A PALEOPATHOLOGICAL INVESTIGATION ON  
THE PRESENCE OF MALARIA IN MEDIEVAL NUBIA AND  
THE ASSOCIATED SKELETAL LESIONS 

By 

Elena Ora Watson 

A DISSERTATION 

Submitted to  
Michigan State University 
in partial fulfillment of the requirements 
for the degree of 

Anthropology – Doctor of Philosophy 

2024 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

The implementation of saqia water-wheel irrigation in ancient Nubia allowed for an 

expansion of cultivatable land and is considered a primary factor behind the settlement growth 

along the Middle Nile Valley prior to and over the course of the medieval period (~375–1500 

CE). However, these anthropogenic influences are believed to have created conditions favorable 

to mosquito populations and concomitantly, may have resulted in elevated malaria during the 

medieval period. The first goal of the dissertation was to investigate this hypothesis in examining 

skeletal remains excavated from two cemeteries at the Fourth Cataract site of Mis Island. 

Towards this objective, a method was used that involved assessing the presence of five skeletal 

lesions that have been implicated in malarial infection. To build on current perceptions of these 

lesions, the second goal of this research was to examine the lesions’ association with age of the 

individual and the association among the lesions. For one of these lesions—femoral cribra—

there is a need to further understand variation in the lesion’s external appearance and the bone 

processes underlying its expression. Therefore, the third focus of this dissertation explored 

several variables visible on the lesion’s surface possibly associated with “activity” and “severity” 

which were also investigated using micro-computed tomography (micro-CT) methods. 

A relatively high estimated prevalence of malaria was observed in the skeletal remains 

excavated from Mis Island. These findings are consistent with the hypothesis of elevated malaria 

in the region during the medieval period associated with the effects of settlement growth and 

expansion of irrigated land that accompanied saqia agriculture intensification. Comparisons were 

also conducted between the two Mis Island cemeteries to examine potential temporal differences, 

and between sex and age cohorts to investigate relative exposure and possible endemicity of 

malaria. In comparing the two cemeteries, no significant differences in estimated malarial 

 
 
infection were found. Additionally, no significant differences in estimated malaria were observed 

between males and females. A significant association between estimated malaria and age was 

found, and the elevated frequencies across age groups may indicate a primarily endemic type of 

malaria in the region during the medieval period. 

In examining the association between the suite of lesions studied in this dissertation and 

age of the individual, the observed age-related patterns of the porous lesions were consistent with 

the hypothesis that their formation may be related to anemia. Furthermore, significant 

associations were observed between these lesions in the subadult age groups, but not among 

adults. These significant associations were between most of the porous lesions included in the 

suite of lesions and suggest shared processes involved in their formation. 

The results of femoral cribra’s possible “activity” and “severity” based on its external 

appearance provided preliminary support for such interpretations. Femoral cribra lesions 

estimated to be active or healing from the surface followed an age-related pattern anticipated for 

such statuses. Additionally, micro-CT analyses demonstrated trabecular differences in lesions 

that appeared healing consistent with more bone-forming processes. Investigating the severity of 

femoral cribra suggested a possible lesion progression of increased trabecular exposure and size 

visible on the surface which may be associated with resorptive processes and/or marrow 

hyperplasia. 

 
 
 
This work is dedicated to my parents. 
Thank you for your incredible support throughout this journey. 

iv 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ACKNOWLEDGEMENTS 

The encouragement and assistance from my mentors, colleagues, family, and friends were 

integral to arriving at this milestone. First, I am immensely grateful for the guidance and support 

from my committee members, Drs. Todd Fenton, Joseph Hefner, Ethan Watrall, and Gabriel 

Wrobel, throughout this research process and over the course of my graduate career. My sincere 

thanks to my committee Chair, Dr. Fenton, for believing in me from the beginning and providing 

so many opportunities for me to grow as a scholar and individual. Your excitement for this 

research spurred me forward and our discussions about these topics were instrumental in its 

development and progression. Thank you, Dr. Hefner, for all the knowledge you shared with me 

that made me a better osteologist and researcher, for teaching me how to use R, and for all the 

humor over the years. Thank you, Dr. Watrall, for your guidance on topics regarding archaeology 

of ancient Nubia and for your support in contextualizing this research within broader 

perspectives. Thank you, Dr. Wrobel, for helping me develop a biocultural lens and for directing 

me towards and facilitating the use of micro-CT methods in this project.  

I would like to express my sincere appreciation for the efforts of everyone involved in the 

initial excavations and subsequent curation of the human skeletal remains and archaeological 

material from Mis Island. Thank you to The British Museum, Sudan Archaeological Research 

Society, and Michigan State University for their roles in these processes and all who were part of 

the archaeology and research teams associated with Mis Island. I am particularly grateful for the 

work by Dr. Angela Soler and Dr. Carolyn Isaac and the impact their research had on my own. To 

Dr. Isaac, thank you also for your insights on this project and your kind mentorship since the 

beginning. I would also like to thank Lilli Antonelli and Savannah Sass for their assistance 

during the data collection process in the lab during the early phases of this research. My thanks 

v 

 
also to alumni Dr. Emily Streetman and Dr. Jared Beatrice for their insights related to my 

research interests. 

The micro-CT component of this dissertation would not have been possible with the 

assistance of Jeremy Hix of the MSU Radiology IQ Advanced Molecular Imaging Facility and 

the Institute for Quantitative Health Sciences and Engineering. Thank you so much, Jeremy, for 

all your help in scanning and your guidance with the software for analysis.  

My many thanks for the support provided by the Department of Anthropology and 

College of Social Science. I would like to thank our current and previous Graduate Program 

Directors, Drs. Stacey Camp and Mindy Morgan, and department staff including Joan Reid, 

Roxanne Moran, Jocelyn Janicek, Kathy McGlynn, and Cathi Pierce. Thank you to everyone 

with the Division of Human Anatomy for supporting me and cheering me on to the finish line. 

My particular thanks to Dr. Nicole Geske, Dr. Lindsey Jenny, Dr. Melanie McCollum, Bill 

McMillan, Alex Pereida, Tim Currie-Dennis, Erin Lyon, and Josh Vincent. 

I have had the great honor to meet and work with amazing fellow graduate students 

during my time at MSU. Dr. Mari Isa and Alex Goots are two of the most caring people I know, 

and I am forever grateful for their friendship. Memories of our adventures will always bring me 

joy. To Dr. Jen Vollner, Dr. Caitlin Vogelsberg, Dr. Emily Streetman, Dr. Amber Plemons, Dr. 

Kelly Kamnikar, Micayla Spiros, Rhian Dunn, Clara Devota, and Aubree Marshall, thank you so 

much for your camaraderie over the years. 

Finally, I cannot thank my family and friends enough for everything, especially my 

parents and brother. It certainly “took a village”—thank you to everyone part of that village not 

named here. 

vi 

 
 
 
TABLE OF CONTENTS 

CHAPTER ONE: INTRODUCTION ..........................................................................................1 

CHAPTER TWO: BIOARCHAEOLOGY OF MALARIA, MALARIA IN THE NILE RIVER 
VALLEY, & ARCHAEOLOGY OF ANCIENT NUBIA ..............................................................9 

CHAPTER THREE: PATHOPHYSIOLOGY OF MALARIA, PALEOPATHOLOGY OF 
ANEMIA, & SKELETAL LESIONS OF INTEREST ............................................................... 38 

CHAPTER FOUR: RESEARCH QUESTIONS & EXPECTATIONS ....................................... 52 

CHAPTER FIVE: MATERIALS & METHODS ....................................................................... 65 

CHAPTER SIX: RESULTS ...................................................................................................... 85 

CHAPTER SEVEN: DISCUSSION AND CONCLUSIONS ................................................... 145 

LITERATURE CITED ............................................................................................................ 177 

vii 

 
 
 
 
 
 
 
 
 
 
CHAPTER ONE: INTRODUCTION 

Malaria has a pervasive presence around the world that has afflicted generations of 

people living in environments conducive to the mosquito, which transmits the disease-causing 

parasite. Within these contexts, there is a dynamic and interdependent interaction between human 

hosts, mosquito vectors, and malarial parasites as human activity within the environment can 

further influence risk and exposure to this disease. For people living in the Nile River Valley, 

malaria has historically had a significant effect on morbidity and mortality. In ancient Nubia, 

malaria is believed to have posed a high burden and risk during the medieval period due to an 

increase in human population and settlement along the Nile River facilitated by an intensification 

of saqia water-wheel irrigation agriculture.  

Prior to and over the course of the medieval period (375–1500 CE), there was a 

significant population increase along the Middle Nile Valley, with a rapid growth of rural 

settlements in Northern Nubia and the Dongola Reach during the sixth century (Obłuski, 2020). 

This expansion is attributed to the heightened implementation of saqia water-wheel irrigation, 

which facilitated farming for extended harvesting seasons and enabled a population growth in 

areas that were previously less inhabited along the Nile (Malcom et al., 2007; Obłuski, 2020). 

However, the intensification of saqia irrigation during this period is believed to have created 

conditions posing a particularly high risk of malaria (Malcom et al., 2007). These shifts in human 

population and farming practice with an accompanying increase in cultivatable land are 

hypothesized to have created conditions favorable for an increase in mosquito populations and 

concomitantly, malaria. With such circumstances favorable for increased malarial risk, this 

irrigation technology may have had deleterious impact on public health (Judd, 2004). 

1 

 
This dissertation research investigates hypothesized malarial infection in Nubia during 

the medieval period and the potential impact of these events using paleopathological approaches 

and a biocultural framework. The dissertation focuses on the skeletal remains excavated from 

two medieval cemeteries at the Upper Nubian site of Mis Island, located near the Fourth Cataract 

of the Nile River. As part of the Merowe Dam Salvage Archaeology Project (MDASP), the Mis 

Island skeletal sample was excavated by the British Museum and the Sudan Archaeological 

Research Society and was curated at Michigan State University. To investigate the presence of 

malarial infection, this research used an approach developed by Smith-Guzmán (2015a) which 

entails examining skeletal remains for a series of five lesions as possible indicators of the 

disease. These lesions are cribra orbitalia, humeral cribra, femoral cribra, spinal porosity, and 

periosteal reaction. A diagnostic outcome algorithm was then used based on these lesions to infer 

how many and which individuals may have been infected with malaria. While some of these 

skeletal lesions have been the focus of research on their expression and co-occurrence, others 

have received less attention and the combination of all five has not been extensively examined 

since Smith-Guzmán’s (2015a) study. 

Within the field of bioarchaeology, there is a need to thoroughly understand the formation 

and expression of paleopathological lesions regarding the processes involved in their 

manifestation and variation among individuals (DeWitte and Stojanowski, 2015). Towards this, 

an evaluation of skeletal lesions encountered in archaeological material can benefit from an 

approach that involves systematically examining the morphology and distribution of lesions, and 

considering the biological processes associated with the anatomical site (Mays, 2018). To build 

on current perceptions of the skeletal lesions implicated in malarial infection, this dissertation 

therefore examines the lesions’ relationship with the affected individual’s age, as well as the 

2 

 
 
association among these lesions, to contribute further insight into their development and patterns 

of expression. For one of these lesions—femoral cribra—there is a need to further understand 

variation in the lesion’s external appearance and the bone processes underlying its expression. 

Femoral cribra is a porous lesion on the anterior-inferior femoral neck that has been 

surrounded by a history of disagreement and ambiguity, which has extended into current 

understandings of the lesion (Göhring, 2021). In addition to uncertainties regarding its etiology, 

variation in its appearance has led to contrasting interpretations of its presence. With current 

understandings of femoral cribra as a pathological stress marker, associated insight into possible 

lesion “severity” and “activity” is needed. Such information aids in interpretating its presence, in 

that assessing whether the lesion might be active or healing can help in separating disease 

processes that were ongoing at the time of death from those that may have been inactive (Wood 

et al., 1992). To address these needs, this dissertation research involved visually examining 

femoral cribra lesions for several variables possibly representing lesion “severity” and “activity” 

and assessed their association with the age of the individual. Additionally, micro-computed 

tomography (micro-CT) methods were used to investigate the relationship between changes 

visible from femoral cribra’s external surface and the types of bone processes occurring at the 

trabecular level. 

