The relationship between proximal long bone shape and activity among four hunter-gatherer populations.

(1)

Gatherer Populations. by

Julia Luba Meyers

Bachelor of Arts and Science, University of Guelph, 2014

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF ARTS

in the Department of Anthropology

 Julia Luba Meyers, 2017 University of Victoria

(2)

Supervisory Committee

The Relationship between Proximal Long Bone Shape and Activity among Four Hunter-Gatherer Populations.

by

Julia Luba Meyers

Bachelor of Arts and Science, University of Guelph, 2014

Supervisory Committee

Dr. Helen Kurki, (Department of Anthropology)

Supervisor

Dr. Lisa Gould, (Department of Anthropology)

Departmental Member

Dr. Lesley Harrington, (Department of Anthropology, University of Alberta)

Outside Member

Dr. Hugo Cardoso, (Department of Archaeology, Simon Fraser University)

(3)

Abstract

Supervisory Committee

Dr. Helen Kurki, (Department of Anthropology) Supervisor

Dr. Lisa Gould, (Department of Anthropology) Departmental Member

Dr. Lesley Harrington, (Department of Anthropology, University of Alberta) Outside Member

Dr. Hugo Cardoso, (Department of Archaeology, Simon Fraser University) External Member

There is an understanding among biological anthropologists that long bone epiphyseal shape is highly regulated by genetic and biomechanical factors. Conversely, long bone diaphyseal geometry and robusticity have been shown to respond to activity in life. The current study examined the assumption of epiphyseal consistency by exploring the relationship between a well established bony response to activity (Cross-Sectional Geometry) and shape change among the proximal humerus and femur. Long bone samples were taken from four hunter-gatherer populations: the Andaman Islanders, the Indian Knoll, Point Hope Alaskans, and the Sadlermiut. Shape was measured through landmark configurations placed on the proximal end of a total of 91 humeri and 84 femora. Cross-sectional Geometry measures (J) were taken from each specimen, as well. Principal Component Analyses were conducted on the landmark shape data to determine where the shape variation was occurring among the sample. These Principal Components were then compared via Bivariate Regression to the J values taken from the diaphysis.

Significant relationships occurred between the development of the lesser tubercle and an increase in J among the humerus sample. Significant relationships were also found among the femur sample; as when J increased the proximal epiphyses were more likely to be more gracile, and the space between the femoral head and the greater trochanter

(4)

increased. The humerus results indicated a more robust proximal epiphysis in groups with activities that rely heavily on the upper body, such as rowing, swimming, and

harpooning. The femur results were more complex, as the relationship between activity and proximal shape is likely heavily influenced by a genetically predetermined body shape. These results indicated that there is a relationship between activity and proximal epiphyseal shape, but that it, like all relationships, is complex, and comprised of multiple factors. Ultimately, proximal long bone shape is the result of multiple influences

including, activity, genetics, population adaptation, health, and many more. Future

research should focus on determining if the relationship between activity and shape exists among other populations, and when and where it begins during growth and development.

(5)

List of Tables

Table 3.1. Male and female humerus and femur used in the study………35 Table 3.2. Information on landmarks used in the proximal humerus landmark

configuration ……….40 Table 3.3. Information on landmarks used in the proximal femur landmark

configuration ……….42 Table 4.1. Average deviation (d) for each landmark between trial 1 and trial 2 in the humerus and femur samples ………..49

Table 4.2. Procrustes ANOVA Results from the Humerus and Femur Samples ...….50 Table 4.3. Eigenvalues, percent variance, and cumulative percentage of the PCA on the proximal humerus ……….52 Table 4.4. Principal Components that demonstrated a significant difference between the average male and female PC scores among the humerus sample ………53

Table 4.5. Analysis of variance among the shape Principal Components of the humerus sample. Significance indicated in bold ……….55

Table 4.6. Significant results from the Post-Hoc Tukey HSD analysis on the humerus sample………56

Table 4.7. Descriptive statistics of the Principal Components that demonstrated

significant clustering among proximal humerus sample ………..64 Table 4.8. Eigenvalues, percent variance, and cumulative percentage variance from the PCA on proximal femur ………66

Table 4.9. Analysis of variance among the shape Principle Components of the femur sample. Significance indicated in bold………..68

(8)

Table 4.10. Significant results from the Post-Hoc Tukey HSD analysis on the femur sample ………...69

Table 4.11. Descriptive statistics of the Principal Components that demonstrated significant differences between male and female mean scores in the femur sample …...71

Table 4.12. Descriptive statistics of the Principal Components that demonstrated significant clustering among proximal femur sample ………..76

Table 4.13. Descriptive statistics of the J values for the humerus samples …………..77 Table 4.14. J values from the humerus samples that demonstrated significant

dimorphism ………...79 Table 4.15. Descriptive statistics of the J values for the femur samples ………...81 Table 4.16. J values from the femur samples that demonstrated significant

dimorphism………....83 Table 4.17. Significant results from the regression analysis on the humerus

sample………85 Table 4.18. Significant results from the regression analysis on the femur sample …...90

(9)

List of Figures

Figure 2.1. Anatomy of the femur (OpenStax Anatomy and Physiology 2016) ...6 Figure 2.2. Right proximal humerus of individual IK140 from the Indian Knoll sample...7 Figure 2.3. Right proximal femur of individual IK140 from the Indian Knoll sample…...8 Figure 2.4. Cells comprising the basic multicellular unit (BMU) and processes

undertaken in bone repair (Raggatt and Partridge 2010: 25105)………...10 Figure 2.5. Regulation of osteogenesis by RANKL/RANK/PG gene expression (Clarke 2008: S133)...11 Figure 2.6. Activation of remodeling via fluid response to mechanical loading (Chen et al., 2010: 109)………12 Figure 2.7. Femur midshaft medullary section showing the areas where Imax and Imin are taken (Stock and Shaw 2007: 415)………16 Figure 3.1. Landmark configuration of the proximal humerus………..39 Figure 3.2. Landmark configuration of the proximal femur………..41 Figure 3.3. a) lollipop diagram, and b) wireframe diagram demonstrating shape change through landmark displacement among Drosophilia (fruit fly) wings (Klingenberg 2013: 19)………..46 Figure 4.1. Eigenvalues of each Principle Component created during the Principle

(10)

Figure 4.2. Humerus shape variation among the samples along Principal Component 2 plotted against Principal Component 4...54 Figure 4.3. Humerus shape variation among the samples along Principal Component 1 plotted against Principal Component 3…...54 Figure 4.4. Principal Components that demonstrated significant cluster among the

humerus samples………...57 Figure 4.5. Wireframe diagrams demonstrating areas of shape change occurring between the average Procrustes configuration (light blue) and PCs 1, 2, 3, 4, 5, and 6 (dark blue) in the humerus sample (superior view). Positive shape change is shown on the left and negative on the left …...……….61 Figure 4.6. Wireframe diagrams demonstrating areas of shape change occurring between the average Procrustes configuration (light blue) and PCs 8, 10, and 11 (dark blue) in the humerus sample (superior view). Positive shape change is shown on the left and negative on the right………...63 Figure 4.7. Eigenvalues of each Principle Component created during the Principle

