodrigo et al., 2009a; Enloe, 2012) . In order to determine human behavior, archa eologists must first rule out natural accumulations, alterations, and effects that would skew the evidence. Archaeologists should not assume that the assemblage is a complete representation of everything that was deposited, much less that everything that w as brought to the location by humans or non human agents (Behrensmeyer et al., 2000) . In order to assess the effects of taphonomy, specimens were examined to determine the completeness of surface, evidence of breaks, etching, an d weathering. If the surface is eroded/removed, it suggests that an agent/process worked to remove it (for example, acidity in the soil/stomach, weathering), and that surface modification s related to nutrient processing (cut/tooth/percussion marks) may have been deleted (Marean et al., 2000) . Fragments were measured for both length and width to compare overall specimen size between aggregates,

PAGE 56

44 which could reflect size sorting by natural or anthropogenic processes, and potentially correlated with degree o f fragmentation . When bones of similar sized animals undergo varying degrees of fragmentation, it may suggest either an intensive hominin processing strategy or more frequent exposure to destructive post depositional processes, such as trampling. When pos sible, bones were identified to element, using the comparative collection in the Mosselbaai Archaeology Project (MAP) lab. This allowed for the determin ation of the frequency of high survival versus low survival elements. Density mediated destruction can a lter interpretations of archaeological assemblages (Brain, 1969; Lyman, 1984; Marean and Spen cer, 1991; Lam et al., 1998; Marean et al., 2004) . High survival, i.e. dense bones, include parts of the skull and long bone shafts, both of which are more likely to survive carnivore attrition and taphonomic processes (Cleghorn and Marean, 2004) . Conversely, low survival elements, (ribs, vertebrae, long bone epiphyses, and pelvic bones) tend to be grease bearing, and consisti ng of spongy trabecular bone, both of which combined reduce the likelihood of survival (Cleghorn and Marean, 2004) . Significant debate s have arisen over the need to assess survivability, as the elements most likely to be destroyed are also the most identifiable, and thus, more likely to be retained and analyzed (Marean and Spencer, 1991; Marean and Kim, 1998; Pickering et al., 2003; Marean et al., 2004) . Currently, many zooarchaeologists are trained to identify shaft fragments, which are not only more likely to survive, but are also more likely to retain anthropogenic surface modificati on. Assessing Nutrient processing Nutrient processing refers to the reduction of long bones to retrieve the marrow contents. Distinguishing between nutritive versus non nutritive breaks (also known as green

PAGE 57

45 versus dry fractures), requires analyzing the an gle and outline of the ends of the shafts, based on methods developed by Villa and Mahieu ( 1991) . Nutritive fractures include curved outlines and oblique angles, while non nutritive fractures consist of straight outl ines and right angles (fig. 4.1 ) (Villa and Mahieu, 1991) . Processing of bones also results in changes in the circumference (Bunn, 1983) . Non nutritive breaks are the resu lt of taphonomic effects, and can include trampling, burning, sediment loading, and excavation (Hodgkins et al., 2016) . Frequently, this analysis is carried out on shaft fragments, for two reasons. First, as mentioned above, these are more likely to survive and have cut/percussion marks. Additionally, trabecular bone breaks in a differ ent manner than cortical bone , rendering the criteria inapplicable (Villa and Mahieu, 1991) . Multiple sets of experiments have been compiled by Marean et al. ( 2000) , and provide a robust dataset with which to compare zooarchaeological data. Hodgkins et al. ( 2016) demonstrated th e potential for using nutritive values across climatic events to understand how hominin subsistence strategies are altered. This method applies the data to broader paleo environmental, paleo ecological, and evolutionary trends. 4.1 Non nutritive fractures (left) and nutritive fractures (right).

PAGE 58

46 Archaeologists must also control for issues of equifinality (Lyman, 2004; Marean et al., 2004) . Once the timing of the fractures has been established, the archaeologist must determine the agent that caused the fractures. Both carnivores and hominin processing will create green fractures on long bones (Marean and Spencer, 1991; Marean et al., 2000) . By examining the bones under a microscope (analysis of surface m odification), researchers can determine the presence of tooth, cut or percussion marks (Marean et al., 2000) . Experimental work has demonstrated that tooth, cut and percussion marks can be reliably distinguished from each other (Brain, 1981; Potts and Shipman, 1981; Capaldo and Blumenschine, 1994; Blumenschine, 1995; Blumenschine et al., 1996; Capaldo, 1997; Marean et al., 2004; Pickering et al., 2004) . Furthermore, students with some training are able to correctly analyze surface modification faunal remains with only incidence lighting and a low power magnification (Blumenschine et al., 1996) . Thus, a sample of KEH 1 faunal remains were examined obliquely with a high intensity light source at 15 20x magnification. Size class and Taxon Faunal specimens were compared to the extensive collection at the MAP lab to identify size class and taxon. Size class es are based on those developed by (Brain, 1981) , for ungulate species. These range from size 1 (2 20 kg), to size 5 (901 2000 kg ), although size 2 (21 113 kg ), size 3 (114 340 kg ) and size 4 (341 900 kg ) are frequently recovered in archaeological sites (Brain, 1981) . Ethnographic and experimental work demonstrates the potential for animals size 1 3 to be carried to the site whole or almost whole (Bunn et al., Castillo, 2014) . Larger animals tend to be processed in the field, or carried back to site if the group size is large (Schoville and Otárola Castillo, 2014) . Several high survival elements (the head and feet) are less likely to

PAGE 59

47 be brought t o the site if the distance is far (Schoville and Otárola Castillo, 2014) . However, high utility bones (including the vertebrae, ribs, humerus, and femur) are more likely to be transported (Binford, 1978; Schoville and Otárola Castillo, 2014) . The mix of high and low survival within this category offers a complicated look at hominin transport strategies, and has incurred much debate over early hominin hunting abilities (Blumenschine, 1986, 1989; Stiner and Kuhn, 1992; Marean and Kim, 1998; Pickering et al., 2003; Cleghorn and Marean, 2004; Domíng uez Rodrigo et al., 2009b, 2014; Schoville and Otárola Castillo, 2014) . Elements were identified to species whenever possible, and subsequently a Minimum Number of Elements (MNE) was established by counting the number of fragments that overlapped on eac h bone (Marean et al. , 2001) . Then Minimum Animal Units (MAU) could be calculated. MAU is calculated on the MNE divided by the number of times the element occurs within the body (Binford, 1978) . Thus, it removes the necessity of siding an element to determine how many individuals are represented (Binford, 1978) . Additional anthropogenic modifications Finally, specimens were asse ssed for degree of burning. Bones were coded as calcined, carbonized, unaltered, or heat altered based on a visual inspection. As a majority (>9,000) were recovered from the DHA, it was expected that a number would show evidence of fire damage. However, ev idence of burning on bone can also be used to track exposure to fire outside of hearth contexts . The placement and percentage of burnt surface may indicate cooking, debris disposal, or accidental/natural conflagration (Andrews and Cook, 1985; Clark and Ligouis, 2010; Thompson and Henshilwood, 2014a) .

PAGE 60

48 Statistical analysis T o determine the if the obser ved patterns between aggregates are statistically significant, S PSS was used to run statistical analys es. T o describe the variation in mean length and width, a n ANOVA was applied . A ll 2434 specimens were used for this analysis. T he stratigraphic aggregates were recoded (1= OBS, 2= DHA, 3= DHA/DS, 4=DS, AND 5= DBS) in order to run the ANOVA. The ANOVA compar ed the length while controlling for stratigraphic aggregate, to describe the dispersion around the mean length. To describe which variables were different, Tukey post hoc test was then run . If there are significant variations between the different stratigraphic aggregates, then it is expect ed that more post depositional damage or anthropog e nic processing occurred in these levels, suggesting more intense site use. Howev er, if there are no significant variations, then all levels were subjected to similar levels of processing and/or post depositional stress. A chi square analysis was used to assess whether there is any significan t va riation in the type of fracture (nutritive or non nutritive) of long bones. O nly long bones (N= 733) were used in this analysis . The fracture angles were reco ded to reflect the number of nutritive breaks per aggregate (O BS= 172, DHA= 307 , DHA/DS= 84 , DS= 131 , DBS= 7 ) . A Cramer s V and contingency coeff icient post hoc test were run, to describe the extent of the association. S ignificant differences between the aggregates suggests that long bones were being processed for marrow differently across a shifting environment , triggered by global climatic c han ges. N o significant differences would suggest that early humans/carni vores did not change their marrow processing strategies in respons e to the expansion and contraction of the Paleo Agulhas Plain.

PAGE 61

49 In order to determine if the n umber of tooth, percussion , and cut marks significant ly vary between aggregates, an ANCOVA was r un . T his statistic has been used in p re vious analysis describing the extent of nutritive processing of long bones between climatic cycles (Hodgkins et al., 2016) . ANCOVA determines the significance while controlling for multiple variables (in this case, surface area of the bone, calculated by Geometric mean). T he sample size was restricted to long bones within the subs ample examined for surface modification (N= 102). G iven the absence of cut, percussion , and tooth marks in the DHA/DS and DBS aggregates, the analysis w as only run on the OBS, DH A, and DS material. S ignificant differences between the aggregates would demon strate a diffe re nce in human and carnivore behavior across time.