The site of Mis Island 

The site of Mis Island was situated upstream of the Nile’s Fourth Cataract in modern day 

Sudan. Excavation of Mis Island, as well as multiple sites along the Fourth Cataract, began as 

part of a large salvage archaeological project initiated by the Sudan Archaeological Research 

Society (SARS) in response to the planned construction of the Merowe Dam. Now inundated by 

the dam, the site lies under the Merowe Reservoir. Mis Island was less than 2km across east to 

3 

 
west and approximately 1km north to south. Before being flooded, the islands in this stretch of 

the Nile were described as having a desert landscape at their centers with oasis-like vegetation 

closer to the riverbanks (Welsh, 2013). The riverbanks of these islands at the Fourth Cataract 

were rocky with narrow areas of fertile soil (Näser, 2007). The Nile Valley adjacent to Mis Island 

and its neighboring islands was closely bordered by desert (Welsh, 2013). 

Figure 1.1: Map of medieval Nubia showing the location of Mis Island, adapted from Hurst 
(2013:28) (star added to mark location of Mis Island). 

In the 1980s, the Sudan government renewed interest in a project first proposed in 1942 

to build a dam at this region of the Nile (Ahmed, 2014a; Welsby, 2003). The National 

Corporation for Antiquities and Museums (NCAM) recognized that the engineering activities 

and concomitant flooding associated with this dam had the potential to affect numerous 

archaeological sites in the Fourth Cataract region (Ahmed, 2014a). In response to the anticipated 

inundation of archaeological sites, the NCAM recruited several international archaeological 

groups to be involved in the Merowe Dam Archaeological Salvage Project (MDASP) before 

completion of the dam in 2009 (Ahmed, 2014a; Welsby, 2003). Mis Island resided within the 

4 

 
 
 
concession headed by the Sudan Archaeological Research Society (SARS) which involved 

approximately 40km of the Nile’s left bank between Amri and Kirkbekan and the immediate 

islands (Ahmed, 2014a; Welsby, 2006). The SARS concession was headed by Project Director 

Derek Welsby and the Field Director for the area including Mis Island was Andrew Ginns 

(Ahmed, 2014a). First surveyed in 1999, a total of fifteen archaeological sites on Mis Island were 

identified, six of which were investigated further through excavations in the 2005–2006 and 

2006–2007 field seasons by SARS and the British Museum (Ginns 2007, 2006). The primary 

sites excavated are as follows: a Late Christian church and associated cemetery referred to as 3-

J-18 near the center of the island; cemetery 3-J-10 located 300m north of the church; cemetery 3-

J-11 at the northern border of the island; cemetery 3-J-20 at the highest elevation on the east side 

of Mis; medieval settlement 3-J-19 located 50m east of the church and north of cemetery 3-J-20; 

and Kerma cemetery 3-J-22 50m south of cemetery 3-J-11 (Ginns 2007, 2006). 

While archaeological findings demonstrate that Mis Island was inhabited in the Kerma 

and Meroitic periods and through the medieval phases, the majority of work on Mis Island 

concentrated on its medieval occupation dating from the mid-fifth to early fifteenth centuries CE 

(Ginns, 2006). Mis Island represented a small rural community situated in the territory of the 

medieval Nubian kingdom of Makuria (Welsby, 2002). Medieval settlement at Mis Island is 

observed by its primary sites of a late Christian church, a settlement, and several cemeteries. 

Also observed on Mis Island are Muslim burials associated with the end of the medieval period 

(Ginns, 2010a) which were not excavated to comply with religious observations (Soler, 2012). 

The Muslim cemetery remained in use by modern inhabitants until construction of the dam, 

demonstrating the continuous occupation of Mis Island (Ginns, 2010a). 

5 

 
Through multiple bioarchaeological research studies, the site of Mis Island has provided 

a unique window into medieval Nubian history and rural life. Previous paleopathological 

research on Mis Island includes examining the effects of stress and disease on the skeletal health 

of adults (Soler, 2012) and subadults (Hurst, 2013) excavated from two cemeteries: cemetery 3-

J-10 (1100–1500 CE) and cemetery 3-J-11 (300–1400 CE). Among the adult individuals from 

these cemeteries, Soler (2012) found a high frequency of paleopathological lesions suggestive of 

stress that are not specific to a certain condition; however, the majority of these lesions 

demonstrated healed expressions. In examining these non-specific lesions in the subadult 

remains, Hurst (2013) observed a higher prevalence of stress indicators compared to the adult 

group and more active expression of the lesions in younger individuals. Soler and Hurst propose 

that the inhabitants of Mis Island during the medieval period experienced high levels of stress 

early in life but were also able to survive and recover from stress experienced in childhood and 

adulthood. 

This study focuses on the following research goals: 

Research goals 

•  To investigate the skeletal indication of malaria at Mis Island during the medieval period 

and how prevalent the disease may have been within the context of settlement patterns 

and irrigation technology.  

•  To examine potential temporal fluctuations in the disease during the medieval period 

potentially associated with changes in the Nile River’s levels. 

•  To assess these individuals’ experiences of malaria as they relate to possible exposure and 

endemicity of the disease.  

6 

 
•  To consider the formation and consistency in the skeletal lesions’ expression associated 

with hypothesized malarial infection. 

•  To explore patterns between variables observed on the external surface of femoral cribra 

potentially related to lesion “severity” and “activity” and if these are supported by the 

underlying trabecular microarchitecture. 

Malarial risk and transmission are heavily influenced by human activity and interactions 

with their environment. This research intends to contribute to our understanding of the host-

pathogen relationship within the context of events taking place in medieval Nubia. Investigating 

the dynamics of malaria in Nubia during the medieval period builds on the scholarship of health 

conditions for these communities and contributes insight into the potential influence of human 

activity on the endemicity of malaria during this time. With further understanding of the 

expression and formation of the skeletal lesions suggestive of malaria, this research also seeks to 

elucidate discussions regarding anemia, porous lesions, and changes in red bone marrow. A more 

coherent awareness of the biological processes underlying femoral cribra can be used in 

conjunction with other lines of evidence to construct appropriate interpretations from 

archaeological material. Recognizing differences in lesion expression possibly related to 

“severity” and “activity” can further aid in investigating disease pathogenesis and 

epidemiological patterns.  

Organization of the dissertation research 

Towards addressing these research objectives, this dissertation is organized into the 

following chapters. Chapter Two discusses bioarchaeological research on malaria and insights 

into the disease in the Nile River Valley, including the dynamic ecology of the Nile River and 

previous research on malaria in this setting. The archaeology of ancient Nubia is then discussed 

7 

 
with a focus on the archaeological understandings and historical context of ancient Nubia 

spanning the Napatan to medieval periods (~800 BCE –1500 CE). Chapter Three presents 

research relevant to the paleopathology of anemia and the skeletal lesions implicated in malarial 

infection. After highlighting skeletal considerations in the pathophysiology of malaria, the 

trajectory of research on anemia in paleopathology is presented. Previous research examining the 

association between the skeletal lesions of interest and age of the individual, as well as the co-

occurrence among the skeletal lesions, is also reviewed. Finally, previous work regarding the 

lesion femoral cribra is described including its proposed etiologies, variation in morphology on 

the lesion’s surface, and its correlation with age of the individual. 

Chapter Four details the research questions examined in this dissertation and the expected 

outcomes of each question with supporting information. Chapter Five describes the materials and 

methods associated with the dissertation and is organized by research questions that share 

common methods, subsamples, and statistical approaches. The skeletal remains that comprise the 

Mis Island sample are first discussed, followed by the data collection procedures, and finally the 

data analysis methods. In Chapter Six, the results of the data analyses are presented and 

organized by research question with descriptions regarding the patterns observed and the 

associated significance of the statistical tests. Finally, Chapter Seven provides a discussion that 

interprets and contextualizes the results of the data analyses. This chapter begins with a 

discussion on the estimated presence and prevalence of malaria at Mis Island during the 

medieval period, followed by interpretations of the investigation into the skeletal lesions 

implicated in hypothesized malaria, and finally a synthesis on the exploration of femoral cribra. 

Concluding thoughts, limitations of the study, and directions for future research are also 

presented at the end. 

8 

 
 
CHAPTER TWO: BIOARCHAEOLOGY OF MALARIA, MALARIA IN THE NILE 

RIVER VALLEY, & ARCHAEOLOGY OF ANCIENT NUBIA 

This chapter begins with a review of researching malaria in bioarchaeology, including the 

use of a biocultural model and the current methodological approaches. The next section provides 

an overview of the ecological context of the Nile River as it relates to the malaria parasite’s 

vector—mosquitoes—and discusses previous research examining malaria in the Nile River 

Valley. The final section presents archaeological insights and historical context of ancient Nubia 

spanning the Napatan to Medieval periods (~800 BCE–1500 CE). After providing a geographical 

and historical background, this section presents the major archaeological understandings 

progressing through the following time periods: Napatan Kingdom, Meroitic Empire, Post-

Meroitic Period, and Medieval Period. 

Bioarchaeology of malaria 

Given the dynamic and interdependent interaction between human hosts, mosquito 

vectors, and malarial parasites in suitable environments, a biocultural approach is well-suited to 

investigations of malaria in past populations. A biocultural model enables exploration of the 

associations between social and cultural practices and their impact on human biology to generate 

inferences regarding archaeological questions and to better understand human-disease systems. 

In bioarchaeological research on malaria, this approach involves the incorporation of 

archaeological, epidemiological, historical, entomological, environmental, and when available, 

genetic insights and information related to the disease process of malaria (Marciniak et al., 2018; 

Smith-Guzmán et al., 2016). Such work thus examines how the presence of malaria and its 

multifaceted epidemiology is influenced by variables related to disease ecology and interactions 

between human hosts and their environment (Marciniak et al., 2018). A biocultural model to 

9 

 
 
studying malaria in archaeological contexts involves examination of sociocultural, biological, 

and environmental variables operating at the macro- and micro-levels that interact to influence 

malarial infection at the archaeological site(s) of interest. Social and economic forces that could 

have influenced the dynamics of malaria in past communities include agriculture, population 

density, migration, trade, and construction activity (Marciniak et al., 2018). If such information is 

available, consideration of the presence and frequency of genetic conditions that provide some 

resistance to malaria and lessen infection severity (e.g., G6PD, thalassemia, sickle cell) should be 

included (Marciniak et al., 2018). However, in their study of malaria in Imperial Italy, Marciniak 

and colleagues (2018) note that the contemporary frequencies of resistance genes may not be 

directly representative of the past population they studied.  

By drawing on these lines of evidence to study past malarial dynamics in a given spatial 

and temporal context, various research questions can be examined that range from more local to 

broad-scale foci. Within the setting of 1st–4th centuries central-southern Italy, Marciniak et al. 

(2018) reflect on comparing the presence of malaria between several sites that differ in 

environmental conditions, population structure and movement, and economic activity—

discussing how such factors can impact experience of the disease within those communities. 

Bioarchaeological inquiry into malaria has also investigated the possibility of this disease as the 

source behind specific epidemics that significantly impacted past populations. Smith-Guzmán 

and colleagues (2016) explore the possibility of malaria as a disease behind the “Hittite plague” 

(1322 BCE) in a study of skeletal remains from Amarna, Egypt (1349–1332 BCE). This study 

also provides an example of considering epidemiological patterns of malaria (i.e., epidemic or 

endemic malaria).  

10 

 
With reference to climate change and its associated impacts on disease, Gowland and 

Western (2012) use a spatial epidemiological approach to explore the prevalence of skeletal 

indicators of poor health as signs of malaria in eastern England during the early to mid-Anglo-

Saxon period—a time of climatic events that resulted in flooding of low-land areas and warmer 

temperatures in the region. The authors contextualize their study with historical records that 

describe endemic outbreaks consistent with malarial infection associated with the region’s low-

lying wetland areas. Gowland and Western map and compare the frequency of cribra orbitalia 

with the distribution of large populations of Anopheline mosquitos, lower altitude and marshy 

environments, and higher incidence of historic references to malaria across England. Gowland 

and Western conclude that the observed pattern of cribra orbitalia is most likely attributable to 

endemic malaria, in conjunction with coexisting diseases that can also contribute to severe 

malarial anemia. Another broad-scale study of malaria in the past, conducted by Smith-Guzmán 

(2015b), investigates the disease’s prevalence and spread within the ancient Nile Valley region. 