Component Analysis conducted on the proximal femur………65 Figure 4.8. Femur shape variation among the samples along Principal Component 1 plotted against Principal Component 2...67 Figure 4.9. Principal Components that demonstrated significant cluster among the femur samples………...70

(11)

Figure 4.10. Wireframe diagram demonstrating the positive shape change between the average Procrustes Configuration (light blue) and PCs 1, 2, 3, 4, and 5 (dark blue) in the femur sample (posterior view). Positive shape change is shown on the left and negative shape change is shown on the right ….………..75 Figure 4.11. Distribution of J values (mm) across the humerus samples………..78 Figure 4.12. Distribution of 20%, 50%, and 80% J value measurements (mm) across the humerus samples………..80 Figure 4.13. Distribution of femoral J values (mm) across the samples………..82 Figure 4.14. Distribution of 20%, 50%, and 80% J value measurements (mm) among males and females among the femur samples………..84 Figure 4.15. Regression scatter plot of humerus sample PC4 scores against J at 20%, 50%, and 80% ………...86 Figure 4.16. Regression scatter plot of humerus sample PC7 scores against J at 20%, 50%, and 80% ………...87 Figure 4.17. Regression scatter plot of humerus sample PC15 scores against J at 20%, 50%, and 80% ………...88 Figure 4.18. Wireframe diagrams demonstrating shape change between the Procrustes coordinates average (light blue) and PC7 and 15 (dark blue) in the humerus sample (superior view). Positive shape change is shown on the left, and negative on the right…89 Figure 4.19. Regression scatter plot of femur sample PC1 scores against J at 20%, 50%, and 80% ………91

(12)

Figure 4.20. Regression scatter plot of PC9 against J at 50%, PC 12 against J at 50%, and PC23 against J at 20% ………...92 Figure 4.21. Wireframe diagrams demonstrating shape change between the Procrustes coordinates average (light blue) and PC 9, 12, and 23 (dark blue) in the femur sample (posterior view). Positive shape change is shown on the left, and negative on the right..94 Figure 5.1. Wireframe diagram demonstrating shape change between Procrustes target configuration (light blue) and PC 4 (dark blue) of the humerus (medial view)…………97 Figure 5.2. Wireframe diagram demonstrating shape change between the target

Procrustes coordinates (light blue) and PC 8, 10, and 11 (dark blue) in the humerus

sample (medial view)……….98 Figure 5.3. Wireframe diagram demonstrating shape change between the target

Procrustes coordinates (light blue) and PC 7, and 15 (dark blue) in the humerus sample (medial view)………..100 Figure 5.4. Wireframe diagram demonstrating shape change between the target

Procrustes coordinates (light blue) and PC 9 (dark blue) in the femur sample (lateral view)………102 Figure 5.5. Wireframe diagram demonstrating shape change between the target

Procrustes coordinates (light blue) and PC 12 (dark blue) in the femur sample (superior view)………103

(13)

Figure 5.6. Wireframe diagram demonstrating shape change between the target Procrustes coordinates (light blue) and PC 9 (dark blue) in the femur sample (lateral view)………104

(14)

Acknowledgments

I would like to acknowledge first and foremost my supervisor Dr. Helen Kurki. Your constant guidance and mentorship are what has made this thesis possible. From the expertise and knowledge you gave me during our field research, to the reassurance and assistance you provided me with during our (almost daily) meetings, you have been a rock in an uncertain environment. Thank you for taking a chance and accepting me as your graduate student, you will never know how much this opportunity has meant to me. I would also like to acknowledge the work of my committee members, Dr. Leslie

Harrington and Dr. Lisa Gould, whose edits and advise have been invaluable to this project. My research would not have been completed without the generosity of Dr. Jay Stock and Dr. Tom Davies, both of whom were kind enough to grant me access to some of their digital specimen data. I want to thank Dr. Janet Young at the Canadian Museum of History and the Inuit Heritage Trust, Dr. George Crothers at the William S. Webb Museum of Anthropology, Dr. Brian Richmond at the American Museum of Natural History, and Dr. Heather Bonney at the Natural History Museum for allowing me access to their collections.

I want to also acknowledge my family for their support. Lucas, Logan, Ari, Mom, and Dad, you were all so helpful and supportive when I decided to uproot my life and move halfway across the country. Your visits helped keep me motivated to complete me work.

I have no clue as to how I got so lucky as to have the group of friends that I had during this degree. Jenny, John, and Colton, I can’t explain how much you all mean to me.

(15)

Without you I would not have completed this research. I think we kept each other sane during this chaotic time and I am so grateful for your friendship. I know you will all go on to do amazing things, and I am so happy I got to know each of you at this time in your lives.

(16)

Dedication

This thesis is dedicated to my mother, Katherine, and my father, Myles. You both pushed me to take school and education seriously as child, and instilled in me the drive and work ethic I needed to complete my thesis. I doubt I would have ever discovered my love of learning if you both had not been there to encourage me during my early years.

I would also like to dedicate this thesis to Carlos. I would not have made it through this process without your love and support. You kept me on track when I felt burnt out, and talked everything out with me when I was stressed. You encourage and inspire me everyday with your strength and your tenacity.

(17)

Chapter 1: Introduction

Certain aspects of skeletal morphology have been accepted as having a plastic or variable response to the conditions of the environment in which individuals live their lives. For example, long bone diaphyseal robusticity (measured via cross-sectional geometry), has been strongly associated with certain types of repetitive, habitual activities that individuals or whole populations engage in over the course of a life time; activities such as repetitive throwing which influences humerus shape, or habitual long-distance walking which influences femur shape (Rhodes and Churchill 2009, Stock and Pfeiffer 2001). In contrast, the epiphyses of long bones are considered to be highly genetically regulated, and thus less likely to respond plastically to activity. This study uses Geometric Morphometric (GM) approaches to analyze epiphyseal shape

characteristics to test the assumption that bone loading as a result of habitual activity has no effect on the proximal epiphyseal regions of the humerus and femur. The GM

approach to measuring shape change relies on removing all non-shape data (i.e. size, position and orientation) from a series of comparable structures so that any variation remaining is a matter of shape difference. Studies utilizing GM have been used through biology and biological anthropology to measure shape change in everything from

Drosophila wing shape to sexual dimorphism in the shape of gorilla scapulae

(Klingenberg 2013, Slice 2007). In this study, proximal shape change was examined in the humerus and femur using archaeologically derived skeletal samples representing four forager populations: Andaman Islanders from South East Asia, Point Hope groups from Alaska, Sadlermiut from the Canadian Arctic, and Indian Knoll from Kentucky. These

(18)

populations have all been studied thoroughly, and archaeological, historical, and/or ethnographic information is available for each (Stock and Pfeiffer 2001, Auerbach 2008, Collins 1956, Ryan and Young 2013, Holland 2014, Cowgill 2010, Nealis and Seeman 2012). The contextual data on these populations allows for a more thorough

understanding of the likely habitual loading patterns that were occurring among each group.