PAGE 62

50 CHAPTER V RESULTS Fragment size The fauna of KEH 1 is heavily fragmented. There are few whole bones recorded in th is analysis (.005%, or 12 of 2400). A majority (97%) have a length <40 mm and width <20 mm (fig. 5.1 ). The most heavily fragmented aggregates are the OBS, DHA, and DBS (mean length = 12.12, 14.04, and 15.19 mm respectively). The DS has a higher mean length (17.31 mm) , while the transitional DHA/DS has a mean of 20.26 mm. W hile the OBS and DHA have low average length/widths, both also have high dispersals around the mean. An ANOVA demonstrates significance variance in length by level at the = 0 .05 level between several aggregates (F= 38.67, df = 4, p value = .000 ). The OBS, DHA, DHA/DS, and DS aggregates are significantly different in length, however, the DBS does not signifi cantly differ from any aggregate except the DHA/DS (table 2 ) . ANOVA results on width (F= 19.31, df = 4, p value = .000), demonstrate a significant variation between several aggregates (table 3 ) . However, there is no significant difference in width between the OBS and the DHA, or the DHA/DS and the DS. In other words, it appears that aggregates with similar climatic trends (I.e. glacial and interstadial periods) have similar widths. Length appears to be more subjected to the individual aggregate. This suggests that there may be di fferent factors controlling fragmentation temporally.

PAGE 63

51 Aggregate Compared aggregate Mean Difference (mm) Standard Error P value OBS DHA 2.33 0.52 0 DHA/DS 8.38 0.74 0 DS 4.84 0.6 0 DHA DHA/DS 6.05 0.67 0 DS 2.51 0.5 0 DHA/DS DS 3.53 0.72 0 DBS 5.87 1.89 0.01 Table 2 : Tukey s post hoc te st on ANOVA results for length . Only significantly different results were included. Aggregate Compared aggregate Mean Difference (mm) Standard Error P value OBS DHA/DS 3.11 1.07 0.03 DS 2.92 0.88 0.008 DBS 20.95 2.54 0 DHA DBS 19.86 2.5 0 DHA/DS DBS 17.84 2.6 0 DS DBS 18.03 2.53 0 Table 3 : Tukey s post hoc test on ANOVA res u lts for width. Only significantly different results were included. Fragment Size

PAGE 64

52 5.1 Box plot showing the lengths and widths of faunal specimens in each aggregate . Boxes represent one standard deviation and whiskers represent two standard deviations , while individual points represent outliers . The analysis was run on all specimens (N= 2434). For individual aggregate counts refer to table 1 (previous chapter). This pattern of high fragmentation persisted in the analysis of the circumference of long bone shafts and epiphyses (N= 733). Fewer than 10% of the long bone shafts in each aggre gate had a complete (100%) circumference (fig. 5.2 ). While 70% or greater of each aggregate had less than 50% circumference (i.e. the shaft fragment would not wrap around more than 50% of a long bone), each aggregate had far fewer fragments with a circumfe rence >50% of the shaft. Smaller fragments frequently occur in scenarios where marrow extraction is required (Bunn, 1983; Marean and Spencer, 1991) . Previous analysis suggests that intensive human bone processing may result in higher frequencies of > %50 circumference (Bunn, 1983) . Overall, both datasets indicate highly fragmented assemblage, across all aggregates. Determining the cause of the fragmentation requires a critica l look at the fracture patterns and surface modification, to assess the potential taphonomic, anthropogenic or carnivore actions that occurred at KEH 1.

PAGE 65

53 5.2 F requency of long bone circumference by aggregate . (N=733) Element and Animal class representation Size class While most of the faunal remains were so badly fragmented as to be unidentifiable, we were able to identify 35% (N= 845) of the specimens to size class (table 4) . Size two animals are consistently well represented across the aggregates, c onsisting of 30 40% of the identified specimens (fig. 5.3 ). During the interstadial occupation (~29 ka 32 ka ), there is less of a reliance on size one (>15%) animals, while greater numbers of both size three and four (>50% combined) are transported to the cav e. Both trends are reversed during the subsequent glacial period (~26 ka 19 ka ). As the frequency of size 4 mammals declines (from 17% to 14 % to 4%), the frequency which size 1 mammals are recovered increases. A notable increase in size one animals occurs du ring the OBS (from 13% to 38%), at the height of the last glacial maximum, possibly indicating a shift in local environment. T his might also be due to 9 47 11 26 1 123 228 122 117 8 4 2 3 8 1 9 3 1 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% OBS DHA DHA/DS DS DBS % LONG BONES STRATIGRAPHIC AGGREGATE Circumference of Long Bone Shafts N/A <50% >50% 100%

PAGE 66

54 carnivore practices, as some of the microfauna in this aggregate have been identified as part of a genet midden (Cleghorn, pers. comm) . 5.3 Graph of the frequency of specimens in each size class (N= 845) . Strat igraphic Agg regate Micromammal Size 1 Size 2 Size 3 Size 4 Non ID Total OBS 4 71 72 36 7 285 475 DHA 3 39 114 104 41 826 1127 DHA/DS 1 12 61 64 24 101 263 DS 23 54 69 29 361 536 DBS 7 10 6 2 7 32 Total 8 152 311 279 103 1580 2433 Table 4 : Specimen counts of s ize class by aggregate Number of Individual Specimens (NISP) The frequency of recovered elements varies across stratigraphic aggregate. Around 39% (N= 944) of the assemblage could be identified to skeletal element, and 34% of the assemblage (N=808) are from the high survival set , or elemen ts most likely to survive the taphonomic processes (fig. 5.4 ). High survival elements are strongly represented across the 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% OBS DHA DHA/DS DS DBS % OF SPECIMENS Animal size class (after Brain 1981) Size 1 Size 2 Size 3 Size 4

PAGE 67

55 aggregates, composing 28 53% of the specimens (fig. 5.4 ). T he frequency of l ow survival elements varied by aggregate . In most aggregat es, low survival elements are less than 10% of the specimens, however, these elements make up 31% of the fauna in the DBS . Unfortunately , the small sample size of the DBS precludes any further conclusions regarding element representation . The DHA has the lowest percentage of low survival elements (3%), possibly indicating differential destruction due to carnivore activity or other post depositional processes like burning . 5.4 F requency of high and low s urvival skeletal elements based on Lam et al. ( 1988 ). No data refers to specimens for which this analysis was not completed (N= 2434) . There is no difference when comparing the overall frequency of high survival and high utility elements (elements attac hed to significant portions of meat, and thus more likely to be transported) across the assemblage (fig. 5.4 , Appendix A ). However, elements which are considered high utility and high survival do not always align (Binford, 1978; Lam et al., 1998; Cleghorn and Marean, 2004) . Elements in cluded in both categories are: the humerus, 182 313 139 165 9 43 33 24 25 11 226 767 94 335 10 20 15 5 7 2 0% 10% 20% 30% 40% 50% 60% 70% OBS DHA DHA/DS DS DBS % OF SPECIMENS STRATIGRAPHIC AGGREGATE Skeletal element representation by high and low survival from Lam et al. (1998) High survival Low survival Non-ID No Data

PAGE 68

56 radi us ulna, tibia, and femur. In other words, longbones, with their hefty marrow content and robust attached muscles, are more likely to be transported to and recovered from archaeological sites. A complete list of KEH 1 faunal skeletal elements by size and their survival/utility status can be found in appendixes A and B . In each aggregate, longbone fragments (LBFs) form a substantial majority of the identifiable specimens ( Appendix A ). Every aggregate contains s izable amounts of high utility elements, mostly resulting from the large numbers of LBFs. Ribs are the second most frequently identified element, although they are considered a low survival element. Other low survival, high utility elements including scapu la, vertebrae, and innominates were recovered ( Appendix A ), although they make up relatively small percentages of the identified fauna. Low utility bones are present in all aggregates (fig. 5.5 , Appendix B ), although in low frequencies (1% 18%). These consist of cranial elements and the distal portions of the limbs, neither of which offer significant portions of nutrition (Cleghorn and Marean, 2004) . 100 92 67 50 11 26 109 77 72 5 55 102 16 43 1 12 12 0 5 1 12 11 2 16 2 18 20 1 4 0 0% 10% 20% 30% 40% 50% 60% OBS DHA DHA/DS DS DBS % OF SPECIMENS Skeletal element representation by High and Low utility bones, from Binford (1980) High Utility Size 1/2 High Utility Size 3/4 High Utility size Non-ID Low Utility Size 1/2 Low Utility size 3/4 Low Utility size Non-ID

PAGE 69

57 5.5 Frequency of high and low utility elements based on Binford ( 1980 ) (N= 2434). Taxonomic representation The highly fragmented nature of the assemblage also impacted the ability to taxonomically identify specimens. Approximately 60% of the specimens were only identifiable as mammal (N= 1472), while only 4% could be identified to family or genus. Diverse anima l families have been recovered from KEH 1, including B ovidae, E quidae, C arnivora, A ves, Testudinidae, and P etromuridae (table 5 ). Out of these, bovids are the most frequently identified family group (more than 50% of the identified specimens). Two identifications were made to the genus level, P rocavia capensis ( in the OBS aggregate ), and raphicerus sp. ( in the DS aggregate ). For MNI and MAU, analysis was restricted to bovids, and the specimens were separated by size class (1 and 2 , and 3 and 4) (Appendix es B and C ). T he small fraction of identifiable specimens hindered any ability to adequately compare specimens (only size 3 and 4 bovids in the DS ha ve a MAU higher than 1) , however, there are changes in the MAU between sizes classes across aggregates. These follow the trends discussed in the paragraph above.