The goal of this research was to analyze the spatiotemporal variation in reported cribra orbitalia 

rates from archaeological sites along the Nile Valley to infer the possible pervasiveness and 

distribution of malaria in the region, while also considering theoretical models of malaria’s 

spread out of Africa.  

While a correlation between cribra orbitalia and malaria has been presented, an analytical 

method developed by Smith-Guzmán (2015a) provides a more specific model for investigating 

the potential presence of malaria in past populations. Smith-Guzmán used an epidemiological 

approach to examine the prevalence of a series of paleopathological lesions in two modern 

skeletal samples. One of these samples was comprised of individuals from Uganda and adjacent 

East African countries where malaria is holoendemic and widespread in the population, and the 

11 

 
 
other sample included individuals from the United States where malaria has been eradicated. In 

comparing these two samples, five lesions were observed more frequently in the sample endemic 

with malaria compared to the malaria-free sample. These five skeletal lesions are cribra orbitalia, 

humeral cribra, femoral cribra, spinal porosity, and periosteal reaction. Based on the patterns and 

associations of these lesions, Smith-Guzmán developed an algorithm of criteria that could 

indicate whether an individual may have been infected with malaria. Along with additional lines 

of evidence (e.g., archaeological, historical, ecological) in a study’s context, the presence of 

these lesions has been interpreted as support for the possible presence of malaria within the 

sample (Buzon and Sanders, 2016; Smith-Guzmán et al., 2016). 

In addition to gross skeletal examination, malarial detection methods in paleopathology 

include ancient DNA (aDNA) and immunological testing. Setzer (2014) describes gross 

examination of skeletal remains as associating skeletal indicators of anemia with ecological and 

historical information. While this approach provides insight into possible prevalence of infection, 

skeletal lesions are not specific to malaria as other disease processes could contribute to their 

formation. To test for malaria directly in archaeological remains, aDNA analysis is used and is 

considered the most definitive method for identifying malaria in archaeological remains. 

Malaria in the Nile River Valley 

Environmental conditions along most of the Nile River provide suitable habitats for the 

mosquito vector Anopheles arabiensis, the local populations of which are constrained to the 

marshy riverbanks and dependent on human population density (Malcolm et al., 2007). The 

annual flooding of the Nile offers particularly favorable breeding conditions for An. arabiensis 

mosquitoes as isolated pools are created from receding river levels (Scheidel, 2001). In a study 

of these mosquito vectors between the Third Cataract and Abu Hammed in the modern Northern 

12 

 
 
State of Sudan, Ageep and colleagues (2009) observed a seasonal pattern in mosquito population 

numbers that is associated with the fluctuations of the Nile River’s levels. Most of the An. 

arabiensis presence in this region is focused along the main Nile River channel. As the river 

level rises, mosquito numbers decrease due to the disruption of breeding sites. However, as the 

river level falls again, stagnant pools of water are created which provide optimal breeding sites 

and a concomitant increase in population size. This ecological relationship creates a seasonal 

pattern observed by Ageep et al. during which larvae populations are lowest between July and 

October and quickly grow when the river recedes in November. While this cycle occurs along the 

river, the authors note how artificial habitats located away from the channel may not be affected 

by such fluctuations and thus, provide additional breeding sites that are more consistent. 

Accounts from ancient Egypt describe a surge in mosquitoes following the Nile’s 

inundation and reference an accompanying risk of illness that implicates malarial infection 

(Sabbatani et al., 2010; Scheidel, 2001). Multiple studies have detected the presence of malaria 

in ancient Egypt through ancient DNA analysis (Hawass et al., 2010; Lalremruata et al., 2013; 

Nerlich et al., 2008) and immunological techniques (Al-Khafif et al., 2018; Bianucci et al., 2008; 

Loufouma-Mbouaka et al., 2020; Miller et al., 1994; Rabino Massa et al., 2000). Similar work in 

ancient Nubia is represented by the detection of falciparum malaria in an individual from North 

Argin near the Second Cataract during the Ballana period (350–550 CE) (Miller et al, 1994). 

While these confirmatory tests provide evidence for the presence of malaria, paleopathological 

approaches examining skeletal lesions have provided insight into its prevalence and extent in the 

ancient Nile River region. 

Smith-Guzmán’s (2015a) method of estimating malaria from a series of skeletal lesions 

has been used in two studies from the Nile River Valley region. In a study of skeletal remains 

13 

 
 
from Amarna, Egypt (1349–1332 BCE), Smith-Guzmán and colleagues (2016) explore the 

possibility of malaria as a disease behind the “Hittite plague” (1322 BCE). The prevalence of 

skeletal lesions in the sample indicates endemic malaria at Amarna during this time based on 

about half of the individuals showing signs of infection and a higher morbidity and mortality of 

subadults and reproductive-age females. This approach was also used in studying the skeletal 

remains excavated from Tombos (~1400–650 BCE), an Upper Nubian site located at the Third 

Cataract (Buzon and Sanders, 2016). Examining the skeletal lesions suggestive of malaria at 

Tombos revealed a relatively prevalent rate of infection with abundant spinal lesions. Whereas 

Amarna and Tombos represent urban centers within their respective contexts, this approach for 

exploring the prevalence of malaria has not been conducted on a rural settlement in the Nile 

River Valley, or at a medieval site when malarial burden is believed to have been elevated in 

Nubia. 

Archaeology of Ancient Nubia 

The geographical scope of Nubia covers the land in modern-day Egypt and Sudan 

between the First Cataract of the Nile River and the convergence of the White and Blue Nile 

Rivers. Coursing through this region is the Nile River with the Eastern and Western deserts on 

either side (Barbour, 1961). The Nile Valley encompasses the river and adjacent alluvial land, 

which typically does not exceed 2km in width (Barbour, 1961). A series of rocky cataracts 

positioned along the Nile River serve as points of reference when dividing the Nubian Nile 

Valley into regions. The main Nile maintains its flow throughout the year with water supply from 

the White Nile and rapidly floods during the summer from the Blue Nile (Barbour, 1961). 

Besides the Nile River, most of the waterways in Nubia do not hold water year-round (Barbour, 

1961). Therefore, the Nile River Valley is vital to human settlement in this region. As 

14 

 
 
demonstrated by Butzer’s (1976) ecological approach to studying ancient Egypt, the Nile and its 

fluctuations in water supply had pervasive effects on inhabitants of the Valley regarding their 

settlement patterns, irrigation strategies, population capacities, and cultural institutions over time.  

The earliest formal archaeological research on ancient Nubia commenced in the 

beginning of the twentieth century, shortly after Sudan became under the control of Britain and 

Egypt (Welsby, 2004). Throughout the history of ancient Nubian archaeology, much of the work 

has been driven by development and construction of dams on the Nile River and thus represent 

rescue operations. While these salvage excavations are a focus of the following discussion, it is 

important to recognize the contributions to our understandings of ancient Nubia from past and 

present archaeological research not associated with such efforts. The first wave of formal 

archaeological investigation into ancient Nubia was initiated by the construction and subsequent 

heightening of the Aswan Low Dam at the turn of the twentieth century. In response to imminent 

flooding, the Egyptian government formed the First Archaeological Survey of Lower Nubia in 

1907 to uncover and document historical sites in danger of being lost south of the First Cataract 

(Reisner, 1910). This work, led by American archaeologist George A. Reisner, marked the first 

systematic archaeological research of ancient Nubia (Edwards, 2004). The Second 

Archaeological Survey of Nubia (1929–1934) expanded archaeological study further south 

towards the border of Sudan (Edwards, 2004). The first substantial archaeological excavation in 

Upper Nubia was led by John Garstang at Meroe around the same time as Reisner’s survey 

(Edwards, 2004).  

A second phase of formal archaeological research was initiated after 1959 in preparation 

for the construction of the Aswan High Dam. This work represented the first regional surveys of 

the Sudan and extended formal archaeological research to between the Second and Third 

15 

 
 
Cataracts (Edwards, 2004). The first published synthesis of contemporary archaeological work 

on Nubia was written by W.Y. Adams (1964, 1965, 1966), who incorporated the accumulation of 

findings into a broad reconstruction of ancient Nubia, largely through the lens of processual 

archaeology. The Aswan High Dam greatly impacted ancient Nubian archaeology as it 

permanently damaged areas along the northern Nubian Valley and directed subsequent attention 

towards more southern sites (Edwards, 2004).  

In the early twenty-first century, a third series of rescue excavations were conducted as 

part of the Merowe Dam Archaeological Salvage Project (MDASP). This international research 

project was in response to the planned construction of the Meroe Dam at the Fourth Cataract and 

provided new insight from the middle Nile Valley. These three general periods of rescue 

archaeology have contributed to the foundation of archaeological work on ancient Nubia and 

played significant roles in shaping its course. However, the salvage nature of this work presents 

its own biases and complications with incomplete sampling and an inability to revisit sites with 

newer approaches. 

The early archaeological investigations of ancient Nubia laid the foundation for future 

scholarship—for example, the chronology and pottery typologies developed by Reisner and the 

Harvard-Boston expedition have reached modern study. However, several issues and 

shortcomings of the initial body of work have had a lasting legacy and should be considered in 

discussions of ancient Nubia. The earliest group of scholars studying ancient Nubia had 

backgrounds in Egyptology and therefore, viewed Nubian culture through an Egyptological lens 

with Egyptocentric biases (Edwards, 2004; Welsby, 2004). Due to this partiality, analyses were 

constructed around an ancient Egyptian narrative and many ancient Nubian indigenous features 

16 

 
were ascribed to Egyptian influence (Edwards, 2004). While Nubia is traditionally viewed as a 

reflection of Egypt, its role as Egypt’s source of wealth should be recognized.  

Due to this Egyptological approach dominant in early works, research focused on 

studying the interactions between Egypt and Nubia (Welsby, 2004). A lasting effect this tradition 

has held over Sudanese archaeology is that it has secluded the region from being incorporated 

into work in other areas of Africa (e.g., Chad Basin, Ethiopian plateau), where important ties 

may have existed (Edwards, 2004; Welsby, 2004). This northwardly emphasis with ancient 

Egypt, coupled with the initial Aswan Dam construction, also contributed to early studies 

concentrating on sites in Lower Nubia. The lack of research and insight from Upper Nubia in 

earlier works is often recognized, but was not comprehensively addressed until recently with an 

increase in southern excavations precipitated by dam construction. Initial investigations also 

tended to gravitate towards visually impressive monumental sites from certain time periods 

(Welsby, 2004). In placing greater attention on royal and elite archaeology, information about the 

non-elite populace is lacking in areas of ancient Nubia’s history. Significant shifts in the 

archaeology of ancient Nubia have occurred since about a century after the first studies started in 

northern Nubia (Edwards, 2004). The following discussion presents the major archaeological 

understandings organized into time periods beginning with the Napatan Kingdom through the 

medieval period. This chronology does not extend more into the past to include a discussion of 

the Kerma and Late Kerma periods as they are outside the focus of the current research. 

Napatan Kingdom (800–300 BCE) 

The Kingdom of Kush is traditionally divided into two main phases—Napatan and 

Meroitic—separated by the relocation of the royal cemetery from Napata to Meroe around 300 

BCE (Edwards, 2004). The Napatan Kingdom was centered in the Dongola Reach and extended 

17 

 
from the upper valleys of the White and Blue Rivers to a northern frontier that fluctuated 

considerably over this period depending on their situation with Egypt (Welsby 1996). The lack of 

evidence pertaining to Napatan origins limits associated interpretations; however, by the eighth 

century BCE, Alara was committed to the cult of Amun which was fundamental in establishing 

the Napatan 25th Dynasty and legitimizing kingship. 

In the early stages of the Kushite state formation, there was conflict in Egypt between the 

pharaoh and the temple of Amun (Welsby, 1996). Egypt was politically divided into different 

states by the eighth century BCE, independently ruled by regional kings (Edwards, 2004). These 

circumstances set the stage for the Kushite advancement into Egypt and the establishment of the 

25th Dynasty Nubian pharaohs as rulers of both Nubia and Egypt. The reverence towards Amun 

shown by the early Napatan rulers was a strong catalyst in their taking control of Egypt (Welsby, 

1996). While the motives of the 25th Dynasty were ideologically deterministic, the political 

aspirations of these rulers should also be acknowledged in their efforts to legitimize authority. 