The purpose of this study was to determine the extent to which the shape of the proximal epiphyses of femur and humerus act as valid indicator of habitual activity in four hunter-gatherer populations. Characteristics of shape, determined via GM, were compared to cross-sectional geometric (CSG) properties of the diaphysis, which are well established as indicators of activity (Stock and Pfeiffer 2001, Shaw and Stock 2007, Davies and Stock 2014). Given their distinctive activity patterns, this study also

examined whether each sample exhibited distinctive proximal shape morphology of the femur and/or humerus. Overall, this study investigated the utility of proximal femur and

humerus shape as an indicator of activity and lifestyle in forager populations.

The main research questions concerned the relationship between proximal long bone shape and activity, and looked to determine whether the proximal epiphyses of the femur and humerus are plastic in their response to different activities and environments:

1. Do the shape characteristics of the proximal ends of the humerus and femur of each sample cluster distinctively among the samples?

(19)

If this were the case it would suggest that epiphyseal shape is significantly different among the sample groups. The next question investigated the link between activity and proximal shape:

2. Is there a relationship between cross-sectional geometry (midshaft J) and shape characteristics of the proximal ends of the femur and/or humerus?

If relationships were detected, it suggests that the shape of the proximal epiphyses are also influenced by the loading regime of the bone, meaning the region is more

biologically plastic than previously thought.

Previous research has shown that among populations with varying activity profiles, diaphyseal shape correlates with activity (Stock and Pfeiffer 2001). It is therefore expected that the different activity profiles represented among the study samples will result in the detection of differences among the groups in both diaphyseal and epiphyseal shape of the femur and humerus.

Chapter 2 discusses some of the research that has already been undertaken within biological anthropology which pertains to this project. It includes a review of how bone is formed and remodeled through life, femoral and humeral anatomy, activity, as well as research that has pushed for a connection between proximal epiphyseal shape and activity. Chapter 3 begins with some contextual information on each of the samples included in the project. Archaeological, historical or ethnographic information that allows for an understanding of the type of activity each group was involved in is reviewed. Chapter 3 also discusses the methods used to obtain and analyze the data for this project. This includes the creation of the 3-D computer models from which the shape and CSG

(20)

measures were taken, the data processing that occurred before the measures could be taken, the measurements themselves, as well as any type of analysis performed on the data. Chapter 4 presents the data acquired from the PCA shape analyses, the CSG measures, and the ANOVAs, T-tests, and bivariate regressions performed on the data to answer each research question. Chapter 5 discusses these findings as they relate to the initial hypotheses, along with some of the factors that may explain the results obtained and any limitations in the study. Finally, Chapter 6 concludes by reflecting over the significance of the results in the context of biological anthropology, and by pointing out any potential areas of future research on this topic.

(21)

Chapter 2: Literature Review

2.1. Morphology of the Humerus and Femur

As this project deals with the femur and humerus, their anatomy is of primary interest. Long bones (Figure 2.1) are comprised of three parts; the hollow shaft or diaphysis, the flared metaphysis between the growth plate and the bone shaft, and the epiphyses on either end of the bone (Clarke 2008). Each section of the bone is made up of trabecular bone and cortical bone. Trabecular bone (or spongy bone) is spongy and lightweight, and is found primarily in the ends of the long bones (White and Folkens 2005). Cortical bone (or compact bone) is solid, comprises most of the diaphysis, but is found outside the marrow, and is less metabolically active than trabecular bone (Clarke 2008).

(22)

Figure 2.1. Anatomy of the femur (OpenStax Anatomy and Physiology 2016). Used with permission from OpenStax.

The shoulder girdle is comprised of three bones, the humerus, the scapula, and the clavicle. The shoulder girdle itself is connected to the axial body via the articulation of the medial clavicle and the lateral manubrium (White and Folkens 2000). The proximal humerus (Figure 2.2) is the site of many points of muscle and ligament attachment, but hosts only one articulation; that of the glenohumeral joint, which it creates with the glenoid of the scapula (White and Folkens 2000). In total 16 muscles affect the movement of the shoulder, which due to its minimal bony articulations and a ball and socket joint is free to rotate in any direction (White and Folkens 2000). The shoulder is also capable of abduction, adduction, flexion, and extension (Bergman et al., 2011). The

(23)

superior glenohumeral ligament attaches along the anterior anatomical neck and the long

head of the biceps tendons runs directly down through the bicipital groove (or

intertubercular sulcus) (Arai et al., 2010). The lesser tubercle is the site of insertion of the

subscapularis and the teres major muscles, while the greater tubercle hosts insertion sites

for the suraspinatus, infraspinatus, and teres minor muscles; all of which are major rotation muscles (White and Folkens 2000).

Figure 2.2. Right proximal humerus of individual IK140 from the Indian Knoll sample.

The hip joint is comprised of two bones; the femur and the os coxa. The os coxa is part of the axial body, connecting to the vertebral column through the sacrum (White and Folkens 2000). The hip joint is made moveable by multiple muscle groups, the main ones being the gluteal and hamstring muscles (White and Folkens 2000). Like the shoulder, the hip is able to perform rotation, adduction, abduction, flexion and extension, but generally has a smaller range of rotational motion due to restrictions caused by the shape of the joint surface (Figure 2.3) (White and Folkens 2000). The femoral head (via the fovea capitis) acts as the insertion site for the ligamentum teres (White and Folkens

(24)

2000). The greater trochanter is the site of insertion of the gluteus minimus and gluteus

medius, both crucial for stable bipedal locomotion (White and Folkens 2000). The lesser

trochanter is the insertion point for the iliopsoas tendon, the iliacus muscle, and the psoas

major muscle; all major flexors (White and Folkens 2000).

Figure 2.3. Right proximal femur of individual IK140 from the Indian Knoll sample. 2.2. Bone Formation and Remodeling

Bone is formed through two main processes; intramembranous and endochondral bone formation. Intramembranous formation involves bone ossifying via deposition of bone tissue on an embryonic connective tissue membrane (White and Folkens 2005). Endochondral formation involves the formation of a cartilage model that is gradually replaced with bone (Mackie et al., 2008). Flat bones, like those of the cranium, are formed by intramembranous processes, while long bones, like the femur and the humerus are formed via a mixture of endochondral and membranous bone formation (Clarke 2008). Both processes result in a mineralized connective tissue (i.e. bone), which contains

(25)

4 basic cells; osteoclasts, osteoblasts, bone lining cells and osteocytes (Florencio-Silva et al., 2015). Osteoclasts are terminally differentiated myeloid cells that act to remove mineralized bone tissue (Raggatt and Partridge 2010). Osteoblast cells deposit bone tissue, make up 4-6% of the total resident bone cells, and lie along the bone’s surface (Florencio-Silva et al., 2015). Bone lining cells are a form of flattened osteoblasts, which line the surface of the bone; their functions are not entirely clear but they appear to play a part in regulating osteogenesis and bone resorption (Florencio-Silva et al., 2015).

Osteocytes are created during bone formation by groups of terminally differentiated osteoblasts after the surrounding osteoid mineralizes and traps them (Raggatt and Partridge 2010). Though bone tissue is mineralized through these processes it does not stay static after growth and development end, and instead remodels over the course of an individual’s life.