PAGE 70

58 Taxon GL ACIAL OBS TRA NS . D H A TRANSITION DHA/DS INTER STADIAL DS GLAC IAL DBS Total Bovid 11 23 1 23 5 63 Equid 1 1 Ungulate 9 1 1 11 Raphicerus sp. 1 1 Carnivore 4 1 5 Procavia capensis 4 4 Aves 6 1 7 Tortoise 2 3 5 Rodent 4 1 5 Microfauna 6 6 Terrestrial Mammal 2 2 Mammal 344 583 258 266 21 1472 No Data 14 10 2 5 1 32 Non ID 79 499 2 235 4 819 Total 475 1127 263 536 32 2433 Table 5 : Number of Individual Specimens by t axonomic representation and by stratigraphic aggregate. B urning The frequency of burned faunal remains varies between aggregate (fig. 5.6 ). Unsurprisingly, the highest frequencies of burned material were recovered from the DHA, which is comprised of more than fifty stacked hearth units. Around 55% of the faunal material from this aggregate is heat altered or burned. The other glacial aggreg ate, the OBS, has similarly high frequencies of heat altered bone. However, the specimens in the DBS, DS and DHA/DS have frequencies of burning lower than 40% . The reasons for this pattern are not entirely clear, although excavations suggest that the DBS ( glacial) represents a low intensity occupation (to date, one hearth has been located in this aggregate). Accumulations during the DS appear much more intense but lack the structured and conspicuous hearth units

PAGE 71

59 that characterize the DHA. It appears that fi re was being utilized during this time, but was either less intense, located elsewhere (possibly down the slope or in the back of the cave), or the evidence of hearths have been destroyed by other processes. It is interesting to hypothesize that fires were being lit on the beach, less than 10 km distant , but the evidence suggests that some level of burning occurred at KEH 1. At any rate, the DHA represents a new use of the cave from the previous interglacial period, and this is reflected in the heat alterat ion of the fauna. The burning process could have affected the survivability of the bones, possibly contributing to the higher degrees of fragmentation (particularly dry fragmentation) observed in the OBS and DHA. 5.6 F requencies of maximum burning stage by aggregate (N= 2434). 249 503 168 333 18 29 156 3 75 0 165 335 72 89 8 29 121 9 35 3 3 12 3 4 3 0% 10% 20% 30% 40% 50% 60% 70% OBS DHA DHA/DS DS DBS % BURNED SPECIMENS Maximum Burning Stage Unburnt Heat altered Calcined Carbonized No data

PAGE 72

60 Long Bone survival In order to determine the degree of nutrient processing, long bone fragments were analyzed for patterns of differential destruction (N= 733). Epiphyseal destruction at KEH 1 is highly variable across the stratigraphic aggregates, reflecting different degrees of post depositional destruction (fig. 5.7 , table 6 ). Both epiphysis and near epiphysis shafts ( sensu Blumenschin ne, 1995 ) include trabecular bone, making them highly susceptible to carnivore consumpt ion or taphonomic destruction (Marean and Spencer, 1991; Marean et al., 2000; Pickering et al., 2004) . Low epiphyseal survival relative to shafts in the OBS is consistent w ith experimental data that includes carnivore ravaging (Marean et al., 2000) . Epiphyseal survival in the OBS, and to some extent in the DHA, are consistent with human marrow processing experiments (Marean et al., 2000) . However, the other major stratigraphic aggregates (the DBS and DS) and the transitional DHA/DS have much higher rates of epiphyseal survival (fig. 5.7 ). These high frequencies (greater than 50%) have no experimental analogue, possibly suggesting that shaft fragments might be so fr agmented as to be unidentifiable, skewing the data in favor of more identifiable epiphyseal ends. Non ID fragments make up 30% 63% of the specimens from these aggregates (fig. 5.7 ). This may also imply that the faunal remains were not exposed to carnivore ravaging.

PAGE 73

61 5.7 Relative f requencies of longbone shaft , near epiphysis, and epiphyse al fragments (N= 733) . Some fragments were identified as longbones, but type was not distinguished. Aggregate No Data Epiphysis Near Epiphysis Shaft Shaft Total DBS 3 2 5 10 DHA 4 16 64 199 283 DHA/DS 1 42 47 47 137 DS 2 36 43 73 154 OBS 2 6 18 123 149 Total 9 103 174 447 733 Table 6 : Count of Longbone shafts (thick cortical), Near Epiphyses (shafts with some trabecula r bone), and Epiphyses (longbone ends). Fragmentation morphology Overall, post depositional breakage is quite high, generally falling outside the confidence intervals from the experimental data by Marean et al. ( 2000 ) . However, while 73% 95% 96% 82% 70% 34% 47% 50% 12% 22% 34% 31% 20% 27% 5% 4% 4% 5% 31% 23% 30% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Hominid only Hominid to carnivore Carnivore only OBS DHA DHA/DS DS DBS % BONE TYPE EXPERIMENTAL DATA Long Bone Shaft vs. Epiphysis representation shaft near epiphysis epiphysis KEH 1 (Based on Marean et al., 2000)

PAGE 74

62 most of the aggregate frequencies cluster between 10 and 20%, the DBS has much higher f requencies of transverse fracture outlines (fig. 5.8 , table 7 ). This further suggests that this aggregate was subjected to higher levels of post depositional destruction, possibly due to trampling and/or compression as the site was occupied for the next 30 40 thous and years. In contrast, nutritive fractures, suggesting marrow consumption by hominins or carnivores, are less frequent at Knysna than the experimental data (fig. 5 . 9 ). This may be a result of less intensive marrow processing, or possibly smaller carnivore s with reduced jaw strength when compared to hyenas (Hodgkins and Marean, 2017) . Nutritive fractures are most frequent in the glacial aggregates where only terrestrial fauna is available, and less frequent during the int erstadial , presumably when marine foraging is the dominant subsistence pattern. Conversely, the DBS (glacial) has a lower frequency of green fractures, falling between the DS and DHA/DS. The breakage data suggests that post depositional damage occurred in all aggregates, and that the long bones were processed to varying extents at different times.

PAGE 75

63 5.8 Frequency of non nutritive fractures compared to data from Marean et al. 2000 (N=733). 0% 5% 10% 15% 20% 25% 30% 35% 0% 5% 10% 15% 20% 25% 30% 35% % RIGHT ANGLES % TRANSVERSE OUTLINE Non nutritive (dry) fracture shape: KEH 1 and experimental data from Marean et al. (2000) CARNIVORE ONLY HUMAN TO CARNIVORE HUMAN ONLY OBS DHA DHA/DS DS DBS 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % OBLIQUE ANGLES % CURVED OUTLINES Nutritive (green) fracture shape: KEH 1 and experimental data from Marean et al. (2000) CARNIVORE ONLY HUMAN TO CARNIVORE HUMAN ONLY OBS DHA DHA/DS DS DBS

PAGE 76

64 5.9 Frequency of the nutritive fractures compared to data from Marean et al. 2000 (N= 733). StratAgg NISP Curved NISP Oblique NISP Transverse NISP Right NISP Shaft Fragments NISP Total Bone Ends OBS 148 172 40 48 149 298 DHA 218 307 93 72 283 566 DHA/DS 55 84 30 27 137 274 DS 86 131 59 49 154 308 DBS 7 7 6 3 10 20 Table 7 : Nutrient and Non nutrient breaks The distal portions of limbs offer significantly fewer marrow nutrients (Binford, 1980) . In order to assess the extent of nutritional stress, Hodgkins et al. (2016) considered the number of distal limb bones which had been processed for marrow. Relatively few bones were identified as metapodial or phalanx in each a ggregate (table 8 ). T he DS has the highest portion of nutrient fractures, and none of the distal long bones in the OBS and DS preserve non nutrient breaks. The DHA has a mix of non nutrient and nutrient fractures, however the sample size is too small to dr aw any significant conclusions. The sample size of the DBS is even smaller (N = 3), and the only fractures are non nutritive. While the sample size is small for all aggregates, it does show that some distal elements were being broken while fresh. To determ ine the agents responsible for the deposition of these and other long bone elements, we considered the surface modification. Aggregate NISP Nutrient breaks Non nutrient breaks Unbroken OBS 7 57% 43% DHA 6 60% 20% 20% DS 8 62% 38% DBS 3 66% 33% Table 8 : NISP a nd frequencies of nutritive/non nutritive fractures for distal limb bones

PAGE 77

65 The higher frequencies of green fractures when the coastline was distant ( OBS= 54% and DHA= 55%) are statistically significantly different from frequencies when the coast was , P=0 ). C ramer s V and a contingency coefficient post hoc suggest that there is a strong association between the variables (table 9) . These results suggest that the long bones were processed or ravaged more intensely during glacial periods, despite the availability of diverse terrestrial resources. Chi Square DF P value Cramer's V Contingency coefficient 2804 16 0 1 0.894 Table 9 : C hi square results . M atrix cover A sub sample of 411 faunal specimens was examined for surface modification. The ability to record surface modification rests on the clarity of the surface. If the surface is obscured by matrix or destroyed by post depositional processes, determining the agents involved becomes difficul t . Within our sample, 54% of the specimens (N= 223) had no matrix cemented to their surface. The DHA has the largest percentage of specimens with some or all surface covered, and the OBS has the fewest (Fig. 5.10) . Thus, analysis of the surface modification required removal of specimens which had their surfaces completely obscured.