The 25th Dynasty included six Nubian rulers who reigned between approximately 740–660 BCE 

(Edwards, 2004). After war broke out with the Assyrians and a period of aggression ensued, the 

Kushite rule in Egypt was eventually ended by the Assyrians. 

Settlement during the Napatan period was focused along the Dongola Reach with centers 

around Tabo and Kawa (Edwards, 2004). Beyond a relatively substantial settlement around the 

Second Cataract, evidence of Napatan settlement in Lower Nubia is minimal (Edwards, 2004). 

The thinly scattered settlement pattern in Lower Nubia may reflect Napatan political and 

economic intentions to create enough of a connection between two foci—the Kushite centers 

further south and Egypt to the north (Edwards, 2004). By the 25th Dynasty, Napatan settlement is 

visible at Meroe, which would later become the Kushite capital (Edwards, 2004). With the 

18 

 
 
revival of the Amun cult, Jebel Barkal expanded from continuous construction of Napatan 

temples and palaces by the Kushite kings (Edwards, 2004). In the Bayuda region and Western 

Desert, remnants of Napatan settlements may suggest some conflict between Kushite and non-

Kushite people (Lohwasser, 2014). This theory of conflict is also indicated by historical records 

describing aggression between Kushites and the “Western Desert People”. 

The earliest royal Nubian pyramid burials were established el Kurru and Nuri, with five 

of the six 25th Dynasty pharaohs interred at el Kurru and Taharqo marking the beginning of the 

cemetery at Nuri (Edwards, 2004). The pyramid tombs of the Napatan pharaohs and their 

associated mortuary features demonstrate the significant ties the royal elite had to Egyptian 

politics and ideologies. Towards the end of the Napatan period after the death of Aspelta, the 

pyramids became increasingly untidy with poorly imitated hieroglyphs, representing a dissolving 

connection with Egyptian culture (Adams, 1964). 

Knowledge of non-royal Napatan burials is broadly limited; however, some insight into 

mortuary ideologies over the Napatan period has been illuminated. A large cemetery at Sanam 

and two smaller ones at Meroe contained an array of mortuary customs, as some conveyed more 

Egyptian traditions and others displayed more indigenous pre-Napatan styles (Edwards, 2004). 

Pyramid burials similar to those at Nuri but on a smaller scale at these sites signify the presence 

of local elites. Non-elite Napatan burials uncovered around the Fourth Cataract also demonstrate 

a great degree of variation in mortuary characteristics and tomb superstructures, with most 

representing a continuation of traditions from previous time periods (Ahmed, 2014b). The 

variation in Egyptian and Nubian traditions visible in Napatan burials preclude the identification 

of a “typical Napatan” mortuary form (Edwards, 2004). 

19 

 
 
 
Meroitic Empire (300 BCE–350 CE) 

In the last centuries of the first millennium BCE, a Kushite state emerged with a new 

center at Meroe between the Sixth and Fifth Cataracts. The beginning of the Meroitic period is 

marked by the royal cemetery relocating from Napata (el-Kurru and Nuri) to Meroe with the 

burial of King Arkamani (Edwards, 2004). This move may represent a symbolic re-focusing of 

power, as some inscriptions allude to dynastic upheaval around this transition (Edwards, 2004). 

While the totality of evidence suggests a degree of cultural continuity between the Napatan and 

Meroitic states, there were significant shifts in cultural traditions, including a greater presence of 

indigenous cults in state religion and the development of Meroitic as the official state language 

(Edwards, 2004). Many previous Napatan settlements were seemingly abandoned while new 

large Meroitic settlements appeared in central Sudan (Edwards, 2004). Around the beginning of 

the Meroitic period, major building projects were carried out at Meroe and dense urban sites with 

Amun temples were constructed nearby (Edwards, 2004).  

A prestige-goods economy is presented as an important component of the Meroitic state 

and its sociopolitical structuring. The distribution of imported goods and their lack of utilitarian 

function reveal that they were reserved for the elite (Edwards, 1996; 2004). Therefore, Meroitic 

trade predicated on sociopolitical hierarchies and motivations rather than commercial needs. This 

prestige-goods economy seemed to operate on a larger scale as royals engaged in long-distance 

elite gift exchange with rulers of other territories in diplomatic contexts, as well as internally 

with local elite in maintaining power relations (Edwards, 1996). Evidence of long-distance elite 

gift exchange is suggested between the Meroitic kings and the Ptolemaic and Roman rulers 

(Edwards, 2004).  

20 

 
The southern and southeastern areas of the Meroitic kingdom are traditionally considered 

peripheries of the state; however, these groups likely played an active role in the Meroitic trade 

system and movement of prestige foreign goods. Meroitic presence has been observed in the 

Gezira region extending to the White and Blue Niles, suggesting some association with the state 

and role in the production of elite goods (e.g., gold) (Edwards, 1996; Lohwasser, 2014). Prestige 

goods from the south traveled to Meroe (e.g., gold, ivory, skins, ostrich feathers) and 

demonstrate trade from this region, perhaps via a retrieval system through Meroitic 

intermediaries in southern Gezira accessing goods with their neighbors to the south (Brass, 

2015). Brass (2015) posits that the mobile pastoralist economy of the southern Gezira, as well as 

the southern Atbai, afforded them the mobility to interact with their borders in trade networks 

between the Meroitic kingdom along the Nile and across the east to the Red Sea. In the Eastern 

Desert, sites such as Tabot have revealed assemblages with Meroitic material culture but did not 

appear under Meroitic rule and were not Kushite sites (Lohwasser, 2014). These sites may 

represent the pathways of the Meroitic trading system with non-Kushite groups. 

Differences between Upper and Lower Nubian settlement patterns and their agricultural 

infrastructure (Edwards, 1999), as well as their visible veneration of political and religious 

figures (Adams, 1974), suggest a lack of institutional uniformity across the Meroitic Kingdom 

despite shared allegiance to the state. The Upper Nubian Western Butana region to the east of the 

Blue Nile is suggested to have been the only area outside of the Meroitic heartlands (i.e., Shendi 

Reach) under direct control and power of the Meroitic royalty over subsistence production. This 

argument is made by Edwards (1999), who suggests that the numerous hafirs associated with the 

settlements in this region signify state control and management of subsistence activities, as their 

construction would entail significant oversight and inscriptions suggest they were state projects.  

21 

 
 
The profound presence of kingship and religious ideologies in Upper Nubia during this 

period is discussed by Adams (1974), who notes the vast majority of Meroitic royal monuments 

(i.e., temples, palaces, and royal burials) were located in the south. Within the southern 

provinces, portrayals of the rulers are numerous and highly associated with Egyptian kingship 

ideology as they depict the rulers connected to the most prominent gods among the southern 

pantheon—Amun and the local deity of Meroe Apedemak, who is not seen further north.  

Contrary to traditional theories, Edwards (1999) argues that settlements in Lower Nubia 

were relatively small, significantly less populated, settled intermittently over the Meroitic phase, 

and designed more for storage than substantial domestic purposes. Previous models on the nature 

of Lower Nubian settlements suggested their primary purpose was agricultural production 

facilitated by saqia (i.e., water wheel) irrigation. However, Edwards cites a reassessment of 

when the saqia wheel could be dated, placing it at the early post-Meroitic instead of the Meroitic 

period. It should be noted though, that at the Meroitic settlement of Umm Muri adjacent to Mis 

Island, numerous fragments of qadus were found possibly indicating the use of saqia irrigation 

during this period (Ahmed, 2014b). Rather than agricultural production, Edwards (1999) 

suggests that the primary purpose of Meroitic settlement in Lower Nubia was to maintain 

communication and state trade between the Meroitic core and Egypt. Compared to Upper Nubia, 

Lower Nubia lacked royal monuments and references to the Meroitic rulers, both in general and 

in relation to deities (Adams, 1974). Adams (1974) suggests that this absence of royal depictions 

and demonstration of state strength indicate that political structuring of the Meroitic north was 

based more on secular institutions than ideologies of divine kingship.  

Mortuary practices during the Meroitic phase are broadly variable. The royal tombs of the 

Meroitic rulers are located near Meroe and display Egyptian mortuary customs with pyramid 

22 

 
 
superstructures and chapels (Adams, 1974). While some elite burials also display these features, 

“lesser” elite burials have been observed as brick mastabas with chapels or niches for offerings 

(Edwards, 2004). Burial forms were not universal and seemed largely reserved for higher status 

individuals, as indicated in non-elite cemeteries in the Meroitic heartland where most burials 

were not visibly marked (Edwards, 2004). Some mortuary forms seemed to have developed from 

Napatan traditions with east-west extended burials in chambers, multiple burials, and grave 

goods associated with socioeconomic status (Edwards, 2004). 

In the transition to the Meroitic phase, the Napatan pottery traditions that were 

substantially influenced by Egyptian features were replaced by new distinctly Meroitic forms 

(Edwards, 2004). The sources of fabrics demonstrate a wide distribution of products and the 

decorations of wheel-made pottery are associated with symbols of the Meroitic state religion. By 

the late Meroitic stage, definite regional variation in ceramic culture is visible with Lower 

Nubian pottery displaying more decoration and assemblages including more imported vessels. 

The disintegration of the Meroitic state likely corresponds to the dissolution of a central 

authority at Meroe and disruptions to the Meroitic economy, rather than traditional theories on 

the Kushite Kingdom’s collapse that present the end as some calamitous natural, military, or 

mass migration event (Edwards, 1996; Welsby, 1996). Some hypothesize that towards the late 

Meroitic phase, Egypt’s shift in trade routes to the Red Sea diminished the need for Kushite trade 

routes (Welsby, 1996). Additionally, as Roman Egypt became poorer and competition from the 

Axumites increased, sea transport may have made trade along the river less desirable (Welsby, 

1996). While the Kushite state faced aggression from desert tribes before its dissolution, their 

lowered economic status in the later Meroitic may have restricted providing the resources 

necessary to ward off attacks (Welsby, 1996). Unrest in the late third century CE from the 

23 

 
 
Blemmyes’ raids on Egypt likely disrupted Meroitic elite access to foreign luxury goods and 

thus, may have affected their prestige economy and central authority’s capital (Edwards, 1996). 

Post-Meroitic Period (350–500 CE) 

Although the circumstances surrounding the Meroitic Empire’s end are unclear, the 

subsequent dissolution of a central authority at Meroe gave way to the replacement of an 

imperial culture with several regional traditions (Edwards, 2004). These post-Meroitic regional 

groups would later emerge as the medieval kingdoms (Edwards, 2004). In the transition from the 

Meroitic to the post-Meroitic period, there are visible signs of discontinuity of the Meroitic state, 

including the abandonment of major monuments, the disappearance of the Meroitic language, 

and an absence of foreign goods in Upper Nubia (Edwards, 2004; Trigger, 1969). Traditional 

models of this transition between periods have been presented as a clear separation in the arrival 

of a new cultural group, particularly considering the replacement of ceramic assemblages and the 

loss of characteristic Meroitic wheel-made pottery (Edwards, 2004; Trigger, 1969). However, 

signs of continuity demonstrate a gradual and direct transition from Meroitic to post-Meroitic 

traditions with a maintained population (Edwards, 2004; el-Tayeb, 2010; Trigger, 1969). 

The Third Cataract seemed to represent a tangible boundary between post-Meroitic 

Lower and Upper Nubian settlement, based on the regions’ distinctive ceramic assemblages 

(Edwards, 1994). The center of post-Meroitic Lower Nubian settlement was focused around the 

south end of the Second Cataract and new settlement extended into areas that were largely 

unoccupied, such as the Batn el-Hajar (Edwards, 2004). The post-Meroitic settlements in Lower 

Nubia were the beginning of the Nobadian kingdom and displayed a distinctive culture with 

Romano-Egyptian influences and imported goods for elites (Edwards, 2004). While cemeteries 

in Lower Nubia represent both Meroitic and post-Meroitic periods, relatively few span the 

24 

 
transition between both (Edwards, 2004). Presence of the Blemmyes at this time in Lower Nubia 

is suggested by handmade pottery associated with the east and Red Sea, as well as examples of 

cemeteries organized in a manner similar to those seen in the Red Sea Hills (Edwards, 2004). 