Bone remodeling occurs via the same processes that occur during initial bone formation. This remodeling is accomplished by a structure called a basic multicellular unit (or BMU), which contains osteoclasts and osteoblasts that are responsible for removing old bone and depositing new tissue (Florencio-Silva et al., 2015). The BMU interacts with different bone components, like osteocytes, within the bony matrix and lining cells which cover the surface of the bone to allow for the deposition new bone tissue (Sims and Gooi 2008) (Figure 2.4).

(26)

Figure 2.4. Cells comprising the basic multicellular unit (BMU) and processes undertaken in bone repair (Raggatt and Partridge 2010: 25105). Used with permission from The American Society for Biochemistry and Molecular Biology.

The basic multicellular unit forms temporarily in bone cavities during remodeling processes, with bone-cutting osteoclasts at the front cone of the structure, and bone forming osteoblasts comprising the closing cone at the rear (Florencio-Silva et al., 2015). The BMU is most likely covered in a group of cells making up a bone remodeling

compartment, through which it can attach to bone lining cells on the bone surface and communicate with cells in the bone matrix (Florencio-Silva et al., 2015).

This process of bone resorption and renewal is highly physiologically controlled and occurs through RANKL/RANK/OPG gene expression (Robling et al., 2006). RANKL (receptor activator of NF-κB ligand) is a biological signal expressed by osteoblasts, osteocytes and stromal cells (which express growth differentiating factors), and attaches to the RANK receptor on an osteoclast precursor causing it to become an osteoclast (Figure 2.5); the final stage in osteogenesis (Florencio-Silva et al., 2015). CSF-1 (colony-stimulating factor) is another biological factor that acts in a similar role as RANKL in osteogenesis (Raggatt and Partridge 2010). Alternatively, the factors OPG,

(27)

NO, and TGFβ act to inhibit osteogenesis but binding to the RANK receptor on the osteoclast precursor thereby preventing the RANKL factor from binding (Niedźwiedzki and Filipowska 2015).

Figure 2.5. Regulation of osteogenesis by RANKL/RANK/PG gene expression (Clarke 2008: S133). Used with permission from the American Society of Nephrology.

Osteocytes are embedded within osteons, arranged circularly around a Haversian canal, and connected with other osteocytes via processes that extend through the

canaliculus (Figure 2.6) (Clarke 2008). Cellular remodeling in response to activity is initiated through a hydrostatic pressure put on osteocytes by fluid flowing through the canalicular space (Chen et al., 2010). This fluid flow is created when mechanical loads are placed on bone (Chen et al 2010).

(28)

Figure 2.6. Activation of remodeling via fluid response to mechanical loading (Chen et al., 2010: 109). Used with permission from Elsevier.

Though bone remodeling as described above occurs as a regular part of bone physiology there are several instances that can increase or decrease the rate of deposition or resorption. Fracture repair can increase the rate of remodeling as more osteoblast and osteoclast precursors are brought to the site of injury through newly created blood vessels (Stegen et al., 2015). Changes in mechanical bone loading and circulating levels of calcium can increase or decrease both the rate of remodeling and the amount of bone deposited (Sims and Gooi 2008). With age remodeling increases, and as less bone is deposited by osteoblasts and more is absorbed by osteoclasts the cortical bone tends to decrease in mass and become more porous (Clarke 2008). High levels of lifelong activity

(29)

can also act to influence the distribution of cortical bone around long bone diaphysis (Gosman et al., 2011).

2.3. What is Activity?

Since the primary questions asked in the current study were about the effects of activity on bone, it is imperative to have a common understanding of what activity is and what the term encompasses. Activity is a conceptual term, and like many of the other conceptual terms used in the study of humans (e.g. health, aging, sickness) a concrete definition is often difficult to come across in the literature. Activity is understood in several ways across different disciplines.

The clinical health research concerning physical activity among large populations tends to focus on activity in terms of its cardiovascular effects, with the types of activities being divided into aerobic and anaerobic (Carlson et al., 2015). In this context, physical activity is any action that elevates an individual’s heart rate above resting rates (Sparling et al., 2015). Within clinical health research activity can thus be defined as actions with raise the heart rate over an extended time span. This definition, while useful for living populations, does not work for archaeological populations.

In addition to the cardiovascular and respiratory effects, physical activity in the context of osteology and sports science can be understood through the effects it has on bone (Maughan et al. 1984, Rubin et al., 1985, Taaffe et al., 1997, Tenforde and Fredericson 2011). For the purposes of the current project activity was defined as any repetitive action that causes mechanical loading and/or strain that may influence bone formation and composition. Therefore, when the term activity is employed within the

(30)

context of this research it refers to physical activity undertaken habitually and over a long period of time, including during growth and development. There are a multitude of different types of activities that impact the body’s bones and muscles differently, but broad categories can be created to describe the biomechanical effect of the mode of activity on the body: low, medium and high intensity activity.

2.4. Activity in Biological Anthropology

2.4.1. Early and Contemporary Bone Theory

The relationships among muscle activity, bone loading, and bone remodeling were first postulated by the German anatomist Julius Wolff in 1892. Wolff’s Law of Bone Transformation states that wherever bone is needed it is deposited and wherever it is not, it will be taken away. This means that bone tissue is placed where there is

functional demand for it due to mechanical force (Larson 1997). Wolff’s Law was foundational in the study of bone remodeling, but since the time of its discovery many have pointed out how the situation is much more complicated (e.g., Stock and Pfeiffer 2001, Pearson and Lieberman 2004, Ruff et al., 2006, Harrington 2010, Carlson and Marchi 2014, among others). The effects of genes, activity, environment and a multitude of other biocultural factors influence bone structure and shape. These additional factors allowed researchers to reconstruct information about bone loading, or activity levels, from bone shape (i.e., bone functional adaptation) (Ruff et al., 2006, Gosman et al., 2011). In 1963 Frost began describing the mechanisms by which bone growth occurs, modeling and remodeling, and how bone morphology is affected by factors such as body mass and bone loading. Since this initial description of modeling during early childhood

(31)

and adolescence and remodeling in adulthood there have been many researchers who have investigated how, where and why these processes of formation occur.

Frost put forth his own theory as to how activity affects bone morphology with his Mechanostat Theory (1987). Mechanostat Theory states that if certain thresholds of non-traumatic strain (i.e. microdamage to bone) are reached bone remodeling will increase bone mass (Frost 2003). This remodeling can result in a number of changes to the bone, including changes to diaphyseal shape, and musculoskeletal stress markers (MSMs), which are small tubercles of bone formed at sites of muscle and ligament attachments (Niinimäki 2012). Since MSMs occur at the point of muscle and ligament attachments, and many muscle and ligament attachment points are at the epiphyseal regions of long bones, they contribute to a change in epiphyseal bone shape.