PAGE 78

66 5.10 F requency of matrix cover on faunal remains (N= 411). Surface modification In order to consider the impact of human/carnivore interactions and the experimental data, all fragments with less than 20% visibility were removed from the analysis, and the comparative analysis was restrict ed to long bones within the surface subsample (N= 1 02 ) . Surface modification frequencies (percussion and tooth marks) for stratigraphic aggregates fall well below the confidence intervals from the experimental data by Marean et al. (2000) (Fig. 5 . 11 , table 10 ) . Withi n our sample, percussion marks on long bones are present in the OBS, DHA, and DS (fig. 5.11 ). They are highest in the DS ( 13 % ) . Tooth marks are absent in the DBS and DS, and infrequently present in the OBS and DHA, ( 6 % and 8 %, respectively). Tooth marks might result from human seasonal occupation of the cave, allowing carnivore access to the faunal remains. However, the low frequencies may also imply year round occupation, which would reduce carnivore access to the remains (Hodgkins et al., 2016) . Cut

PAGE 79

67 marks vary across the assemblage, with the DS having a substantially higher fr equency ( 40 %) than the OBS ( 3 %), DHA ( 16 %) and DBS (0%). Aggregate / Mark OBS / PM OBS / TM DHA / PM DHA / TM DS / PM DS / TM Long Bone Fragment 1 2 1 1 Metapodial 1 1 Tibia 1 1 Total 2 3 1 1 2 0 Table 10 : S urface modification on long bones. PM = percussion mark; TM = tooth mark . 5.11 F requencies of tooth and percussion marks compared to data from Marean et al. ( 2000 ) (N= 102). In all instances, the ANCOVA found the number of marks to be non significant. However, the small number of fragments with surface modifications (PM=5, TM=4, and

PAGE 80

68 CM= 11 ) in the sample may suggest that there are insufficient degree of freedom based on the number of variables in this analysis. A larger surface modification sample is needed to tackle the question of how the long bones are being broken. Aggregate Tooth Percussion Cut N OBS 6% 3% 3% 63 DHA 8% 8% 16% 12 DS 0% 13% 40% 22 DBS 0% 0% 0% 5 Table 11 : frequencies of Lone Bone surface modification within the subsample Figure 5.12 Count of the specimens with surface modification by aggregate . One specimen in the DS had both a percussion mark and tooth mark (N= 12 in a sample of 102).

PAGE 81

69 CHAPTER VI DISCUSSION Subsistence strategies of KEH 1 The analyses suggest that climatic shifts impacted human access to terrestrial and marine resources, and thus, how terrestrial resources were utilized. Initially, during the ~46 ka 34 ka occupation, early humans hunted terrestrial mammals. A much more intense occupation occurred as the coastline drew near the cave ~34 ka 29 ka , although foragers intensely utiliz ed the marine resources they continued to prey upon terrestrial mammals. When the coastline shifted, exposing the Paleo Agulhas, site use intensified, as d id use of terrestrial fauna. Foragers acquired more diverse species as the glacial progresse d , and there was an increasing the number of smaller animals brought to the site. Additionally, foragers appear to have begun processing their prey more intensely during the Last Glacial Maximum.

PAGE 82

70 Figure 6.1 : Coastal distance from KEH 1 based on modeling by Fisher et al. (2010) , and changes in the frequencies of the animal size classes, fracture patterns, and surface modification. Our analysis of size class highlights several trends concerning transport decisions. We collapsed sizes one/two, and three/four to describe transport decisions. High utility bones (which, in many cases, are also high survival bones) are more likely to be transported during the entire occupation sequence. H owever, t he size of the animal impacted the frequency with which it was brought to the cave . During the DBS, when foragers focused on ter restrial fauna, more size one and two fauna was carried to the cave overall. As the coastline progressed towards the cave, the trend reversed, and the DS has greater frequencies of larger ( size three and four ) fauna, including low utility elements. During the transitional DHA/DS

PAGE 83

71 and into the DHA as the coastline recedes, both size groups are transported frequently, although size three and four are still prioritized. When the coastline reaches nearly 70 kilometers distant, smaller animals were utilized more frequently . T hus, t hroughout the Last Glacial Maximum , low utility elements of all sizes were transported at similar frequencies . It is interesting to note the spike in low utility elements of larger size mammals during the interstadial period. It coul or elements transported along with the nutrient heavy portions. This is referred to as the (Faith and Gordon, 2007) , and may indicate either a relatively adjacent kill site or a large hunting party transporting all elements to the site (Schoville and Otárola Castillo, 2014) . Currently, the Knysna Heads ( < 2km west of KEH 1) sit at the intersection of multiple waterways, including an ocean passage through the rocky cliffs on which KEH 1 is perched, the lagoon, and the Knysna river. Phytolith analysis suggests that the vegetation during the interglacial ~34 ka 29 ka containe d large numbers of C4 water loving species, making it likely that the vegetation and landscape were similar to current conditions (Cleghorn et al., 2018) . Possibly a landscape of this nature would have attracted animals and humans, thus making it easier (in terms of distance) to transport all portions of larger carcasses . I t is also p ossible that early humans cleaned the site by discard ing the larger pieces of fauna outside the cave. T he steep slope leading to the cave would have made it easy to rapidly remove skeletal elements from the cave , especially considering the proximity of the hearths to the entrance. N utritional stress and a distant coastli ne Were the foragers of KEH 1 nutritionally stressed? Analysis indicates that this is a possibility. A lthough the increase i n smaller prey may be a result of the Genet midden

PAGE 84

72 (Cleghorn, pers. comm), c utmarks on one bird and two carnivore elements (~19 ka ) suggest that humans obtained and consumed meat from diverse sources. L arge ungulates are frequently discussed in the context of early human subsistence practices, smaller animals, including birds, are frequently overlooked (Val et al., 2016) . It has been argued that population s will intensify their use of smaller animals when the larger ones have been over exploited (Klein and Cruz Uribe, 1983; Thompson and Henshilwood, 2014b; Armstrong, 2016) . It is possible that an increased taxonomic diversit y during the glacial period (~26 ka 19 ka ) suggests that the foragers of KEH 1 were nutritionally stressed. However, it is also possible that this is an opportunistic exploitation of a diversified resource base, as the Paleo Agulhas plain is exposed. Nearby si tes suggest that earlier foragers were utilizing a broad resource base (Wadley, 2010; Thompson and Henshilwood, 2014b; Armstrong, 2016; Val et al., 2016) , although in several cases, this has been linked to compensation for loss of marine resources. The apparently reduced capture rate of large mammals later in the sequence is noteworthy, as the Paleo Agulhas is reconstructed as offering space for large, migratory mammals (C opeland et al., 2016) . Thus, if the Last Glacial Maximum offers extensive bioavailable pasture south of the cave, it is curious as to why KEH 1 foragers do not intercept these animals as frequently during the DHA and OBS, as the proceeding interstadial period. The answer may lie in the territoriality of LSA hunter gatherers. If the Paleo Agulhas is carved into territories, and the large mammals are migrating E W (as proposed by Copeland et al., 2016), then the foragers might only hunt them s easonally, and focus on smaller prey at other times. Additionally, it has been proposed that smaller animals (rabbit, tortoise, ect) offer a higher return in balanced micro nutrients, which would be lacking in

PAGE 85

73 larger game (Brown et al., 2011) . Smaller game may have offered glacial foragers a more diverse array of nutrients and offset the loss of marine resources . The fauna was fragmented through a combination of human action and post depositional processes. Initially, the smaller fragment size (in the DBS) is related to taphonomic effects. However, the subsequent interglacial period h as reduced post depositional breakage, and overall lower frequencies of nutrient fractures. The reduction in length during the glacial period (~26 ka 19 ka ), results from statistically significantly higher frequencies of nutrient fractures. Some level of non nutrient breakage is occurring, possibly due to the intensity of site use. Additionally, burned bones may be more susceptible to fragmentation (Stiner et al., 1995) . This suggests that there may have been multiple factors influe ncing the fragmentation of long bones. The surface modification data presents a more complex picture of the site. Tooth marks are absent from our sample of the DS, while present (albeit not to a large extent) in the DHA and OBS. However, anthropogenic mod ifications on bones are more frequent (at least double) in the DS. Moreover, some of these occur on the distal portions of the limbs. If early humans were nutritionally stressed, it is expected that they would utilize even the distal elements of the limb, which do not offer large amounts of protein (Hodgkins et al., 2016) . It appears t hat both humans and carnivores are responsible for the nutrient fractures during the Last Glacial Maximum at KEH 1. This appears to be a shift from the interstadial (the DS) , when carnivores had less access to the site, and humans appear to be responsible for the nutritive fractures. Taken together, does the increased focus on smaller animals during the OBS , coupled with more intense nutritive processing of long bones, demonstrate nutritional stress during

PAGE 86

74 the glacial period? The previous literature sugges ts that these are valid proxies for gauging human subsistence strategies (Thompson and Henshilwood, 2014b; Armstrong, 2016; Hodgkins et al., 2016) . However, th e frequency of anthropogenic surface modification in the interstadial exceeds the amount in the subsequent glacial periods. This specific proxy suggests that hominins were more stressed for nutrients during the interstadial ( per Hodgkins et al., 2016) . However, it should be noted that s urface analysis has only been conducted on a small subsample of the long bone assemblage. An expanded analysis of the surface modification may alter this conclusion (for example, a subsample N= 1 2 for the DHA when the total number of faunal remains in the DHA > 9 ,000, is not exactly representative). However, if a higher sample size demonstrates that the significantly higher results of nutrient related fragmentation are more frequently a result of anthropogenic activity, then it would suggest that early human s are more stressed for marrow during the glacial period. Increasing the sample size will reveal higher frequencies of surface modification in all aggregates and permit a more complete understanding of the human subsistence. Th e explanation that multiple subsistence strategies were utilized by foragers is certainly plausible. Lower levels of cut/percussion marks in the peak glacial period, interspersed with tooth marks, may suggest that humans were moving across the landscape, but returning consis tently to KEH 1. This m eans that carnivores would have been able to access the site . The location, with diverse mammalian resources, provide d substantial shelter and offer significant terrestrial resources for both humans and carnivores . While long bones were processed more frequently, it may be through a combined effort of humans and carnivores. If this is the case, then humans were still highly mobile, possibly traveling from

PAGE 87

75 the coastline into the interi or and making use of the entire offerings of the Paleo Agulhas plain.