In Upper Nubia, patterns of post-Meroitic settlement shifted from before as the new 

kingdom of Makuria was emerging (Edwards, 2004). The post-Meroitic period saw a scarcity of 

settlements in the Dongola Reach and minimal evidence of occupation at previous major Kushite 

centers (e.g., Jebel Barkal) (Edwards, 2004). A greater density of post-Meroitic settlement is 

observed downstream around Dongola, which would later become the capital of Makuria 

(Edwards, 2004). The presence of numerous post-Meroitic settlements in the Fourth Cataract 

region, which was void of any permanent settlement for most of the first millennium BC, 

suggests that there was a significant expansion into this area (Edwards, 2004). This post-Meroitic 

territorial extension is partly attributed to a widespread presence of the saqia wheel in the 

beginning of the period, which enabled increased cultivation and settlement, year-round crops, 

more dependable food supplies, and introduction of new crop species (Edwards, 2004). By the 

late Meroitic period at Qasr Ibrim, sorghum appeared in the region and subsequently developed 

durra features (Edwards, 2004). Additional new crops observed included different wheat forms, 

termis beans, peas, bulrush or pearl millet, and sesame (Edwards, 2004). This new technology 

and its effects are largely credited for the population growth in the late post-Meroitic/early 

medieval periods (Edwards, 2004). 

The changes in ceramic assemblages during the post-Meroitic differed between Upper 

and Lower Nubia. Upper Nubian workshops for wheel-made pottery were abandoned with the 

decline of the Meroitic core and by the early post-Meroitic period, black and dark brown 

handmade wares were the most prevalent (el-Tayeb, 2010). Pottery uncovered in northern Nubia 

25 

 
represents a different situation in the transition to the post-Meroitic period, as the end of the 

Meroitic Empire did not seem to significantly affect production. The manufacture of wheel-made 

pottery continued through the post-Meroitic to the end of the Christian period in the north and 

the presence of imported wheel-made pottery in Lower Nubian burials demonstrates their 

continued relationship with Upper Egypt (el-Tayeb, 2010). During the late third and fourth 

centuries, Romano-Egyptian wares were increasingly prevalent in the region; however, by the 

late fourth century, a more distinctly “Nubian” set of materials was developing (Edwards, 2004). 

Mortuary practices shifted in the transition to the post-Meroitic period, during which time 

distinct regional variations are visible (el-Tayeb, 2010). The Meroitic custom of multiple 

individuals interred in the same grave disappeared in the post-Meroitic period, which may 

represent shifting beliefs of the afterlife (Edwards, 2004). Elite burials significantly changed 

from the Meroitic to post-Meroitic periods, as pyramid and mastaba superstructures were 

replaced by massive tumuli (Edwards, 2004). The largest and richest burials from post-Meroitic 

Lower Nubia are located at Qustul and Ballana; however, elite tumuli have also been observed in 

smaller settlements (e.g., Sesibi) (Edwards, 1994). 

Interpretations of the royal burials at Qustul and Ballana have suggested that these post-

Meroitic rulers in the north were Nubian and had some connection to the previous Meroitic state, 

perhaps considering themselves as a continuation of the Meroitic kings (Kirwan, 2002; Trigger, 

1969). In the earliest tombs at Qustul, ties to Meroe are demonstrated in the prevalence of 

Meroitic culture and religious features (Kirwan, 2002). At Ballana, the royal silver crowns are 

nearly identical to those belonging to the Meroitic rulers (Trigger, 1969). Kirwan (2002) posits 

that following the end of Meroe, Nubians already integrated in Meroitic culture became the 

26 

 
 
 
predominant power in the north and displayed themselves as the successors of the Meroitic 

royalty.  

Christianization of Nubia 

Over an extended and dynamic process, Christianity spread across Nubia and was 

adopted as state religion by the medieval period. At the time of Nubia’s Christianization, two 

main denominations were present in the Byzantine Empire—the Melkites and Monophysites 

(Welsby, 2002). These two forms of Christianity were tied to differing political ideologies in 

Egypt, as Monophysites were Egyptian “nationalists” against Byzantine rule and Melkites 

supported Byzantine administration. Both denominations were introduced to Nubia, with 

Nobadia and Makuria seemingly adopting Monophysite and Melkite Christianity, respectively 

(Kirwan, 1984; Welsby, 2002). However, by the mid-to-late seventh century AD all three 

medieval Nubian kingdoms were regarded as Monophysite Christian (Welsby, 2002). 

Across the three kingdoms, the royal families were presumably the primary focus of 

attention during the first Christian envoys (Kirwan, 1984). Religion was highly political within 

the Roman Empire’s operations as a cohesive force to unite people under their influence (Welsby, 

2002). Justinian and the emperors of his time were concerned about imminent Persian invasion 

into Egypt taking a route across the Red Sea and Sudanese Plain, as well as the attraction this 

might hold for Nubians (Kirwan, 2002). Therefore, introducing Byzantine Christianity into the 

Nubian kingdoms was seen as a means of integrating them into the Empire and securing the 

southern Egyptian border (Kirwan, 2002). However, these local rulers undoubtedly had their own 

motivations for adopting Christianity in their kingdoms and Christianization should not be 

viewed as an inevitable, passive event. Christianity significantly affected traditional 

27 

 
 
sociopolitical institutions by essentially ending ideographic worship of the divine monarch 

(Żurawski, 2006) and the production of material goods associated with such veneration.  

By incorporating the archaeological evidence from this Christianization process, the 

complexity and variation of Christianity’s impact on Nubian societies is recognizable (Edwards, 

2001). Archaeological indicators of the spread of Christianity include Christian symbolism on 

artifacts, production of artifacts associated with Christianity (e.g., ceramic assemblages), 

construction of churches, and Christian funerary practices (Edwards, 2001; Welsby, 2002). While 

the appearance of churches reflects more official and state-driven acts, burials displaying 

Christian mortuary customs are more indicative of personal beliefs and the adoption of 

Christianity by individuals (Welsby, 2002). Much of the scholarship on the Christianization of 

Nubia has been conducted through a lens of state worship and institutions which lacks 

recognition of private worship. 

With the spread and adoption of Christianity as state religion, mortuary practices became 

more uniform across Nubia as they followed a unified ideological system (Welsby, 2002; 

Żurawski, 2006). This new religion offered a different set of eschatological views on death and 

the afterlife which required new practices and rituals surrounding funerary contexts (Żurawski, 

2006). Over the adoption of Christianity, local funerary traditions gave way to a distinctive 

mortuary repertoire that embodied the Christian religious structure (Żurawski, 2006). 

The typical Christian Nubian burial was predominantly oriented in an east-west direction, 

with the head on the western end and body in an extended position (Adams 1998; Welsby, 2002; 

Żurawski, 2006). Following Christian eschatological views on the transmigration of the soul 

rather than of a corporeal afterlife, Christian Nubians included minimal grave goods in burials, 

marking a profound divergence in funerary tradition (Adams, 1998; Welsby, 2002; Żurawski, 

28 

 
2006). Additionally, whereas pre-Christian mortuary rituals for elite individuals were intimately 

associated with striking monuments and grave goods, Christian ideologies deterred such 

practices as all people became equal at death and the rewards of the afterlife were possible for 

those who adhered to the Christian faith (Welsby, 2002; Żurawski, 2006). Visibly high-status 

tombs were therefore abandoned, although some burials within churches suggest interments for 

individuals of canonical status (Żurawski, 2006). 

Although the belief in corporeal preservation of the deceased ended with Christianity, 

many burials included the mortuary practice of protecting the body and guarding it from direct 

exposure to the earth with slabs and bricks (Adams, 1998; Welsby, 2002; Żurawski, 2006). This 

practice may have connections to pre-Christian Nubian and Egyptian traditions and has been 

interpreted as a superstitious measure to protect the face and prevent evil spirits entering the 

body (Welsby, 2002). Shrouding of the decedents seems to be one of the most prevalent practices 

among Christian burials across Nubia during the entire medieval period (Adams 1998). 

The use of superstructures in Nubian mortuary behavior has a deep history and while the 

presence of superstructures continued through Christianization, they took notably different forms 

from previous traditions (Adams, 1998). Over the course of Christianization, the tumulus 

superstructure was largely abandoned in areas where churches or monasteries were implemented; 

however, outside of these centers, the tumulus burial is observable for longer (Żurawski, 2006). 

Christian superstructures were predominantly two general types—mastabas and flat pavements 

(Adams, 1998; Welsby, 2002; Żurawski, 2006). The new manifestation of superstructures varied 

considerably in form and became a common aspect of Christian Nubian burials, particularly 

during the Early and Classic Medieval periods (Adams, 1998; Welsby, 2002; Żurawski, 2006). 

The cross as a religious symbol appeared in the medieval Nubian mortuary complex with 

29 

 
Christianization and was a principal component of funerary behavior (Welsby, 2002). In addition 

to cross-shaped superstructures, crosses were incorporated into mortuary customs as cross-

shaped stelae and crosses made of terracotta or wood installed on top of tombs. 

The arrival and subsequent adoption of Christianity in Nubia was a gradual process that 

began during the post-Meroitic period and continued through the medieval period in variable 

ways across Nubia. During this time, Christian cemeteries began where pre-Christian burials had 

ended, demonstrating the complexity of continuity and change over this process (Żurawski, 

2006). This transition is also marked by burials that display a combination of pre-Christian and 

Christian mortuary features aligned with a transition in ideological views (Welsby, 2002). The 

replacement of pre-Christian temples with churches also reveals an extended process, as some 

were modified or built before the “official conversion”, as well as centuries afterwards (Welsby, 

2002). This evidence demonstrating the inconsistency within the Christianization process 

prompts a reconsideration of presenting medieval Nubia as “Christian Nubia” (Edwards, 2001). 

Medieval Period (500–1500 CE) 

The beginning of the medieval period in Nubia is traditionally marked by the “official 

conversion” and adoption of Christianity as state religion by the three Nubian kingdoms—

Nobadia, Makuria, and Alodia (Alwa). The relationship between these Nubian polities and the 

Byzantine Empire were harmonious into the beginning of the seventh century (Welsby, 2002). 

However, Byzantine control in Egypt was interrupted by the Persian Empire in 619 CE (Welsby, 

2002). As the Byzantine and Persian empires competed against each other, another power 

emerged in this struggle—the Arab Muslim Empire. In a vast campaign, Arab Muslim forces 

defeated the Byzantines and Persians around 673 CE and continued into Egypt (Welsby, 2002). 

30 

 
 
 
After taking Egypt, the Arab Muslim forces traveled west and south, taking over the Byzantine 

powers in Libya, Tunisia, and Algeria; however, were stopped at Nubia.  

In 652 CE, a campaign into Nubia led by Abdallah b. Sa’id Abi Sahr advanced to 

Dongola and laid siege to the capital (Welsby, 2002). After fighting ensued between Muslim and 

Nubian forces, historical accounts describing what followed differ depending on the source’s 

sympathies or affiliation (Spaulding, 1995; Welsby, 2002). However, they generally depict the 

Makurian king seeking peace with the Muslim Egyptian forces and both parties striking an 

agreement referred to as the baqt by external sources. (Spaulding, 1995; Welsby, 2002). In these 

baqt terms, both parties were allowed freedom of movement in the other country, although 

residency was restricted (Spaulding, 1995). Additionally, Nubia would send enslaved peoples to 

Egypt and repatriate Arab prisoners, and Egypt would give Nubia foodstuff and likely additional 

elite goods (Spaulding, 1995). For medieval Makuria, Spaulding (1995) suggests that the baqt 

was an expression of the Nubian royal gift exchange that extended into their past forms of 

diplomacy. From this perspective, the agreement was therefore an acknowledgement of the new 

Islamic rule in Egypt with which medieval Nubia would interact. 

The Medieval Kingdoms 

Located between the First and Third Cataracts, the medieval Kingdom of Nobadia grew 

from post-Meroitic settlement and established a capital at Faras (Edwards, 2014). In northern 

Nubia, numerous walled settlements date to early medieval construction and may represent an 

unstable northern frontier for Nobadia around this time (Edwards, 2014). Another early medieval 

settlement pattern is a scattering of small sites, particularly in the Batn el-Hajar and Third 

Cataract regions. These settlements may represent farming communities associated with the post-

Meroitic establishment of unoccupied areas facilitated by saqia irrigation (Edwards, 2014). A 

31 

 
shift in late medieval Nobadian settlement is seen with the abandonment of settlements in the 

Second Cataract region and sudden increase in occupation of the Batn el-Hajar and Third 

Cataract regions (Edwards, 2014). This regional abandonment may have been instigated by 

threats of Egyptian aggression in the north (Adams, 2001). 