2.4.2. Cross-Sectional Geometry

Cross-sectional robusticity represents the distribution of cortical bone around a central (or neutral) axis (Figure 2.7). In long bones, robusticity refers to the external size or circumference of the diaphysis relative to the length of the bone (Ruff et al., 1993). The distribution of bone around and away from the central axis reflect loading history since it is the bone furthest from this point that resists loading more than that which is closest to the central axis (Stock and Shaw 2007). Since bone formation and remodeling respond to mechanical loading the size and shape of the diaphysis can be used to examine habitual activity within populations (Macintosh et al., 2013). Anthropologists use

biomechanics (the applications of engineering principles to biological materials), and components of beam theory to examine torsion and bending properties within diaphyseal patterns (Ruff and Hayes 1983, Larson 1997, Stock and Shaw 2007).

(32)

Figure 2.7. Femur midshaft medullary section showing the areas where Imax and Imin are taken (Stock and Shaw 2007: 415). Used with permission from John Wiley and Sons.

There are five primary loading forces that affect bone, tension: wherein equal and opposite forces are applied outwardly from the surface of a slice of bone; compression: where equal and opposite forces are directed toward the surface; and shear force: which is applied parallel to the bone slice (Larson 1997). Bending forces are produced by tension being applied to one side of the bone making the side convex, and by compression being applied to the other side, making it concave (Larson 1997). Torsional loads on the bone are caused by tension, compression and shear force on the bone and result in a twisting of the bone around a central axis (Larson 1997). The further the bone tissue is distributed from the central axis in the cross section the more the whole long bone is able to resist both bending and torsional forces that act upon it (Larson 1997). The cross-sectional area and the distribution of bone around the central axis reflects any mechanical behaviour the bone was involved in (Larson 1997). In order to quantify these forces anthropologists can take measures of total subperiosteal area (TA), maximum second moment of area (Imax), minimum second moment of area (Imin), and polar second moment of area (J) from bone cross-sections (Figure 2.7). Mechanically speaking TA represents the external measure of

(33)

the combined cortical bone and medullary cavity (Stock and Shaw 2007). The maximum and minimum bending strengths, Imax and Imin, represent the maximum and minimum distances of bone distribution from the central axis (Lieberman et al., 2004). Torsional strength is represented by J and is equal to Imax + Imin (Macintosh et al., 2013).

Traditional calculations of these CSG properties have taken into account both the internal and external contours of the diaphysis, however recently researchers have begun to debate whether the internal contours are necessary for these calculations, as it is more difficult to obtain endosteal and periosteal measures. X-rays or CT scans are needed to acquire the internal dimensions and can be difficult to obtain. Davies and colleagues (2012) tested the use of periosteal-only CSG measures against periosteal plus endosteal CSG measures to determine if it was necessary to include endosteal measures in studies of robusticity. The perceived benefit of using both periosteal and endosteal measures was that endosteal measures tend to remodel more quickly than periosteal and were thought to reflect biomechanical loading better than periosteal alone (Szulc et al., 2005, Stock and Shaw 2007). The results of Davies and colleagues (2012) study indicates that both

periosteal and endosteal, and periosteal-only methodologies will give similar measures of CSG, meaning that periosteal measures alone can be used. This is significant for multiple reasons including, reduction in the cost and time associated with gaining robusticity measures, as well as the implications for virtual 3-D bone models. While Sparacello and Pearson (2010) maintain that the use of the internal contours is necessary in skeletal samples with older members wherein some bones may be subject to degradation and resorption, periosteal-only measures have proven useful on specimens that are neither extremely juvenile nor aged (Davies et al 2012). Macintosh and colleagues (2013) found

(34)

similar results in their study of “true” v. periosteal only measurement. They took cross-sectional geometric measures from several places along the humeral, femoral and tibial diaphyses and found that the CSG properties (TA, I and J) of the periosteal only measures were comparable to the “true” diaphyseal measures. The standardized biometric positions along the proximal-distal axis (z-axis) of the diaphysis from which CSG measures are taken are based on the type of long bone being examined. Cross sections are taken at the same location on every long bone and are only compared amongst each other at standard locations. Locations commonly used in measuring CSG are, 20%, 35%, 50%, 65%, and 80% of the diaphysis, beginning at the distal end of the long bone (Ruff 2002, Macintosh et al., 2013).

Stock and Pfeiffer (2001) examined the relationship between bone remodeling and activity by comparing the CSG measures of robusticity of the long bones of Later Stone Age South African and Andaman Islanders foraging populations, who were involved in different habitual activity, but were of a similar body size. The purpose of the study was to determine the CSG measures taken from their respective upper limbs (humeri,

clavicles) and lower limbs (femora, tibiae and first metatarsals) matched the types of habitual behaviour that occurred in each population. The researchers found that the measurements of robusticity they observed followed patterns predicted by the known types of behaviour among each group. Stock and Pfeiffer used measures of total

subperiosteal area of the section (TA) to derive Imax, Imin, and J. From these measures, the researchers were able to determine the distribution of bone around central x- and y- axes. The CGS measures were taken at different intervals (20%, 35%, 50%, 65%, 75%, and

(35)

80%) along the z-axis (parallel to the diaphysis) in order to gain a more holistic knowledge of bone distribution along the long bone shafts.

Shaw and Stock used this method to examine the differences in cross-sectional patterning between early Homo, mid-Holocene hunter gathers, and modern athletes in the ulna and the humerus (2009), and the humerus and tibia (2013). In the earlier study, they were looking to determine how the different loading patterns created by playing cricket and swimming manifested in diaphyseal robusticity. They found that there was a higher degree of asymmetry in robusticity measures among cricket players, who traditionally have a dominant playing arm, while there was little to no asymmetry found among the swimmers. They also found that the effects of this type of loading were more evident in the proximal limb segments rather than the distal segments. In their 2013 study, Stock and Shaw looked to further understand activity patterns among Pleistocene hominins, by comparing their cross-sectional robusticity with mid-Holocene hunter gatherers, and modern long-distance runners and swimmers, the activity patterns of which are well known. They found that the tibial loading patterns of the Pleistocene hominins were closest to the terrestrially mobile Later Stone Age South African foragers and modern endurance runners. These results indicate that Homo sapiens and Neanderthals from that time were highly mobile.

Sládek and colleagues (2015) used 3-D computer long bone models to gain robusticity measures of the lower limb (femora and tibiae) from two Mid-Holocene hunter-gatherer groups from the Cis-Baikal, Siberia in an effort to understand changes in diet observed through isotopes. The virtual models were just as useful as the skeletal remains themselves thanks to the utility of periosteal measures in determining robusticity.

(36)

These studies indicate that one can, with a high level of confidence, use 3-D periosteal models of bones to gain CSG measures of robusticity, and to detect and predict activity within samples. Therefore, in this project, only the periosteal surfaces were used to obtain the measures of robusticity.

2.4.3. Recent Research on Shape, Robusticity and Activity

Research on activity has generally fallen under two, sometimes overlapping, domains: athlete-based studies, and skeletal-based. The difference comes down to those being studied, with skeletal-based research focusing on the remains of past populations of

Homo sapiens, and sometimes other hominins and hominoids, and athlete-based on living

humans.

The earliest studies done on athletes compared the limb bones of trained and untrained individuals, and the differences in physiology between the two. Jones and Colleagues (1977) examined the cortical bone distribution in the dominant arms and non-dominant arms of professional tennis players and found significant growth of osteon cells and widening of the medullary cavity in response to the constant mechanical loading in the dominant arm.