PAGE 88

76 CHAPTER VII CONCLUSION The Paleo Agulhas plain provided early humans with alternating terrestrial and marine resources. KEH 1 offers a glimpse into human subsistence during a tumultuous climatic period, when the coastline and ecology underwent rapid change. KEH 1 is one of a handful of sites preserving occupational sequences throughout MIS2 , providing a unique opportunity to understand early human foraging . Taphonomic analysis shows a significant difference of the fragmentation (size of fauna) between the aggregates . Aggregates with the highest fragmentation are also those with the highest frequency of fire modification. Bones which are high survival (dense and compact) are more frequently identified to element across all aggregates . These bones also contain large amounts of meat/marrow, making them highly desirable (Binford, 1978) . Long bone fragments are likewise fragmentated, and during the later sequences (DHA and OBS) this is more frequently associated with fresh breaks, as observed in cases of marrow processing. T his p ilot assessment of the fauna provides some compelling initial results. Over time, the fauna brought to the cave trend towards smaller bodied animals. Evidence suggests that the interstadial (late MIS3) offered substantial resources, with shellfish transported to the cave, while foragers continued to consume a variety of mammals of differing body sizes. The increased marrow processing represented in the glacial period is consistent with nutritional stress. Only a small subsample was assessed for surface modification, and the specimens with cut/percussion/tooth marks were so few as to be statistically insignificant. Further research is necessary to clarify the extent to which this represents hominin or carnivore

PAGE 89

77 activity . This research can contribute to the reconstruction of hominin behavior on the paleo A gulhas plain .

PAGE 90

78 REFERENCES Andrews, P., Cook, J., 1985. Natural Modifications to Bones in a Temperate Setting. Man. 20, 675 691. Armstrong, A., 2016. Small mammal utilization by Middle Stone Age humans at Die Kelders Cave 1 and Pinnacle Point Site 5 6, Western Cape Province, South Africa. Journal of Human Evolution. 101, 17 44. Arnold, J., 1992. Complex Hunter Gatherer Fishers of Pr ehistoric California: Chiefs, Specialists, and Maritime Adaptations of the Channel Islands | American Antiquity | Cambridge Core [WWW Document]. URL https://www.cambridge.org/core/journals/american antiquity/article/div classtitlecomplex hunter gatherer fi shers of prehistoric california chiefs specialists and maritime adaptations of the channel islandsdiv/93BE358F07E49A652557DAB7543C9B98 (accessed 4.10.18). Athreya, S., Ackermann, R., in press. Colonialism and narratives of human origins in Asia and Africa. 10.31730/osf.io/jtkn2. Current Anthropology: Vol 43, No 1 [WWW Document]. URL https://www journals uchicago edu.aurarialibrary.idm.oclc.org/doi/full/10.1086/33829 2 (accessed 5.6.18). Bar Matthews, M., Marean, C.W., Jacobs, Z., Karkanas, P., Fisher, E.C., Herries, A.I., Brown, K., Williams, H.M., Bernatchez, J., Ayalon, A., 2010. A high resolution and continuous isotopic speleothem record of paleoclimate and paleoen vironment from 90 to 53 ka from Pinnacle Point on the south coast of South Africa. Quaternary Science Reviews. 29, 2131 2145. Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. Paleobiology. 4, 150 162. Behrensmeyer, A.K., Kidwell, S.M., Gastaldo, R.A., 2000. Taphonomy and paleobiology. Paleobiology. 26, 103 147. Beier, J., Anthes, N., Wahl, J., Harvati, K., 2018. Similar cranial trauma prevalence among Neanderthals and Upper Palaeolithic modern humans. Nature. 563, 686. Ber ger, T.D., Trinkaus, E., 1995. Patterns of trauma among the Neandertals. Journal of Archaeological Science. 22, 841 852. Binford, L.R., 1962. Archaeology as Anthropology. American Antiquity. 28, 217 225. Binford, L.R., 1978. Nunamiut: Ethnoarchaeology. New York: Academic Press.

PAGE 91

79 Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity. 45, 4 20. Binford, L.R., 1985. Human ancestors: changing views of their behavior. Journal of Anthropological Archaeology. 4, 292 327. http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/41053227?pq origsite=summon (accessed 9.14.17). Blumenschine, R .J., 1986. Early hominid scavenging opportunities: implications of carcass availability in the Serengeti and Ngorongoro ecosystems. British Archaeological Reports. Blumenschine, R.J., 1989. A landscape taphonomic model of the scale of prehistoric scavengin g opportunities. Journal of Human Evolution. 18, 345 371. Blumenschine, R.J., 1995. Percussion marks, tooth marks, and experimental determinations of the timing of hominid and carnivore access to long bones at FLK Zinjanthropus, Olduvai Gorge, Tanzania. Jo urnal of Human Evolution. 29, 21 51. Blumenschine, R.J., Marean, C.W., Capaldo, S.D., 1996. Blind tests of inter analyst correspondence and accuracy in the identification of cut marks, percussion marks, and carnivore tooth marks on bone surfaces. Journal o f Archaeological Science. 23, 493 507. Bousman, C.B., 2005. Coping with risk: Later stone age technological strategies at Blydefontein Rock Shelter, South Africa. Journal of Anthropological Archaeology. 24, 193 226. Bousman, C.B., Brink, J.S., 2018. The em ergence, spread, and termination of the Early Later Stone Age event in South Africa and southern Namibia. Quaternary International, In Honour of James Brink. 495, 116 135. Brain, C.K., 1969. The contribution of Namib Desert Hottentots to an understanding o f australopithecine bone accumulations. Scientific Papers of the Namib Desert Research Station. 1969, 13 22. Brain, C.K., 1981. The hunter or the hunted. An Introduction to African Cave Taphonomy, The University of Chicago press, Chicago et Londres. Brown, J.L., 1964. The evolution of diversity in avian territorial systems. The Wilson Bulletin. 160 169. Brown, K., Fa, D.A., Finlayson, G., Finlayson, C., 2011. Small game and marine resource exploitation by Neanderthals: the evidence from Gibraltar. In: Trekk ing the Shore. Springer, pp. 247 272. Brumfiel, E.M., 1992. Distinguished Lecture in Archeology: Breaking and Entering the Ecosystem Gender, Class, and Faction Steal the Show. American Anthropologist. 94, 551 567.

PAGE 92

80 Brumm, A., Moore, M.W., 2005. Symbolic rev olutions and the Australian archaeological record. Cambridge Archaeological Journal. 15, 157 175. Bunn, H.T., 1983. Comparative analysis of modern bone assemblages from a San hunter gatherer camp in the Kalahari Desert, Botsuwana, and from a spotted hyena den near Nairobi, Kenya. Animal and Archaeology, 1. Hunters and Their Prey; BAR International Series. 163, 141 148. Bunn, H.T., Bartram, L.E., Kroll, E.M., 1988. Variability in bone assemblage formation from Hadza hunting, scavenging, and carcass processin g. Journal of Anthropological Archaeology. 7, 412 457. Butzer, K.W., 19 73. Geology of Nelson Bay Cave, Robberg, South Africa. The South African Archaeological Bulletin. 28, 97 110. Capaldo, S.D., 1997. Experimental determinations of carcass processing by Plio Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Oldu vai Gorge, Tanzania. Journal of Human Evolution. 33, 555 597. Capaldo, S.D., Blumenschine, R.J., 1994. A quantitative diagnosis of notches made by hammerstone percussion and carnivore gnawing on bovid long bones. American Antiquity. 59, 724 748. Chase, B.M zone. Earth Science Reviews. 84, 103 138. Clark, A., 1997. The MSA/LSA Transition in Southern Africa: New Technological Evidence from Rose Cottage Cave on JSTOR [WWW Docu ment]. URL http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/3889076?casa_token=TkfUt6lLJXYA AAAA:mMeocsx4xrZiFb5RDnU81SidGfWnx7jkXoQRuub8XjIhlLXh_ED kthRODqjOi5 uXgJM7EWzF7PCIAXyoUw9HiF5EQoQ6tXHLk4WBQOilxZo3g_wQ (accessed 3.19.18). Clark, J.L., 2017. within Middle Stone Age industries. Journal of Human Evolution. 107, 49 70. Clark, J.L., Kandel, A.W., 2013. The Evolutionary Implications of Variation in Human Hunting Strateg ies and Diet Breadth during the Middle Stone Age of Southern Africa | Current Anthropology: Vol 54, No S8 [WWW Document]. URL https://www journals uchicago edu.aurarialibrary.idm.oclc.org/doi/abs/10.1086/673386 (accessed 4.1.18). Clark, J.L., Ligouis, B., Stone Age deposits at Sibudu (South Africa): behavioral and taphonomic implications. Journal of Archaeological Science. 37, 2650 2661. Clark, J.L., Plug, I., 2008. Animal exploitati on strategies during the South African Middle Stone Age: Howiesons Poort and post Howiesons Poort fauna from Sibudu Cave. Journal of Human Evolution. 54, 886 898. Cleghorn, N., Keller, H., McGrath, J.R., Shelton, C.M., Watson, S., Schoville, B.J., Peart, D ., Bickerton, N., Esteban, I., 2018. Assessing the development of societal complexity at the Middle to