The Kingdom of Makuria was located from the Third Cataract to around the Fifth 

Cataract with a capital at Dongola (Adams, 1991). The first figures of authority in Makuria’s 

early stages of formation may have been interred in the burial grounds at Tanqasi and ez-Zuma, 

located downstream of the Fourth Cataract (Godlewski, 2014). These burials demonstrate a 

significant departure in social differentiation with multi-chamber tombs, complex grave shafts, 

and large mounds. After a substantial population increase during the fourth and fifth centuries, 

the appearance of fortified sites in the fifth and sixth centuries indicates an urbanization process 

and perhaps economic organization by the early ruling elites (Godlewski, 2014). A series of 

defensive fortresses occupied the Fourth Cataract region, the largest of which were built at the 

beginning of the seventh century and then abandoned after a century, perhaps from external 

threats and/or combining of regional powers (Żurawski, 2014).  

In the beginning of the medieval period, Nobadia and Makuria had an adverse 

relationship, perhaps related to the rival Christian denominations they had adopted (Adams, 

1991). However, at some point the two kingdoms united under the single Makurian king (Adams, 

1991). While subservient to the kings in Makuria, Nobadia had a regional administrative 

figure—the Eparch of Nobadia. The process and timing surrounding this union is unclear, but the 

alliance was likely in place by the baqt agreement and endured over six centuries. Although both 

kingdoms were joined under the king at Dongola, cultural, economic, and political differences 

32 

 
 
existed between them, suggesting they did not become a uniform entity after merging (Adams, 

1991).  

In the mid-eleventh century, Makuria entered into an alliance with Alodia and the 

Kingdom of Dotawo was established as a new form of Makuria (Godlewski, 2014). The 

Kingdom of Dotawo was seemingly synonymous with Makuria; however, in the thirteenth 

century, kings of Makuria and kings of Dotawo existed (Adams, 1991). Adams suggests that the 

kings of Dotawo may have lived near Qasr Ibrim at this later date and acted as authoritative 

figures in Nobadia, superior to the Eparch (Adams, 1991). While the Eparch apparently dealt 

with trade and relations with Egypt, the king of Dotawo seemed to operate within the civil and 

legal sphere (Adams, 1991). The Kingdom of Dotawo brought the establishment of Old Nubian 

as the predominant language in the church, state administration, and community while traditional 

Greek and Coptic were used in monastic groups (Godlewski, 2014). 

The Kingdom of Alodia (Alwa) is described as the most wealthy and powerful of the 

medieval Nubian kingdoms in historical accounts (Welsby, 2014). As the Kushite power center 

Meroe was diminishing, massive tumuli with elite grave goods appeared at el-Hobagi (~70km 

upstream) which likely represents the early Alodian power base (Welsby, 2014). The majority of 

Alodian material culture seemingly continued from earlier post-Meroitic traditions (Edwards, 

2004). By the sixth century, the new settlement of Soba was occupied and may have already 

served as Alodia’s capital (Welsby, 2014). Soba was a large urban center and a religious and 

political focus, containing at least seventeen churches with two large churches located in the city 

center adjacent to a palatial structure (Edwards, 2004; Welsby, 2014). Many fortifications have 

been observed downstream and may signify a larger defense system. The kingdom’s borders 

were likely fluid and dependent upon their interactions with neighboring groups and the high 

33 

 
 
mobility of peripheral groups (Welsby, 2014). However, the terms in the baqt agreement suggest 

that Alodia and Makuria were contiguous. Along the Atbara, the Dihiyyun were Alodian subjects 

and may represent a union between Alodia and Beja. 

Medieval Nubian economy 

Given the lack of seluka land naturally available for cultivation, the spread of saqia 

irrigation resulted in an agricultural revolution, in which arable land significantly increased 

(Welsby, 2002). Many of the crops grown in the region today were cultivated during the 

medieval period (Welsby, 2002). The staple grain in medieval Nubia was dhurra/sorghum. 

Dhurra was used to make beer (mizr) and grains including dhurra and wheat were used to make 

flat unleavened bread called kirsa. Other crops included millet, wheat, dates, olives, palm trees, 

bananas, citrus, and cotton (Welsby, 2002). Crops that had been previously cultivated in Egypt 

appeared and the introduction of new crop species likely made year-round production possible 

(Welsby, 2002).  

The later sixth century saw a new type of pottery, some of which displayed clear 

Christian symbolism (Edwards, 2004). These “Early Christian” wares represented a series of 

“Christian” pottery that would last for almost 1,000 years (Edwards, 2004). Two major 

production centers for “Early Christian” wares were located at Faras and Dongola, as well as a 

third production center for the distinctive “Soba ware” (Edwards, 2004). Compared to the pottery 

manufactured at Old Dongola, wares from Soba primarily stayed within the capital’s reaches 

rather than exchanged over long distances (Welsby, 2014). In the ninth century, new “Classic 

Christian” styles appeared (Edwards, 2004). Production at Faras was abandoned in the late 

tenth/eleventh centuries and the Nobadian pottery demands were increasingly met by Egyptian 

imports (Edwards, 2004). A significant rise in the amount of imported pottery from Aswan and 

34 

 
 
 
Lower Egypt is observed at Meinarti during this time (Adams, 2001). At Dongola, “Post-Classic” 

wares appeared in the eleventh century but are not observed far beyond this region (Edwards, 

2004). In the late medieval period, a highly variable “Terminal Christian” collection of wares 

appeared and is suggestive of more local production (Edwards, 2004). At the end of the medieval 

period, wheel-made pottery disappeared and was replaced by a range of handmade wares. 

The medieval Nubian kingdoms engaged in trade networks with Egypt and eastern 

polities over the medieval period. In the beginning of the medieval period, Makuria enjoyed a 

trade partnership with Byzantine Egypt, as suggested by the extensive presence of amphorae 

from the end of the sixth century (Godlewski, 2014). In turn, Makuria provided Byzantine Egypt 

with the goods they had historically presented to Egypt. After the baqt agreement, evidence 

suggests that Nubian and Muslim royalty engaged in diplomatic exchange of elite goods, in 

which Nubia imported luxury foods, vinegar, wine, horses, and textiles and provided Muslim 

princes with ivory, animal skins, dates, ebony, emery, alum, and exotic live animals (Spaulding, 

1995). While Nubian economic dealings were tightly controlled by the elite, Muslim merchants 

seeking to trade in medieval Nubia after the baqt created avenues for more private commerce 

that decreased the exclusivity of foreign items (Spaulding, 1995).  

The northern zone of Nobadia became open to free trade as Muslim merchants could 

establish their commerce and even own land (Adams, 1991). The coexistence of Christian 

Nubians and a Muslim minority is demonstrated by the presence of Muslim tombstones from the 

tenth and eleventh centuries in Wadi Halfa’s predominantly Christian cemeteries (Shinnie and 

Shinnie, 1965). Egyptian coinage discovered in Lower Nubia suggests its circulation in the 

Nobadian market and historical documents indicate that the Nubian cash economy was directly 

connected to Egypt (Adams, 1991; Ruffini, 2012). Acting as a kind of viceroy, the Eparch of 

35 

 
 
Nobadia was primarily responsible for overseeing trade and foreign relations with Nubia’s 

Muslim neighbors, as well as upholding the terms of the baqt (Adams, 1991). Documents from 

Qasr Ibrim discuss how independent Muslim agents and the Eparch acted as trade intermediaries 

for the kings of Makuria and Alodia (Welsby, 2002).  

At Soba East, a large Muslim community lived within the city and likely included a 

group of merchants (Welsby, 2002). Given Makuria’s restrictions on the presence of Muslim 

traders, this may indicate a direct connection between Alodia and Egypt and/or Red Sea ports 

(Welsby, 2002). Texts describe routes between Alodia and the Red Sea and a relatively small 

amount of Islamic imports have been uncovered at Soba including rare “Islamic” glazed wares 

and fine glassware, as well as Chinese porcelain sherds from the twelfth and thirteenth centuries 

(Edwards, 2004; Welsby, 2002). 

Collapse of the medieval Nubian Kingdoms 

Beginning in the thirteenth century, a series of raids and campaigns was carried out 

between Makuria and Muslim Egypt (Welsby, 2002). After a succession of dynastic upheavals, 

the first Muslim Nubian ruler ascended to the Makurian throne in the early fourteenth century 

(Welsby, 2002). During his reign, the first evidence of a mosque at Dongola appeared and 

Egypt’s imposed jizya ended. The Makurian centralized government weakened and perhaps 

spurred by aggression from the desert Arabs (likely the Beja), the royal court presumably 

departed Dongola and relocated to Nobadia in the mid-fourteenth century (Godlewski, 2014; 

Welsby, 2002). Evidence of increasing insecurity within the Makurian kingdom is represented by 

the presence castle-houses between Qasr Ibrim and the Third Cataract (1100–1500 CE), 

fortification of structures, appearance of fortresses, and relocation of settlements into more 

defensible and desolate areas (Welsby, 2002). The final manifestation of this polity, described as 

36 

 
 
a continuation of the Dotawo Kingdom, was likely focused at Daw with one final bishopric at 

Qasr Ibrim and lasted until the end of the fifteenth century (Edwards, 2004; Godlewski, 2014). 

Little is known regarding the decline and end of the Kingdom of Alodia. Some burials 

that possibly date to around its demise demonstrate neither Christian nor Muslim funerary rites 

and may represent decreased centralization (Welsby, 2014). In the late medieval period, an 

independent Kingdom of el-Abwab in the northern province of Alodia is described (Welsby, 

2002) and may represent a fractured centralized Alodian authority. Alodia eventually fell to the 

Funj (Abdallab Arabs) from the south (Welsby, 2002). Accounts describe how Soba was already 

in ruins when the Funj entered Alodia and established their kingdom at Sinnar in 1504 CE 

(Edwards, 2004; Welsby, 2002). 

37 

 
 
 
 
CHAPTER THREE: PATHOPHYSIOLOGY OF MALARIA, PALEOPATHOLOGY OF 

ANEMIA, & SKELETAL LESIONS OF INTEREST 

The first section in this chapter discusses the pathophysiology of malaria as it relates to 

possible effects on the skeleton, including the anemia involved in the disease. The 

paleopathology of anemia is then discussed, and a historical overview is provided regarding 

studying anemia in skeletal remains. The last section presents previous findings on the skeletal 

lesions studied in this dissertation, including their association with age of the individual and co-

occurrence. Additionally, a review of femoral cribra is provided which discusses the proposed 

etiologies associated with the lesion and observations on variation in the lesion’s appearance. 

Pathophysiology of malaria 

Malaria induces anemia in infected individuals through hemolytic anemia from extensive 

destruction of parasitized and non-parasitized red blood cells (White, 2018). Hemolytic anemia, 

which can be inherited or acquired, has been proposed as one of the possible anemias responsible 

for producing skeletal lesions, as it initiates increased erythropoiesis that leads to marrow 

hyperplasia (Walker et al., 2009). However, the hemolytic anemia of malaria is also accompanied 

by suppressed erythropoiesis and dyserythropoiesis (Lamikanra et al., 2007; White, 2018), rather 

than increased erythropoiesis. This inability to produce an adequate erythroid response may be 

due to the production of mediators that inhibit erythropoiesis, including proinflammatory 

cytokines, and perhaps the presence of hemozoin, a parasite byproduct of hemoglobin digestion 

(Lamikanra et al., 2007; White, 2018). In acute malaria, erythropoiesis is perturbed and reduced 

and the bone marrow “may be hypocellular, normocellular, or mildly hypercellular” (Bain et al., 

2019:133; Wickramasinghe and Abdalla, 2000). In chronic malaria, however, total erythropoietic 

activity rises with increased marrow cellularity and erythroid hyperplasia (Wickramasinghe and 

38 

 
 
Abdalla, 2000). This suggests that chronic infection and severe anemia from malaria may be 

more likely to produce visible skeletal lesions than acute malarial infections.  