More recent studies on athletes have focuses on the effects of different types of habitual loading in different sports on bone morphology. A study by Nikander and colleagues (2010) on female athletes demonstrated that activities involving ground impacts such as endurance running, ball games, or jumping increased cortical bone robusticity in the tibial diaphysis at a higher rate than high-magnitude (powerlifting) or non-impact (swimming) activities. Another study by Hind and colleagues (2012)

(37)

examined cross-sectional robusticity of the femur in male swimmers, runners, gymnasts and non-athletes. The researchers found that runners and gymnasts had increased cross-sectional areas (resistance to axial loads), and higher bone mineral densities than the swimmers and non-athletes.

Studies that utilize non-living hominins and hominoids have used similar CSG methods to analyze activity (see Section 2.4.2.), and have found diaphyseal robusticity to be strongly associated with detection of habitual activity. Though robusticity has a well-recognized relationship with activity many researchers believe the epiphyseal portions of long bones are constant and highly regulated by genetic and biomechanical limitations. Ruff and colleagues (1993) observed a general decrease in recent H. sapiens post-cranial robusticity from archaic H. sapiens, which they attributed to less active recent

populations, but saw no significant change to proximal end size between the two groups. There may well be no significant changes to size in these areas, but what about shape? During that same year Trinkaus (1993) was investigating the relationship between neck-shaft angle and activities among Early Modern Humans and Neanderthals, and found that evidence that loading patterns during development created plastic variations in the

femoral neck. Since that time there has been a push towards a relationship between activity and bone shape in the literature. Harmon (2007) linked variation in neck-shaft angle with different activity levels among australopithecines and different types of bipedalism. Frelat and colleagues (2012), in their recent examination of hominoid tibia epiphyseal shape using GM, suggested that explorations of shape would be beneficial in distinguishing differences among humans based on habitual activity. Other researchers who have focused on hominoid long bone shape have suggested a relationship between

(38)

bone bowing and activity type or level among each species (Holliday and Friedl 2013). Studies on more recent hominin groups have also alluded to a relationship between activity and bone shape. Rhodes and Churchill (2009) examined humeral torsion and shape in Middle and Upper Paleolithic populations that practiced spear throwing. They, like others, suggested that population-wide habitual activity could be related to humeral asymmetry and overall humerus shape. There is also evidence that mechanical strain can influence the growth of tubercles and pits on the epiphyses at points of muscle attachment and lead to the creation of MSMs (Hirschberg 2005, Niinimäki 2012). The current study examined the relationship between proximal epiphyseal shape and diaphyseal shape, the latter being known to reflect loading regimes, to determine whether epiphyseal shape is also influenced by bone loading and activity. Since habitual loading patterns are known to differ among different human populations, and since bone responds to increased

loading by laying down more bone and decreased loading by resorbing bone, there should be differences among peoples that practiced different habitual activity. This connection is already well established within the bone structure of the diaphysis, but the connection between epiphyseal shape and habitual loading is less well understood.

2.4.4. Geometric Morphometrics

Geometric Morphometrics (GM) is a relatively new outgrowth of Morphometrics, which is a field of study that uses multivariate statistical analysis to address questions about changes and variation in shape (Slice 2007). It has been used frequently throughout biological and anthropological analysis of morphology. Geometric Morphometrics involves the use of morphometric methodologies on a Cartesian coordinate system of landmarks, which allows for all the landmark data (both size and shape information) to be

(39)

retained throughout any type of shape analysis and changes in 3D space (Slice 2007). This process is superior to traditional shape measuring techniques that utilize multiple linear measurements because the shape data is not affected by bone size, position or orientation. GM has become more and more common in the past few years with the increased availability of advanced computing and 3D scanning technologies. Biological anthropologists have used it to examine everything from femoral curvature in

Neanderthal and modern humans, to cranial shape difference in modern male humans born with fetal alcohol syndrome (De Groote 2011, Slice 2007).

Geometric Morphometrics analyses rely on landmarks to represent shape structure in 3D space. Landmarks are homologous points shared by each specimen in the analysis. There are several types of landmarks that can be used when making configurations each with their own significance: Type I, II, and III. Type I landmarks mark areas that are biologically significant (e.g. muscle insertion points like the fovea capitus), Type II landmarks denote areas that are both biologically and geometrically significant (e.g. the most superior point on the anterior point of the greater trochanter, which is both a muscle insertion point and a geometrically significant point), and Type III landmarks indicate an area is geometrically significant (e.g. the most proximal point on the bone) (Klingenburg 2013). Landmarks are positioned based on the aspects of shape variation being captured. In their 2014 study of the proximal humerus, Arias-Martorell and colleagues lay out two distinct landmark configurations; one designed to capture shape differences based on growth and developmental, and one to capture shape differences based on functional constraints. They determined there was more shape change associated with their functional landmark configuration than the developmental one. Thus, landmark

(40)

configurations must be in sync with the type of shape change the researcher is looking for.

After landmarks are determined, a General Procrustes Analyses (GPA) must be undertaken in order to ensure that only differences in shape are examined. This method was first described by Sneath in 1967, but has been around long enough (in different forms) to have been theorized by Boas (Slice 2007). General Procrustes Analysis involves the removal of all non-shape related variation via the creation of a centroid (an overall size component of the specimen) and the continual “fitting” of each landmark configuration to said centroid, which will remove all information from the comparisons of the specimen (Klingenberg 2013). Through GPA all models are realigned so as to be in the same orientation and position (Slice 2007). The process of removing differences between size, position, and orientation ensures that all landmark configurations of all specimens are brought onto a common coordinate system, where any differences in coordinate values is the result of differences in shape among the sample. After the GPA is conducted the landmark configurations are now called Procrustes coordinates

(Klingenburg 2013). These coordinates can then be subjected to additional analysis. Shape analyses on these new Procrustes coordinates typically takes the form of a Principle Component Analysis (PCA) or a Canonical Variates Analysis (CVA), though other methods of analysis can be conducted as well. Both methods are used to produce ordered data which simplify descriptions of shape data change, and both can be used as tools for exploratory data analysis (Zelditch et al., 2004). PCA creates a new set of ordinations from the original Procrustes coordinates and then scores each individual in the sample as to where their shape falls on these new ordinations (Zelditch et al., 2004).

(41)

PCA is not a tool that can test hypotheses statistically, but is instead used to observe patterns in shape variation that may occur in the data.

(42)

Chapter 3: Materials and Methods

3.1. Study Samples

The samples used in this research are representative of forager populations living across various places in space and time. They share an anthropological classification of “hunter-gatherer” groups. This term is generally given to groups which have existed and subsisted in small, egalitarian, non-state controlled societies (Larson and Daly 1999). Foragers, in general, have also managed to sustain their lifestyles for prolonged periods of time and with little large-scale environmental degradation when compared to

agricultural and industrial societies (Larson and Daly 1999). However, how individual groups live as foragers is highly variable.