PAGE 93

81 Later Stone Age transition in the context of the Economic Defensibilty Model: Evidence from Knysna, coastal South Africa. Cleghorn, N., Marean, C.W., 200 4. Distinguishing selective transport and in situ attrition: a critical review of analytical approaches. Journal of Taphonomy. 2, 43 67. Collins, B., Steele, T.E., 2017. An often overlooked resource: Ostrich (Struthio spp.) eggshell in the archaeological r ecord. Journal of Archaeological Science: Reports. 13, 121 131. Colonese, A.C., Mannino, M.A., Bar Yosef Mayer, D.E., Fa, D.A., Finlayson, J.C., Lubell, D., Stiner, M.C., 2011. Marine mollusc exploitation in Mediterranean prehistory: An overview. Quaternar y International, Shell Midden Research: An Interdisciplinary Agenda for the Quaternary and Social Sciences. 239, 86 103. Compton, J.S., 2011a. Pleistocene sea level fluctuations and human evolution on the southern coastal plain of South Africa. Quaternary Science Reviews. 30, 506 527. Compton, J.S., 2011b. Pleistocene sea level fluctuations and human evolution on the southern coastal plain of South Africa. Quaternary Science Reviews. 30, 506 527. Copeland, S.R., Cawthra, H.C., Fisher, E.C., Lee Thorp, J.A., Cowling, R.M., le Roux, P.J., Hodgkins, J., Marean, C.W., 2016. Strontium isotope investigation of ungulate movement patterns on the Pleistocene Paleo Agulhas Plain of the Greater Cape Floristic Region, South Africa. Quaternary Science Reviews. 141, 65 84 . Conus shells from Border Cave. Journal of Human Evolution. 93, 91 108. Higham, T.F.G., Colombini, M.P., Beaumont, P.B., 2012a. Early evidence of San material culture represented by organic artifacts from Border Cave, South Africa. Proceedings of the National Academy of Sciences. 109, 13214 13219. ifkin, R.F., 2012b. Technological, elemental and colorimetric analysis of an engraved ochre fragment from the Middle Stone Age levels of Klasies River Cave 1, South Africa. Journal of Archaeological Science. 39, 942 952. Deacon, H.J., 1993. Planting an Ide a: An Archaeology of Stone Age Gatherers in South Africa. The South African Archaeological Bulletin. 48, 86 93. Deacon, J., 1978. Changing patterns in the late Pleistocene/early Holocene prehistory of Southern Africa as seen from the Nelson Bay Cave Stone artifact sequence. Quaternary Research. 10, 84 111. year old Middle Stone Age levels at Blombos Cave, South Africa: implications for the origin of symbolism and lang uage. Antiquity. 75, 309 318.

PAGE 94

82 with archaeological data. American Antiquity. 62, 629 651. Domínguez Rodrigo, M., Bunn, H.T., Mabulla, A.Z.P., Baquedano, E., Uribelarrea, D., Pérez González, A., Gidna, A., Yravedra, J., Diez Martin, F., Egeland, C.P., Barba, R., Arriaza, M.C., Organista, E., Ansón, M., 2014. On meat eating and human evolution: A taphonomic analysis of BK4b (Upper Bed II, Olduvai Gorge, Tanz ania), and its bearing on hominin megafaunal consumption. Quaternary International, The Evolution of Hominin Behavior during the Oldowan Acheulian Transition: Recent Evidence from Olduvai Gorge and Peninj (Tanzania). 322 323, 129 152. Domínguez Rodrigo, M. , de Juana, S., Galán, A.B., Rodríguez, M., 2009a. A new protocol to differentiate trampling marks from butchery cut marks. Journal of Archaeological Science. 36, 2643 2654. Domínguez Rodrigo, M., Mabulla, A., Bunn, H.T., Barba, R., Diez Martin, F., Egelan d, C.P., Espílez, E., Egeland, A., Yravedra, J., Sánchez, P., 2009b. Unraveling hominin behavior at another anthropogenic site from Olduvai Gorge (Tanzania): new archaeological and taphonomic research at BK, Upper Bed II. Journal of Human Evolution. 57, 26 0 283. Domínguez Rodrigo, M., Musiba, C.M., 2010. How accurate are paleoecological reconstructions of early paleontological and archaeological sites? Evolutionary biology. 37, 128 140. Dusseldorp, G.L., 2010. Prey Choice During the South African Middle Sto ne Age: Avoiding Dangerous Prey or Maximising Returns? African Archaeological Review. 27, 107 133. marine resources during the South African Middle Stone Age. Southern African Humanities. 25, 105 135. Enloe, J.G., 2012. Middle Palaeolithic Cave Taphonomy: Discerning Humans from Hyenas at Arcy sur Cure, France. International Journal of Osteoarchaeology. 22, 591 602. Faith, J.T., 2008. Eland, buffalo, and wild pigs: were Middle Stone Age humans ineffective hunters? Journal of Human Evolution. 55, 24 36. Faith, J.T., 2011. Ungulate community richness, grazer extinctions, and human subsistence behavior in s ScienceDirect [WWW Document]. URL http://www.sciencedirect.com.aurarialibrary.idm.oclc.org/science/article/pii/S00310182110 02276 (accessed 12.7.17). Faith, J.T., 2013. Taphonomic and paleoecological change in the large mammal sequence from Boomplaas Cave, western Cape, South Africa. Journal of Human Evolution. 65, 715 730. Faith, J.T., Gordon, A.D., 2007. Skeletal element abundances in archaeofaunal assemblages: economic utility, sample size, and assessment of carcass t ransport strategies. Journal of Archaeological Science. 34, 872 882.

PAGE 95

83 Faith, J.T., Thompson, J.C., 2018. Low survival skeletal elements track attrition, not carcass transport behavior in Quaternary large mammal assemblages. In: Zooarchaeology in Practice. S pringer, pp. 109 126. Faure, H., Walter, R.C., Grant, D.R., 2002. The coastal oasis: ice age springs on emerged continental shelves. Global and Planetary Change. 33, 47 56. Fisher, E.C., Bar Matthews, M., Jerardino, A., Marean, C.W., 2010. Middle and Late Pleistocene paleoscape modeling along the southern coast of South Africa. Quaternary Science Reviews. 29, 1382 1398. Flannery, K.V., 1968. Archaeological systems theory and early Mesoamerica. Anthropological archaeology in the Americas. 67, 87. Gifford Gon zales, D., Damrosch, David, Damrosch, Debra, Pryor, J., Thunen, R., 1985. The Third Dimension in Site Structure: An Experiment in Trampling and Vertical Dispersal | American Antiquity | Cambridge Core [WWW Document]. URL https://www.cambridge.org/core/jour nals/american antiquity/article/div classtitlethe third dimension in site structure an experiment in trampling and vertical dispersaldiv/0F9657D5DAA7DA1F1946B0E5FD3A9C09# (accessed 9.7.17). Grine, F.E., Wurz, S., Marean, C.W., 2017. The middle stone age hu man fossil record from Klasies River main site. Journal of human evolution. 103, 53 78. Hardy, B.L., Moncel, M. H., 2011. Neanderthal Use of Fish, Mammals, Birds, Starchy Plants and Wood 125 250,000 Years Ago. PLOS ONE. 6, e23768. Haws, J.A., Funk, C.L., B enedetti, M.M., Bicho, N.F., Daniels, J.M., Minckley, T.A., Denniston, R.F., Jeraj, M., Gibaja, J.F., Hockett, B.S., 2011. Paleolithic landscapes and seascapes of the west coast of Portugal. In: Trekking the Shore. Springer, pp. 203 246. Helm, C.W., McCrea , R.T., Cawthra, H.C., Lockley, M.G., Cowling, R.M., Marean, C.W., Thesen, G.H.H., Pigeon, T.S., Hattingh, S., 2018. A New Pleistocene Hominin Tracksite from the Cape South Coast, South Africa. Scientific Reports. 8, 3772. Watts, I., 2009. Engraved ochres from the Middle Stone Age levels at Blombos Cave, South Africa. Journal of Human Evolution. 57, 27 47. from the Middle Stone Age at Blombos Cave, South Africa: implications for the origins of modern human behaviour, symbolism and language. Journal of Human Evolution. 41, 631 678. Hodgkins, J., Marean, C.W., 2017. New information from old excavations: A comparative analysis of Paleolithic zooarchaeological assemblages from the Zagros Mountains. In: Biglari F, Mashkour M, Shidrang S (Eds.) The Pleistocene Archaeology of the Iranian Plateau, Iraq and the Caucasus, National Museum of Iran Paleolithic Studies Series. National Mu seum of Iran, Tehran.

PAGE 96

84 Hodgkins, J., Marean, C.W., Turq, A., Sandgathe, D., McPherron, S.J.P., Dibble, H., 2016. Climate (Dordogne Valley, France). Journal of Human Ev olution. 96, 1 18. Hodgkins, J., Roux, P. le, Marean, C.W., Penkman, K., Crisp, M., Fisher, E., Lee Thorp, J., 2018. The role of ostrich in shaping the landscape use patterns of humans and hyenas on the southern coast of South Africa during the late Pleist ocene. Multispecies Archaeology. 333 346. Hublin, J. J., Ben Ncer, A., Bailey, S.E., Freidline, S.E., Neubauer, S., Skinner, M.M., Bergmann, I., Le Cabec, A., Benazzi, S., Harvati, K., 2017. New fossils from Jebel Irhoud, Morocco and the pan African origin of Homo sapiens. Nature. 546, 289. Jacobs, Z., 2010. An OSL chronology for the sedimentary deposits from Pinnacle Point Cave 13B a punctuated presence. Journal of Human Evolution. 59, 289 305. Jacobs, Z., Roberts, R., Galbraith, R., Deacon, H.J., Grun, R. , Mackay, A., mitchell, P., Vogelsang, R., Wadley, L., 2008. Ages for the Middle Stone Age of Southern Africa: Implications for Human Behavior and Dispersal | Science [WWW Document]. URL http://science.sciencemag.org/content/322/5902/733 (accessed 5.21.18) . Jerardino, A., Marean, C.W., 2010. Shellfish gathering, marine paleoecology and modern human behavior: perspectives from cave PP13B, Pinnacle Point, South Africa. Journal of Human Evolution. 59, 412 424. Karkanas, P., Brown, K.S., Fisher, E.C., Jacobs, Z ., Marean, C.W., 2015. Interpreting human behavior from depositional rates and combustion features through the study of sedimentary microfacies at site Pinnacle Point 5 6, South Africa. Journal of Human Evolution. 85, 1 21. Kelly, R.L., 1995. The foraging spectrum. Washington, DC: Smithsonian Institution Press. Kidwell, S.M., Holland, S.M., 2002. The quality of the fossil record: implications for evolutionary analyses. Annual Review of Ecology and Systematics. 33, 561 588. Klein, R., 1976. The Mammalian Fau na of the Klasies River Mouth Sites, Southern Cape Province, South Africa on JSTOR [WWW Document]. URL http://www.jstor.org.aurarialibrary.idm.oclc.org/stable/3887730 (accessed 5.27.18). Klein, R.G., 1989. Why does skeletal part representation differ between smaller and larger bovids at Klasies River Mouth and other archeological sites? Journal of Archaeological Science. 16, 363 381. Klein, R.G., 1994. Southern Africa before the iron age. Integrative paths to the past: Paleoanthropological advances in honor of F. Clark Howell. 471 519. Klein, R.G., 1995. Anatomy, behavior, and modern human origins. Journal of World Prehistory. 9, 167 198.