In addition to anemia, malarial infection seems to cause an imbalance in osteoclast and 

osteoblast activity resulting in heightened bone resorption. This resorptive effect from malarial 

infection may be due to a surge in free heme (Moreau et al., 2012) and parasitic sequestering in 

the extravascular space of bone marrow (Farfour et al., 2012; Joice et al., 2014; Obaldia et al., 

2018). In chronic cases, the prolonged buildup of parasitic products in marrow is suggested to 

ultimately result in increased osteoclast production and bone resorption (Lee et al., 2017). In two 

experimental studies involving mice infected with malaria, resorptive microarchitectural changes 

included reduced trabecular bone volume, increased trabecular spacing, and decreased trabecular 

numbers compared to control subjects (Lee et al., 2017; Moreau et al., 2012).  

Malarial infection has a synergistic relationship with other pathologies which can result 

in comorbidities as host immune responses are lowered. In communities experiencing moderate 

to high transmission of malaria, conditions observed to manifest concurrently include bacterial, 

viral, and helminth infections, and malnutrition such as vitamins A and B12 deficiencies (White, 

2018). Additionally, the prevalence of haemoglobinopathies that offer some protection against 

malaria, but also contribute to anemia, often parallel the incidence of malaria (White, 2018). 

Therefore, as these multiple conditions can affect individuals at risk or experiencing malarial 

infection, the resulting anemia can be “multifactorial” (White, 2018:7). In addition to this 

influence on morbidity, the combination of malarial infection with other infections or conditions 

can have profound effects on mortality (Carter and Mendis, 2002). 

39 

 
 
 
Paleopathology of anemia 

In response to anemia, skeletal changes are considered to primarily result from the need 

to expand hematopoietic marrow to promote red blood cell production (Brickley, 2018). Due to a 

lack of healthy red blood cells, red bone marrow becomes overreactive to compensate for the 

disruption to hematopoiesis caused by blood loss, inhibited erythropoiesis, and/or increased 

hemolysis (Rivera and Lahr, 2017). This stress on the marrow tissue leads to expansion of 

hematopoietic centers (i.e., marrow hyperplasia), which in turn places pressure on the 

surrounding bone and initiates osteoblastic and osteoclastic activity (Rivera and Lahr, 2017). The 

resulting changes include expansion of the marrow space and openings from the marrow cavity 

to the external cortical surface (Wapler et al. 2004). Hyperplasia is also associated with 

interrupted remodeling of the external cortical bone that can produce porous lesions (Rivera and 

Lahr, 2017). Porous lesions of the cranium—porotic hyperostosis of the vault and cribra orbitalia 

of the orbital roof—have long been linked to anemia among possible etiologies. The types of 

anemia that could potentially be responsible for causing the lesions have been a source of debate, 

affecting interpretation of the lesions’ presence in human skeletal remains. 

Early paleopathological inquiry into the skeletal manifestation of anemia and interpreting 

its presence in archaeological remains is marked by J. Lawrence Angel’s work. Among his 

publications on the subject is his 1966 article that first applied the term porotic hyperostosis to 

describe cranial porosities while discussing these lesions observed from prehistoric Eastern 

Mediterranean sites. Based on a history of endemic malaria, Angel links porotic hyperostosis to 

the hemolytic anemia involved in genetic conditions conferring some resistance to malaria—

specifically, thalassemia and sickle cell anemia. However, he notes that thalassemia or sickle cell 

are unlikely causes of porotic hyperostosis in areas that lacked falciparum malaria. Subsequent 

40 

 
researchers focused on the cause of these lesions in such regions where the hypothesis of 

thalassemia or sickle cell anemia did not fit due to relative low levels of these genetic variants in 

endemic populations (Hengen, 1971) and the presence of these cranial lesions in areas where 

malaria likely did not exist at the time (El-Najjar et al., 1976). 

Following Angel’s work, the iron deficiency anemia hypothesis was a primary etiological 

model and was attributed to various sources. Hengen (1971) suggests that the cribra orbitalia 

observed in crania from different time periods and geographic locations is most likely linked to 

iron deficiency anemia arising from insufficient iron intake, inadequate sanitation, and parasitic 

infections. El-Najjar and colleagues (1976) interpret patterns of porotic hyperostosis and cribra 

orbitalia prevalence from prehistoric and historic North American Southwest to iron deficiency 

anemia due to hypothesized differences in dietary iron. Stuart-Macadam (1992) later argues that 

iron deficiency anemia is not strongly associated with nutrition and diet. Instead, iron-deficiency 

anemia is presented as the result of an adaptive mechanism as the body reserves iron in response 

to pathogens. In this process, a chronic or high pathogen load can trigger a response by the 

immune system in which iron is relocated and stored in the liver in defense of invading 

pathogens whose growth is facilitated by iron. Therefore, the lack of iron in blood strengthens 

the host’s defense against microbial invaders. Holland and O’Brien (1997) contend that this 

parasite model is too simplistic in its application of evolutionary processes that focus on short-

term outcomes and ignore long-term costs. The authors suggest that this response is more of a 

maladaptation rather than successful adaptation and conclude that porotic hyperostosis and cribra 

orbitalia are likely the result of a combination of synergistic and interacting factors. 

Among several arguments against the iron deficiency hypothesis, the most noted criticism 

is articulated by Walker and colleagues (2009). In their paper, Walker et al. insist that iron-

41 

 
deficiency anemia cannot cause marrow hypertrophy in these lesions because it inhibits red 

blood cell production, which requires iron to perpetuate. These researchers instead contend that 

megaloblastic anemia and hemolytic anemia are the likely etiologies of the cranial lesions 

because they result in premature red blood cell destruction, initiating a compensatory response 

from the body to produce more red blood cells. Walker and colleagues discuss how porotic 

hyperostosis and many cases of cribra orbitalia may likely be caused by megaloblastic anemia 

acquired by infants during nursing. This is due to low maternal vitamin B12 levels and unsanitary 

living conditions that contribute to additional nutrient deficiencies and gastrointestinal infections. 

While marrow hypertrophy from anemia is a likely common cause of cribra orbitalia, Walker et 

al. suggest another possible etiology—subperiosteal hematomas, which are associated with 

multiple diseases (e.g., scurvy). 

This position countering iron deficiency anemia was in turn contradicted by Oxenham 

and Cavill (2010), who claim that Walker et al.’s (2009) assertions on the relationship between 

iron deficiency anemia and erythropoiesis are not supported by the clinical literature. In their 

discussion, Oxenham and Cavill (2010) contend that increased erythropoietic activity can 

increase from iron deficiency anemia (resulting from blood loss and/or a low-iron diet). The 

authors explain that in iron deficiency anemia, there is a lack of iron to supply the stimulated red 

marrow and there is insufficient iron for erythroblasts. A large portion of these cells is then 

unable to make enough hemoglobin in time to become mature red blood cells, and are thus 

destroyed in the marrow. Following this is a significant increase in erythropoiesis, albeit an 

ineffective response. The authors conclude that hemolytic, megaloblastic, and iron deficiency 

anemias are all possible etiologies for porotic hyperostosis and cribra orbitalia. 

42 

 
In further response to the dismissal of iron deficiency as a potential etiology, McIlvaine 

(2015) argues that excluding it altogether from paleopathological interpretations relating to 

porotic hyperostosis could be premature. McIlvaine’s concern is that if iron deficiency does 

restrict erythropoiesis, hidden heterogeneity may result in skeletal samples. Vitamin B12 

deficiency in megaloblastic anemia and iron deficiency anemia often co-occur given their shared 

causes. When these two deficiencies occur at the same time, iron deficiency may restrict marrow 

hypertrophy and thus prevent the expression of porotic hyperostosis and cribra orbitalia, despite 

the marrow being stimulated by B12 deficiency. Therefore, McIlvain argues that in excluding the 

iron deficiency anemia hypothesis, skeletal populations may demonstrate hidden heterogeneity 

as iron deficiency could mask the expression of vitamin B12 deficiency. In such cases, individuals 

who have multiple nutritional deficiencies and are thus the most nutritionally stressed would 

resemble the non-stressed individuals because neither sub-group would display evidence of 

porotic hyperostosis. Ultimately, McIlvaine supports the inclusion of multiple indicators of stress 

and a biocultural framework in bioarchaeological interpretations. 

The following predominant model for interpreting anemia from the skeleton has centered 

on hemolytic and megaloblastic anemias and iron deficiency anemia, the latter often being noted 

as disputed. An observable absence of marrow expansion underlying these lesions has been 

interpreted as lack of support for a diagnosis of anemia, such as Wapler and colleagues’ (2004) 

histological examination of cribra orbitalia lesions in a group of remains from ancient Nubia. 

Although, based on recent research, Rivera and Mirazón Lahr (2017) suggest other forms of 

anemia as sources of cribra orbitalia when the lesions demonstrate a lack of bone marrow 

expansion, including anemia of chronic disorder or disease that result in marrow hypoplasia 

rather than hyperplasia. 

43 

 
The causes behind these lesions, and the associated implications regarding their 

interpretation, is an area of contestation and further research. Most recently, Brickley (2018) 

argues for the implementation of a biological approach when analyzing lesions associated with 

anemia. This approach entails drawing on clinical literature and examining the relationship 

between the location of lesions across the skeleton and the age of the individual. Porous lesions 

resulting from anemia are thought to manifest only in regions of red or mixed marrow with 

potential for hyperplastic responses. As the distribution of hematopoietic marrow changes over 

time, lesions are therefore age-dependent, and age should be considered when forming possible 

diagnoses. 

Lesions of interest: association with age and co-occurrence 

Previous studies examining the relationship between the lesions of focus in this 

dissertation and age of the individual have found patterns regarding the cribrous lesions. In a 

modern skeletal sample from Uganda, where malaria is holoendemic, Smith-Guzmán (2015a) 

reports an association between age-at-death and humeral and femoral cribra with higher 

prevalences of the lesions in younger individuals. Additionally, while humeral and femoral cribra 

were significantly more frequent in younger individuals, Smith-Guzmán (2015a) notes that cribra 

orbitalia was observed across age groups. Schats (2021) also reports a significantly higher 

prevalence of humeral and femoral cribra, as well as cribra orbitalia, in subadults compared to 

adults in a skeletal sample from the medieval/early modern Netherlands (800–1600 CE). 

Furthermore, Schats (2021) notes a common pattern in the prevalence of cribra orbitalia, humeral 

cribra, and femoral cribra across age groups with the lesions’ frequencies at their highest in the 

child age group (4–12 years) and decreasing with age. While adult individuals demonstrated a 

decrease in the chances of displaying humeral and femoral cribra with increasing age, Schats did 

44 

 
 
 
not find a similar trend for cribra orbitalia. In assessing the hypothesized activity of the three 

cribrous lesions in subadult and adult remains from late medieval archaeological sites in NW 

Spain, Mangas-Carrasco and López-Costas (2021) observe more signs of healing from the 

lesion’s surface with increasing age for cribra orbitalia, humeral cribra, and femoral cribra. 

The co-occurrence between cribra orbitalia, humeral cribra, and femoral cribra has 

presented in varying patterns for all three lesions, although a relatively strong connection 

between humeral and femoral cribra has been observed. An association between cribra orbitalia, 

humeral cribra, and femoral cribra was described by Miquel-Feucht and colleagues (1999) as 

part of a “cribrous syndrome” upon noting similar external and internal appearances of the 

postcranial cribrous lesions compared to cribra orbitalia. In addition to observing cortical 

reduction and trabecular expansion in the lesions, the authors describe a correlation between the 

three lesions in a sample of subadult skeletal remains from Spain. Djuric and colleagues (2008) 

also describe a co-occurrence of the three lesions and report a frequency of their combined 

expression in 33.33% (5/15) of subadult (0–14 years) remains studied. Among the lesions 

investigated in a malaria-endemic skeletal sample, Smith-Guzmán (2015a) reports a significant 

association between humeral cribra and femoral cribra; however, cribra orbitalia was not 

correlated with the cribrous postcranial lesions in the study sample. Smith-Guzmán therefore 

suggests the potential for varying processes behind the expression of cribra orbitalia in the 

population. In examining the association between pairs of the three cribrous lesions, Schats 

(2021) also observes a significant association between humeral cribra and femoral cribra in 

subadult individuals. While more frequent in subadults, Schats reports that the co-occurrence of 

all three lesions was uncommon in the study sample. Schats concludes that the findings do not 

45 

 
 
support a “cribrous syndrome”; however, a relatively small subadult sample size may not have 

captured the potential for this lesion pattern. 