In this project, femora and humeri from four different hunter-gather groups were examined; the Andaman Islanders of South Western Asia, Alaskan populations from the Point Hope peninsula, Indian Knoll from the Green Valley region of Kentucky, and the Sadlermiut from Native Point in the Eastern Canadian Arctic. The samples examined took the form of 3-D virtual models. These models were constructed by taking surface scans of each bone using the Konica Virtuoso and Next Engine laser scanners. The scans were then compiled into computer models using GeoMagic Design X and ScanStudio Pro software, respectively. The scans taken of the Sadlermiut, Point Hope and Indian Knoll samples were collected as part of a larger research project of Dr. Helen Kurki and Dr. Lesley Harrington, and the scans from the Andaman sample were taken by Dr. Jay Stock and Dr. Tom Davies, of Cambridge University, Department of Anthropology and

(43)

3.1.1. Andaman Islanders

The Andaman Islands in the Eastern Indian Ocean are home to several indigenous populations, mainly the Great Andamanese, the Onge, and the Jarawa, who migrated from the South Asian mainland roughly 50,000 to 70,000 years ago, most likely dispersing among the archipelago as Homo sapiens were populating Asia for the first time (Thangaraj et al., 2005). These populations were extremely genetically isolated for most of their occupation of the islands (Flower 1880). Andaman Islanders are generally described as “negritos”; a suite of phenotype characteristics common among populations in South Asia, and comprising small body size, dark skin and tightly curled hair (Stock 2013). The Asian “negrito” phenotype most likely developed separately from similar groups in Africa, rather than being an ancestral trait shared between the populations (Stock and Migliano 2009). The islands themselves have a long history of British

colonialism beginning in 1858 and ending when the islands gained independence in 1947 (Stock and Migliano 2009). Research on their cultural and morphological traits began with British colonial rule, which brought a variety of ethnographers, archaeologists, naturalists, and other colonial agents to the islands (Flower 1880). The dietary change and food shortages caused by the colonial period created a dramatic increase in mortality among the indigenous populations and a generational change in body size among groups who were in contact with the British (Stock & Migliano 2009). The sample used in this study contains individuals who died between 1860 and 1900 A.D. (Kurki 2013), and thus body size may be affected by the strains put on the population during the colonial period. The collection is held at the Museum of Natural History in London (Stock and Pfeiffer 2001).

(44)

The Andaman Islanders are often described as a seafaring people. They exploited offshore resources either by swimming or by using watercraft to hunt (Stock and Pfeiffer 2001). To a lesser extent, they also exploited land-based resources, like roots, fruit, honey and wild boars using bow (Radcliffe-Brown 1964). Canoes were the main mode of transportation utilized, both for hunting and for visiting neighbouring people (Radcliffe-Brown 1964). Hunting from canoes took the form of harpooning, while small nets were used to catch small fish and prawns (Radcliffe Brown 1964). The bow and arrow were also important in both marine and land-based hunting, with spears being used less

frequently (Cipriani 1966). As with the Sadlermiut and the sample from Native Point, the Andaman’s watercraft-based hunting methods would put significant mechanical loads on the shoulder joint, as well as torsional loading on the diaphysis of the humerus. An interesting point of comparison between the study samples that were marine reliant is that of swimming. Unlike the other groups, the Andaman Islanders would hunt by swimming in shallow waters. Stock and Shaws’ (2009) study of robusticity studies conducted on the upper limbs of swimmers demonstrate little to no asymmetry in diaphyseal robusticity in the humerus since mechanical loading involved in the activity is bilateral, but as of yet there is no indication of how joint shape will be affected, if at all. Swimming has also been shown to produce less mechanical loads on the lower limbs when compared to other activities, such as habitual long distance running (Stock and Shaw 2013). Robusticity comparisons have been made between the Andaman Islanders and populations that more closely resemble the Indian Knoll, semi-sedentary foragers. Stock and Pfeiffer (2001) compared the cross-sectional geometry of Andaman Islander long bones to those from a highly mobile, but terrestrial Later Stone Age South African sample to show the

(45)

relationship between upper and lower limb robusticity and habitual activity. They found that the Andaman exhibited stronger humeri and clavicles, but less robust lower limbs when compared to the LSA foragers.

3.1.2. Indian Knoll

Of the several archaeological forager populations originating within the Green River Archaic Period, the material from the site of Indian Knoll is the largest, containing 3 meters of accumulated deposits, 55,000 recovered artifacts and over 1000 human burials (Nealis and Seeman 2015). A shell-midden dates the Indian Knoll hunter-gather occupation of the Green River valley in Kentucky to roughly 6,415 to 4,143 years BP (Cowgill 2010). The Indian Knoll habitat consisted of relatively moderate uplands, occupied seasonally, and lowlands, which were characterized by streams and wetlands (Moore and Thomspon 2012). The area around the Green River Valley constituted a stable mid-Holocene riverine environment, wherein shellfish exploitation was on the rise among Indian Knoll populations (Nealis and Seeman 2015).

The site itself is dated to the pre-agricultural and pre-pottery Eastern Archaic period (Webb 1946). Based on the burials recovered from the Green River Valley shell midden mound, the Indian Knoll population was fairly large, and relied on essential resources such as deer, turkey, mussels, nuts and locally-sourced plant materials (Cowgill 2010). Mussels remained a steady fresh-water food source for Indian Knoll populations, even as technologies changed. Early population dental wear patterns demonstrate the use of teeth to pry open mussels, while later groups show a marked decrease in this type of tooth wear as new food-processing technologies were adapted (Nealis and Seeman 2015). Several types of projective points have been unearthed in the Green River Valley, as well

(46)

as a number of food processing tools such as stone mortars and pestles (Moore and Thompson 2012), suggesting that bow and arrow, and spears were probably used frequently.

This sample will be interesting when compared with the other study samples, since the Indian Knoll sample is arguably the most terrestrially active population, which may influence their femur shape. Auerbach (2008) examined robusticity across multiple Holocene hunter-gather groups across North America, including Indian Knoll and Point Hope. He found that broad-spectrum hunter-gatherers, such as the Indian Knoll foragers, had significantly stronger femora when compared to fresh-water and marine-based groups. In terms of the humerus, it is unclear how they will compare to the Sadlermiut and Andaman collections, as they had access to water but may not have been as reliant on aquatic food sources. Also, they were not hunting the type of large-scale marine

mammals that these populations exploited. Cowgill’s (2010) study compared the cross-sectional robusticity of Indian Knoll juveniles to Point Hope juveniles and found that Indian Knoll samples were less robust overall than the Point Hope sample. Since

developmental loading is influential on adult morphology, Cowgill’s results may predict a lower-level loading pattern in the upper limbs of the Indian Knoll adults.

The Indian Knoll site was excavated during the first half of the 19th_{Century by} Clarence B. Moore and the Work Progress Administration (Herrmann 2002), and William S. Webb (Webb 1946). The collection is now part of the W.S. Webb Museum and is currently housed at the University of Kentucky in Lexington.