PAGE 97

85 Klein, R.G., 2000. Archeology and the evolution of h uman behavior. Evolutionary Anthropology: Issues, News, and Reviews. 9, 17 36. Klein, R.G., Avery, G., Cruz Uribe, K., Halkett, D., Parkington, J.E., Steele, T., Volman, T.P., Yates, R., 2004. The Ysterfontein 1 Middle Stone Age site, South Africa, and ear ly human exploitation of coastal resources. Proceedings of the National Academy of Sciences. 101, 5708 5715. Klein, R.G., Avery, G., Cruz Uribe, K., Steele, T.E., 2007. The mammalian fauna associated with an archaic hominin skullcap and later Acheulean art ifacts at Elandsfontein, Western Cape Province, South Africa. Journal of Human Evolution. 52, 164 186. Klein, R.G., Cruz Uribe, K., 1983. Stone Age Population Numbers and Average Tortoise Size at Byneskranskop Cave 1 and Die Kelders Cave 1, Southern Cape P rovince, South Africa. The South African Archaeological Bulletin. 38, 26 30. Klein, R.G., Cruz nearby sites, Western Cape Province, South Africa. Papers in the prehistory of the wes tern Cape, South Africa. 1, 132 163. Klein, R.G., Steele, T.E., 2013. Archaeological shellfish size and later human evolution in Africa. Proceedings of the National Academy of Sciences. 110, 10910 10915. Kusimba, S.B., 2005. What Is a Hunter Gatherer? Vari ation in the Archaeological Record of Eastern and Southern Africa. Journal of Archaeological Research. 13, 337 366. Lam, Y.M., Chen, X., Marean, C.W., Frey, C.J., 1998. Bone Density and Long Bone Representation in Archaeological Faunas: Comparing Results f rom CT and Photon Densitometry. Journal of Archaeological Science. 25, 559 570. Lam, Y.M., Pearson, O.M., Smith, C.M., 1996. Chin morphology and sexual dimorphism in the fossil hominid mandible sample from Klasies River Mouth. American Journal of Physical Anthropology: The Official Publication of the American Association of Physical Anthropologists. 100, 545 557. Lee Thorp, J.A., Ecker, M., 2015. Holocene Environmental Change at Wonderwerk Cave, South Africa: Insights from Stable Light Isotopes in Ostrich E ggshell. African Archaeological Review. 32, 793 811. Loftus, E., Sealy, J., Lee Thorp, J., 2016. New radiocarbon dates and Bayesian models for Nelson Bay Cave and Byneskranskop 1: Implications for the South African later stone age sequence. Radiocarbon. 58 , 365 381. Lyman, R.L., 1984. Bone density and differential survivorship of fossil classes. Journal of Anthropological archaeology. 3, 259 299. Lyman, R.L., 2004. The Concept of Equifinality in Taphonomy. Journal of Taphonomy. 2, 15 26. Lyman, R.L., Lyman, C., 1994. Vertebrate Taphonomy. Cambridge University Press.

PAGE 98

86 Mackay, A., 2010. The Late Pleistocene Archaeology of Klein Kliphuis Rock Shelter, Western Cape, South Africa: 2006 excavations. The South African Archaeological Bulletin. 65, 132 147. Mackay, A. , Jacobs, Z., Steele, T.E., 2015. Pleistocene Archaeology and Chronology of Putslaagte 8 (PL8) Rockshelter, Western Cape, South Africa. Journal of African Archaeology. 13, 71 98. Mackay, A., Welz, A., 2008. Engraved ochre from a Middle Stone Age context at Klein Kliphuis in the Western Cape of South Africa. Journal of Archaeological Science. 35, 1521 1532. Mallick, S., Li, H., Lipson, M., Mathieson, I., Gymrek, M., Racimo, F., Zhao, M., Chennagiri, N., Nordenfelt, S., Tandon, A., 2016. The Simons genome div ersity project: 300 genomes from 142 diverse populations. Nature. 538, 201. Marean, C.W., 2010. Pinnacle Point Cave 13B (Western Cape Province, South Africa) in context: The Cape Floral kingdom, shellfish, and modern human origins. Journal of Human Evoluti on. 59, 425 443. Marean, C.W., 2011. Coastal South Africa and the Coevolution of the Modern Human Lineage and the Coastal Adaptation. In: Trekking the Shore, Interdisciplinary Contributions to Archaeology. Springer, New York, NY, pp. 421 440. Marean, C.W., 2014. The origins and significance of coastal resource use in Africa and Western Eurasia. Journal of Human Evolution, The Role of Freshwater and Marine Resources in the Evolution of the Human Diet, Brain and Behavior. 77, 17 40. Marean, C.W., 2016. The tr ansition to foraging for dense and predictable resources and its impact on the evolution of modern humans. Phil. Trans. R. Soc. B. 371, 20150239. Marean, C.W., Abe, Y., Frey, C.J., Randall, R.C., 2000. Zooarchaeological and taphonomic analysis of the Die K elders Cave 1 Layers 10 and 11 Middle Stone Age larger mammal fauna. Journal of Human Evolution. 38, 197 233. Marean, C.W., Abe, Y., Nilssen, P.J., Stone, E.C., 2001. Estimating the minimum number of skeletal elements (MNE) in zooarchaeology: a review and a new image analysis GIS approach. American Antiquity. 66, 333 348. Marean, C.W., Bar Matthews, M., Bernatchez, J., Fisher, E.C., Goldburg, P., Herries, A.I.R., Jacobs, Z., Jerardino, A., Karkanas, P., Minichillo, T., Nilssen, P., Thompson, E., Watts, I., Williams, H., 2007. Early human use of marine resources and pigment in South Africa during the Middle Pleistocene | Nature [WWW Document]. URL https://www.nature.com/articles/nature06204 (accessed 5.6.18). Marean, C.W., Cawthra, H.C., Cowling, R.M., Esler, K.J., Fisher, E., Milewski, A., Potts, A.J., Singels, E., Vynck, J.D., 2014. Stone Age people in a changing South African Greater Cape Floristic Region. In: Allsopp, N., Colville, J.F., Verboom, G.A. (Eds. ), Fynbos. Oxford University Press, pp. 164 199.

PAGE 99

87 Marean, C.W., Dominguez Rodrigo, M., Pickering, T.R., 2004. Skeletal element equifinality in of Taphonomy. 2, 69 9 8. Implications for Neanderthals and Early Modern Humans. Current Anthropology. 39, S79 S114. Marean, C.W., Spencer, L.M., 1991. Impact of Carnivore Ravaging on Zooar chaeological Measures of Element Abundance. American Antiquity. 56, 645 658. modern human behavior. Journal of Human Evolution. 39, 453 563. Mellars, P., S tringer, C., 1989. The human revolution. Edinburgh University Press Edinburgh. Milo, R.G., 1998. Evidence for Hominid Predation at Klasies River Mouth, South Africa, and its Implications for the Behaviour of Early Modern Humans. Journal of Archaeological S cience. 25, 99 133. Mitchell, J.P., 1994. Understanding the MSA/LSA transition: the pre 20 000 BP assemblages from new excavations at Sehonghong:rock shelter, Lesotho. Southern African Field Archaeology. 3, 15 25. Mitchell, P., 2008. Developing the archaeo logy of marine isotope stage 3. Goodwin Series. 52 65. Nettle, D., Gibson, M.A., Lawson, D.W., Sear, R., 2013. Human behavioral ecology: current research and future prospects. Behavioral Ecology. 24, 1031 1040. Nowell, A., 2010. Defining Behavioral Moderni ty in the Context of Neandertal and Anatomically Modern Human Populations | Annual Review of Anthropology [WWW Document]. URL https://www.annualreviews.org/doi/abs/10.1146/annurev.anthro.012809.105113 (accessed 4.9.18). N.B., 1990. Reanalysis of large mammal body part transport among the Hadza. Journal of Archaeological Science. 17, 301 316. Opperman, H., 1978. Excavations in the Buffelskloof Rock Shelter near Calitzdorp, Southern Cape. The South African Archaeological Bu lletin. 33, 18 38. Owen, R.B., Muiruri, V.M., Lowenstein, T.K., Renaut, R.W., Rabideaux, N., Luo, S., Deino, A.L., Sier, M.J., Dupont Nivet, G., McNulty, E.P., 2018. Progressive aridification in East Africa over the last half million years and implications for human evolution. Proceedings of the National Academy of Sciences. 201801357. Parkington, J., Poggenpoel, C., Halkett, D., Hart, T., 2004. Initial observations on the Middle Stone Age coastal settlement in the western Cape, South Africa. Settlement dyn amics of the middle paleolithic and Middle Stone Age. 2, 5 51.