(47)

3.1.3. Point Hope Alaska

The Point Hope sample is made up of several populations from along the Northwest Alaskan coastline, specifically from the Point Hope peninsula. The sample represent varying time periods, from the relatively recent Tigara site (1300-1800 A.D.), to the older Birnirk (500-1000 A.D.), Ipiutak (100 B.C. – 500 A.D.), and Norton (900 B.C. – 100 A.D.) sites (Auerbach 2008). The populations from Ipiutak and Tigara show some variation in resource exploitation, where Ipiutak period individuals hunted both marine and terrestrial animals, while Tigara period individuals subsisted almost completely on marine animals (Dabbs 2011). Despite the rather larger temporal span represented within the collection and the marked cultural changes that occurred between the Ipiutak and Tigara periods, all populations are representative of an extensive marine-based exploitation strategy, wherein large and small marine animals were hunted (Larson and Rainey 1948, Hilton et al., 2014). Past research has also indicated little changes in biomechanical loading between the time periods (Shackelford 2007). Whale hunting as a subsistence strategy in the Northwestern Arctic involves the hunting of bowhead and other baleen whales found in the region from both sea ice and the open water (Whitridge 1999). The whale hunting tradition and has occurred for thousands of years in this area (Hilton et al., 2014). In order to accomplish this task hunting technologies such as toggle harpoon heads, whale harpoon foreshafts, whale lance heads, and other projectile and spear-like hunting tools were used regularly (McCartney 1980). In addition to hunting whale, the Point Hope populations would have also exploited seal and walrus resources as importance sources of food, fuel, clothing and other essential raw materials (Whitridge 1999).

(48)

In keeping with Bergmann’s and Allen’s rules on the effects of cold climate on the evolution of body shape and size, the Ipiutak and Tigara peoples of the Point Hope Peninsula tended to possess a shorter, heavier, cold-adapted body (Holliday and Hilton 2010). There is also evidence that Point Hope individuals demonstrate a relatively round femoral shaft early in life when compared to other hunter-gather groups (Cowgill 2014). But when looking at the relationship between severity of the cold and body shape, not all groups follow the expectations, therefore activity and lifestyle are integral to

understanding bone morphology (Holliday and Hilton 2010). Since Point Hope

populations relied heavily on ocean-based hunting they were extremely marine mobile. In order to hunt such large marine mammals these individuals frequently used umaiks (a large, open, canoe-like boat), kayaks, floats and drag gear (McCartney 1980). This type of activity creates heavy loading patterns on the arms and shoulders, with less (or perhaps differently-patterned) loading on the lower limbs. Of interest is the comparison with the other Arctic foragers, the Sadlermiut, as well as the marine-mobile but geographically distinct Andaman Islanders, since these individuals are shorter, but not cold adapted like those from Point Hope or Sadlermiut. Cowgill’s comparisons between Indian Knoll and Point Hope juvenile humeri have already been discussed, but she also compares juvenile femora between the two populations. Juvenile femora were significantly more elliptical, and more robust, in Indian Knoll when compared to Point Hope (Cowgill 2014).

The four sites that now comprise the Point Hope collection were excavated by Helge Larsen of the Danish National Museum in Copenhagen and Froeligh Rainey of the University of Pennsylvania between 1939 and 1941 (Rainey 1941). Overall, 500

(49)

the research team, the oldest being radiocarbon dated to roughly 1,600 to 1,300 years ago, and the youngest 600-500 years ago (Hilton et al., 2014). The collection is now housed at the American Museum of Natural History in New York.

3.1.4. Sadlermiut

Until the start of the 20th century the Sadlermiut Inuit people lived at Native Point on the southeastern portion of Southampton Island, as well as Walrus and Coats Islands in the Canadian Arctic (Ryan and Young 2013). Frans Boas described their relationship with other Hudson’s Bay Inuit groups. According to him, the Sadlermiut’s neighbors described them as the isolated inhabitants of Sagdlirn, about which they knew little (Boas 1964). In the summer of 1902 a single Sadlermiut man carried an unknown European disease back from a whaling station on the south-western point of Southampton, and by the winter of 1903 the population was almost completely decimated with all but one Sadlermiut woman and four children had died (Boucher 2012). The recent date of the Sadlermiut collections makes them unique in the context of the current study, since in addition to archaeological evidence, we have some historical accounts of how they lived, from both neighbouring Inuit groups and European explorers, but we have relatively little ethnographic information compared with the other recent collection; the Andaman

islanders. The Sadlermiut remains used in this project were excavated in between 1954 and 1959 by Dr. Henry B. Collins, Dr. William Laughlin and Dr. Charles Merbs (Boucher 2012). The individuals excavated died en mass as a result of the widespread disease that wipeout nearly the entire population in 1903. They are curated in the Canadian Museum of History in Ottawa.

(50)

The Sadlermiut are believed to have relied heavily on marine resources, as well as some land animals, like the caribou. During the summer and up until the ice made it impossible, the Sadlermiut would use kayaks to hunt seals, walruses and whales, while in the winter these animals were hunted using ice edges and breathing holes where the mammals could surface for air (Raye Wood 1992). Excavation of the archeological site at Native Point by a number of researchers produced many artifacts (kayak rests and kayak lances) which tied the Sadlermiut to a way of life dependent on watercraft and rowing (Holland 2014). While men were traditional involved in the hunting, women took part in food preparation activities such as processing hides (Raye Wood 1992).

This reliance on watercraft as means by which to hunt makes diaphyseal robusticity and proximal head shape in the humerus of great interest among the

Sadlermiut. Loading patterns are created by both paddling and hide processing. Habitual marine-transport creates mechanical loading regimes on the shoulders and upper limbs (Stock and Pfeiffer 2001, Shaw and Stock 2013). A great amount of bilateral force is put on the upper limbs during hide preparation (Shaw et al., 2012, Boucher 2012). A

comparison between the robusticity and shape measures of the Sadlermiut and the Point Hope foragers would be interest, as both populations were cold adapted, marine hunters living in arctic environments, and thus should show similar body size and bone

morphology characteristics. These same measures can be compared to other marine-mobile populations, such as the Andaman Islanders, or to terrestrial groups such as the Indian Knoll collection.

(51)

3.1.5 Sample Summary

Table 3.1 provides information on the distribution of humerus and femur for each sample used in this study.

Table 3.1. Male and female humerus and femur used in the study.

Humerus Femur

Sample Total N Male N Female N Total N Male N Female N

Andaman Islanders 25 12 13 19 10 9

Indian Knoll 21 10 11 20 9 11

Point Hope 25 12 13 28 15 13

Sadlermiut 20 10 10 17 9 8

Total 91 44 47 84 43 41

These samples were chosen for two main reasons. The first is that each

archaeological sample is representative of forager societies, meaning they were highly active throughout growth and development and throughout life. The second reason is preservation. Each bone in the sample had to be preserved well enough to be able to take cross-sectional measurements of the diaphysis and to be able to set shape-determining landmarks on the proximal end. If the proportion of adult bones in a collection that are damaged in those areas is too high, the collection could not be used. These two issues combined means that there were a set number of collections that could be used in this project. Being that the bone samples taken from these collections belonged to highly active foraging peoples, they were ideal for conducting this type of shape and diaphyseal analysis.