PAGE 100

88 Pickering, T.R., Dominguez Rodrigo, M., Egeland, C.P., Brain, C.K., 2004. New data and ideas on the ar ticle. South African Journal of Science. 100, 215 219. Pickering, T.R., Marean, C.W., Dom nguez Rodrigo, M., 2003. Importance of limb bone shaft M.C. Stiner. Journal of Archaeological Science. 30, 1469 1482. Pierce, G.J., Ollason, J.G., 1987. Eight Reasons Why Optimal Foraging Theory Is a Complete Waste of Time. Oikos. 49, 111 118. Porraz, G., Parkington, J.E., Rigaud, J. P., Miller, C.E., Poggenp oel, C., Tribolo, C., Archer, W., Cartwright, C.R., Charrié Duhaut, A., Dayet, L., Igreja, M., Mercier, N., Schmidt, P., Verna, C., Texier, P. J., 2013. The MSA sequence of Diepkloof and the history of southern African Late Pleistocene populations. Journal of Archaeological Science, The Middle Stone Age at Diepkloof Rock Shelter, Western Cape, South Africa. 40, 3542 3552. Potts, R., Shipman, P., 1981. Cutmarks made by stone tools on bones from Olduvai Gorge, Tanzania. Nature. 291, 577. Rector, A.L., Reed, K .E., 2010. Middle and late Pleistocene faunas of Pinnacle Point and their paleoecological implications. Journal of Human Evolution. 59, 340 357. Renfrew, C., 1996. The sapient behaviour paradox: how to test for potential. Modelling the early human mind. 11 14. Schoville, B.J., Otárola Castillo, E., 2014. A model of hunter gatherer skeletal element transport: The effect of prey body size, carriers, and distance. Journal of human evolution. 73, 1 14. Scott, L., Neumann, F.H., 2018. Pollen interpreted palaeoen vironments associated with the Middle and Late Pleistocene peopling of Southern Africa. Quaternary International. 495, 169 184. Southernmost Africa. Current Anthropo logy. 47, 569 595. Smith, E.I., Jacobs, Z., Johnsen, R., Ren, M., Fisher, E.C., Oestmo, S., Wilkins, J., Harris, J.A., Karkanas, P., Fitch, S., Ciravolo, A., Keenan, D., Cleghorn, N., Lane, C.S., Matthews, T., Marean, C.W., 2018. Humans thrived in South Af rica through the Toba eruption about 74,000 years ago. Nature. Steele, T., Mackay, A., Fitzsimmons, K., 2016. Varsche Rivier 003: A Middle and Later Stone Age Site with Still Bay and Howiesons Poort Assemblages in Southern Namaqualand, South Africa. 64. St eele, T.E., Klein, R.G., 2013. The Middle and Later Stone Age faunal remains from Diepkloof Rock Shelter, Western Cape, South Africa. Journal of Archaeological Science, The Middle Stone Age at Diepkloof Rock Shelter, Western Cape, South Africa. 40, 3453 34 62.

PAGE 101

89 Stewart, K.M., 2010. The case for exploitation of we tlands environments and foods by pre sapiens hominins. Human Brain Evolution. The Influence of Freshwater and Marine Food Resources. 137 171. Stewart, K.M., 2014. Environmental change and hominin exploitation of C4 based resources in wetland/savanna mosaic s. Journal of human evolution. 77, 1 16. Stiner, M.C., Kuhn, S.L., 1992. Subsistence, technology, and adaptive variation in Middle Paleolithic Italy. American Anthropologist. 94, 306 339. Stiner, M.C., Kuhn, S.L., Weiner, S., Bar Yosef, O., 1995. Different ial Burning, Recrystallization, and Fragmentation of Archaeological Bone. Journal of Archaeological Science. 22, 223 237. Stringer, C.B., Finlayson, J.C., Barton, R.N.E., Fernández Jalvo, Y., Cáceres, I., Sabin, R.C., Rhodes, E.J., Currant, A.P., Rodríguez Vidal, J., Giles Pacheco, F., 2008. Neanderthal exploitation of marine mammals in Gibraltar. Proceedings of the National Academy of Sciences. Tattersall, I., 1995. The Fossil Trail: How We Know what We Think We Know about Human Evolution. Oxford Universit y Press. Texier, P. J., Porraz, G., Parkington, J., Rigaud, J. P., Poggenpoel, C., Miller, C., Tribolo, C., Cartwright, C., Coudenneau, A., Klein, R., Steele, T., Verna, C., 2010. A Howiesons Poort tradition of engraving ostrich eggshell containers dated t o 60,000 years ago at Diepkloof Rock Shelter, South Africa. Proceedings of the National Academy of Sciences. 107, 6180 6185. Texier, P. J., Porraz, G., Parkington, J., Rigaud, J. P., Poggenpoel, C., Tribolo, C., 2013. The context, form and significance of the MSA engraved ostrich eggshell collection from Diepkloof Rock Shelter, Western Cape, South Africa. Journal of Archaeological Science, The Middle Stone Age at Diepkloof Rock Shelter, Western Cape, South Africa. 40, 3412 3431. Thompson, J.C., 2010. Taphon omic analysis of the Middle Stone Age faunal assemblage from Pinnacle Point Cave 13B, Western Cape, South Africa. Journal of Human Evolution. 59, 321 339. Thompson, J.C., Henshilwood, C.S., 2011. Taphonomic analysis of the Middle Stone Age larger mammal fa unal assemblage from Blombos Cave, southern Cape, South Africa. Journal of Human Evolution. 60, 746 767. Thompson, J.C., Henshilwood, C.S., 2014a. Tortoise taphonomy and tortoise butchery patterns at Blombos Cave, South Africa. Journal of Archaeological Sc ience. 41, 214 229. Thompson, J.C., Henshilwood, C.S., 2014b. Nutritional values of tortoises relative to ungulates from the Middle Stone Age levels at Blombos Cave, South Africa: Implications for foraging and social behaviour. Journal of Human Evolution. 67, 33 47. Tribolo, C., Mercier, N., Douville, E., Joron, J. L., Reyss, J. L., Rufer, D., Cantin, N., Lefrais, Y., Miller, C.E., Porraz, G., 2013. OSL and TL dating of the Middle Stone Age sequence at Diepkloof Rock Shelter (South Africa): a clarification. Journal of Archaeological Science. 40, 3401 3411.

PAGE 102

90 Trinkaus, E., 2012. Neandertals, early modern humans, and rodeo riders. Journal of Archaeological Science. 39, 3691 3693. Ubelaker, D.H., Owsley, D.W., 2003. Isotopic evidence for diet in the seventeenth c entury colonial Chesapeake. American Antiquity. 68, 129 139. Val, A., de la Peña, P., Wadley, L., 2016. Direct evidence for human exploitation of birds in the Middle Stone Age of South Africa: The example of Sibudu Cave, KwaZulu Natal. Journal of Human Evo lution. 99, 107 123. Villa, P., Mahieu, E., 1991. Breakage patterns of human long bones. Journal of Human Evolution. 21, 27 48. Villa, P., Soriano, S., Teyssandier, N., Wurz, S., 2010. The Howiesons Poort and MSA III at Klasies River main site, cave 1A. Jo urnal of Archaeological Science. 37, 630 655. Wadley, L., 2010. Were snares and traps used in the Middle Stone Age and does it matter? A review and a case study from Sibudu, South Africa ScienceDirect [WWW Document]. URL https://www.sciencedirect.com/sci ence/article/pii/S004724840900219X (accessed 3.30.18). Wadley, L., Williamson, B., Lombard, M., 2004. Ochre in hafting in Middle Stone Age southern Africa: a practical role. Antiquity. 78, 661 675. Hunter Gatherer Settlement Systems. American Antiquity. 47, 171 178. Will, M., Kandel, A.W., Conard, N.J., 2019. Midden or Molehill: The Role of Coastal Adaptations in Human Evolution and Dispersal. Journal of World Prehis tory. 1 40. Will, M., Kandel, A.W., Kyriacou, K., Conard, N.J., 2016. An evolutionary perspective on coastal adaptations by modern humans during the Middle Stone Age of Africa. Quaternary International, The African Quaternary: environments, ecology and hum ans Inaugural AFQUA conference. 404, 68 86. Will, M., Parkington, J.E., Kandel, A.W., Conard, N.J., 2013. Coastal adaptations and the Middle Stone Age lithic assemblages from Hoedjiespunt 1 in the Western Cape, South Africa. Journal of Human Evolution. 64, 518 537.

PAGE 103

91 A PPENDIX A : Number of Individual Specimens Tables 1 5: The N umb er of Individual Specimens (NISP) in each stratigraphic aggregate by size class. R ed indicates high survival elements , and blue low surviva l elements (foll owi ng L am et al., 1988) .

PAGE 104

92 Table 1: NISP from the OBS aggregate (glacial) . Table 2: NISP from the DHA aggregate (transitional).

PAGE 105

93 Table 3 : NISP from the DHA/DS aggregate (transitio nal). Table 4: NISP from the DS aggregate (interstadial).

PAGE 106

94 Table 5 : NISP from the DBS aggregate (glacial ).

PAGE 107

95 A ppendix B: Minimum Number of Elements T able 1 : M NE o f size class 1 and 2 Bovids .

PAGE 108

96 Table 2: MNE of size class 3 and 4 Bovids.

PAGE 109

97 APPENDIX C: Mi nimum Animal Units for Bovids Table 1: T he MAU f or size 1 and 2 Bovids . T he DHA/DS is excluded from this analysis, a s no taxonomic identifications of bovids were made in this aggregate.

PAGE 110

98 T able 2: the MAU of size class 3 and 4 bovids. T he DHA/DS was excluded from this an alysis, a s no taxonomic identifications of bovids were made in this aggregate.