Femoral cribra 

The earliest cited source describing femoral cribra is often attributed to Harrison Allen, 

who stated that “the neck is marked in front near the articular surface by a faint depression, 

which is often cribriform in appearance and may receive the name of the cervical fossa” 

(1882:190). Since then, such porous features observed in this area of the femoral neck have been 

referred to as various iterations of “cervical fossa of Allen” (Angel 1964; Finnegan, 1978; Kate, 

1963; Meyer, 1924, 1934), but is also discussed under the names Empreinte (Odgers, 1931; 

Schofield, 1959), anterior cervical imprint (Kotsick, 1963), cribra femoris (Blondiaux et al., 

2015; Molleson et al., 1998), and femoral cribra (Djuric et al., 2008; Djuric et al., 2010). Adding 

to this inconsistency are conflicting definitions and studies that conflate femoral cribra with other 

features, such as Poirier’s facet (e.g., Odgers, 1931), thus making it difficult to distinguish the 

findings between discrete features. The dynamic nature of the femoral neck with its propensity to 

display several features of interest led J. Lawrence Angel to refer to its anterior surface as the 

“reaction area,” which “in its most dramatic form is an eroded fossa, the cervical fossa of Allen” 

(1964:130). This lesion has been categorized as a postcranial nonmetric trait (Finnegan and 

Faust, 1974; Finnegan, 1978), but is among a subset of nonmetric traits whose “origin and 

development is either unknown or debated” (Mann et al., 2016:xii). 

Furthermore, porous lesions observed at the same area of the femoral neck have been 

presented as disparate lesions based on variations in their appearance and labeled either Allen’s 

fossa or femoral cribra (Göhring, 2021). Göhring (2021) describes Allen’s fossa as a demarcated 

concave area of exposed trabeculae that is a nonmetric trait created by friction of the zona 

46 

 
orbicularis, and femoral cribra as a more diffuse area of convex trabecular expansion that is 

considered pathological in nature. Others offer a more encompassing definition for femoral 

cribra that attributes these variations in appearance to different expressions of the lesion (Radi et 

al., 2013). In the current research, the lesion of interest is referred to as femoral cribra and 

defined as an area of cortical disruption or porosity perforating the cortical surface on the 

anterior-inferior femoral neck (Miquel-Feucht et al. 1999; Radi et al., 2013) with varying 

expressions that are investigated to better understand possible underlying processes. 

Proposed etiologies 

Various etiologies have been proposed for femoral cribra and its related monikers that are 

attributed to pathological origins or associated with biomechanical stress, either from general 

contact with soft-tissue structures or specific activities. These different etiological categories 

subsequently affect the interpretation of this lesion’s presence in skeletal remains. The non-

pathological explanations suggested to produce femoral cribra include prolonged acetabular 

contact while sleeping (Meyer, 1924, 1934), hyperflexion from a squatting posture (Schofield, 

1959), contact with the rectus femoris tendon associated with an erect bipedal posture (Kate, 

1963), and tightening of the zona orbicularis of the hip joint capsule (Odgers, 1931) which is 

included in recent discussions of porous lesions at the femoral neck. Göhring (2021) presents 

Allen’s fossa as a nonmetric trait separate from femoral cribra that is caused by biomechanical 

factors including rubbing of the zona orbicularis on the femoral neck during activity. Göhring 

attributes the lesion femoral cribra, however, as indicative of pathology. Other recent 

examinations of porous lesions at the femoral neck cast doubt on interpretations of femoral cribra 

as a specific activity-related marker (Radi et al., 2013), and propose that the lack of anatomical 

47 

 
insertion at this site renders propositions of biomechanical alteration or injury as dubious 

(Miquel-Feucht et al., 1999). 

Current etiological understandings of femoral cribra describe it as a pathological stress 

marker. Femoral cribra has been likened to cribra orbitalia in morphology and is occasionally 

grouped together with both cribra orbitalia and humeral cribra in an assemblage of lesions 

referred to as “cribrous syndrome” (Miquel-Feucht et al., 1999). The presence of femoral cribra, 

in conjunction with other cribrous lesions, has been associated with hypothesized etiologies 

including magnesium deficiency from malnutrition and/or malabsorption (Djuric et al., 2008; 

Miquel-Feucht et al., 1999), tubercular granuloma within bone marrow (Blondiaux et al., 2015), 

anemia from pathogenic stress (Djuric et al., 2008; Djuric et al., 2010; Smith-Guzmán, 2015a), 

inherited anemia (Mendiela et al., 2014), and anemia from some combination of nutritional, 

infectious, or genetic origin (Paredes et al., 2015).  

Previous microscopic observations of femoral cribra support anemia as a possible 

etiology. In a qualitative histological analysis on twelve subadult cases of femoral cribra, Djuric 

and colleagues observed “thinning and porosity of the cortical bone, enlargement of 

intertrabecular space, and growth of the trabeculae perpendicular to the bone surface” 

(2008:469). Based on their two-dimensional assessment, Djuric and colleagues (2008) propose 

that these characteristics could be due to marrow expansion and are thus consistent with an 

etiology of chronic anemia. 

Variation in morphology 

Variation in femoral cribra’s appearance has been observed in previous examinations of 

the lesion within skeletal samples. In these discussions, proposed classifications of this variation 

are based in terms of general size, location, and “severity” of the lesion. Miquel-Feucht and 

48 

 
colleagues (1999) classify femoral cribra into three categories based on the location and extent of 

the lesion distributed over the femoral neck’s anterior and inferior surface. Other classificatory 

systems describe varying degrees of femoral cribra’s expression that allude to “severity” or 

“intensity”. Odgers’s categories of the lesion include “slight” or “marked erosions” (1931:355). 

Kotsick divides the range of variation into two types—a demarcated “ulcer-like excavation” or a 

diffuse, cortical discontinuity displaying a “moth-eaten appearance” (1963:395). More recently, 

Djuric et al. (2008) apply Nathan and Haas’s (1966) classifications of cribra orbitalia to femoral 

cribra with varying intensities of “porotic, cribrotic, and trabecular”; however, they do not 

compare these different grades in their analyses. Mangas-Carrasco and López-Costas (2021) also 

follow Nathan and Haas’s (1966) scoring system and analyzed the severity of femoral cribra in 

their study sample using a four-grade scale that ranges from small pores on the surface, which 

become increasingly connected, to exposure of trabecular structure. Radi and colleagues (2013) 

propose that the lesion can be described as either a cluster of pores on the cortical surface, or an 

area of cortical erosion that exposes trabeculae and may be depressed. The authors note that this 

latter expression of femoral cribra is consistent with Finnegan’s (1978) definition of Allen’s 

fossa, which Göhring (2021) suggests is a separate lesion unassociated with femoral cribra. 

Another process in the pathogenesis of femoral cribra possibly contributing to differences 

in the lesion’s appearance may be the amount of remodeling of the affected bone, or stage of 

“activity” at time of death. Mangas-Carrasco and López-Costas (2021) present stages of healing 

in their analysis of femoral cribra as active, mixed, or inactive that correspond to the relative 

visibility of the lesion’s morphology on the bone’s surface. For this system, active lesions are 

described as readily apparent, mixed lesions are recorded as less defined, and inactive lesions are 

observed as barely discernable. A raised, thickened border which sometimes surrounds the 

49 

 
porous bone of femoral cribra has been linked to possible lesion “activity” with Angel (1964) 

and Finnegan (1978) likening it to an inflammatory response and perhaps “a partial healing 

process” (Angel, 1964:135). In presenting a description of Allen’s fossa and distinguishing it 

from femoral cribra, Göhring states that Allen’s fossa “does not show a bony ridge or rim,” but 

does have a clear cortical margin that is proposed to be due the zona orbicularis eroding the 

cortical surface during activity (2021:4). 

One factor that may affect the presence of femoral cribra and its differential appearance is 

the individual’s age. Kotsick (1963:395) links variation in the lesion’s expression to age, even 

referring to one form of femoral cribra that has a “worn cancellous appearance” as the “teenage 

imprint”, which may represent an early stage of the lesion. In examining the prevalence of 

femoral cribra among age groups in skeletal samples, previous studies report higher frequency of 

this lesion in subadults and younger adults compared to older age groups (Miquel-Feucht et al., 

1999; Radi et al., 2013; Schats, 2021; Smith- Guzmán, 2015a). Additionally, Mangas-Carrasco 

and López-Costas (2021) report more active-looking lesions in subadult individuals and more 

lesions that appear healed in adult individuals within their skeletal sample from medieval 

northwest Spain (12th–15th centuries). This age-related pattern, as well as the location of 

femoral cribra in the region of epiphyseal union, has led some to suggest that it manifests earlier 

on during growth (Lewis, 2018; Smith- Guzmán, 2015a). Porous lesions resulting from anemia 

are thought to manifest only in regions of red or mixed marrow with potential for hyperplastic 

responses and thus, are age-dependent (Brickley, 2018). In the proximal femur, hematopoietic 

marrow is retained past twenty-five years of age (Kricun, 1985). This retention, coupled with red 

marrow’s heightened reactivity in subadults, could contribute to a hyperplastic change in 

response to a pathological stimulus (Lewis, 2018). Considering anemia as a possible etiology of 

50 

 
femoral cribra, the lesion could develop from the need for increased erythropoiesis and 

concomitant marrow hyperplasia (Smith- Guzmán, 2015a). The lesion may also manifest later in 

cases of chronic anemia, especially hemolytic anemia, as non-converted red marrow can become 

hyperplastic and result in expansion (Kricun, 1985). 

51 

 
 
CHAPTER FOUR: RESEARCH QUESTIONS & EXPECTATIONS 

This dissertation investigates hypothesized malarial infection in Nubia during the 

medieval period and the potential impact of these events using paleopathological approaches and 

a biocultural framework. This research focuses on the skeletal remains excavated from the site of 

Mis Island, located near the Fourth Cataract of the Nile River, which is believed to have 

practiced saqia agriculture during the medieval period based on excavated archaeological 

material. This research investigates the potential extent of the disease at Mis Island using Smith-

Guzmán’s (2015a) approach for inferring malarial infection from the skeleton, which entails 

examining skeletal remains for a series of five lesions associated with the disease and using a 

diagnostic outcome algorithm to infer how many and which individuals may have been infected 

with malaria. 

The prevalence of hypothesized malaria is explored between the sites’ cemeteries and 

across age and sex cohorts. Cemeteries 3-J-10 and 3-J-11 represent discrete burial areas on Mis 

Island and differ in time of use. While cemetery 3-J-10 was active during the late medieval 

period (1100–1500 CE), cemetery 3-J-11 spanned pre-medieval to late medieval years (300–

1400 CE). By comparing the prevalence of posited malarial infection between these cemeteries, 

this research examines potential temporal fluctuations in this disease. The prevalence of 

hypothesized malaria is compared between age and sex groups to generate inferences about 

relative exposure and the endemicity of malaria in this area during the medieval period. 

Investigating the dynamics of malaria in medieval Nubia builds on the scholarship of health 

conditions for populations living in the Nile River Valley at the time and contributes insight into 

the potential influence of posited agricultural and populational trends on malarial risk during the 

medieval period. 

52 

 
Given the context of hypothesized prevalence of malaria, this dissertation explores the 

manifestation of the skeletal lesions implicated in the disease to contribute insights into this suite 

of lesions and discussions about their etiology. This research examines the lesions’ relationship 

with the affected individuals’ age and the association among these lesions to contribute further 

understanding into their formation and patterns of expression. For one of these lesions—femoral 

cribra—there is a need to further understand variation in the lesion’s external appearance, the 

bone processes underlying its expression, and the role an individual’s age may have in the 

lesion’s manifestation. The research therefore visually examines femoral cribra lesions for 

several variables possibly representing lesion “activity” and “severity” and assesses their 

association with the age of the individual. By examining the variation in femoral cribra’s 

appearance across age groups, this analysis aims to assess lesion formation and progression in 

conjunction with age-related variables and possible healing activity. This assessment also seeks 

to better understand potential patterns in femoral cribra’s appearance that may be related to 

different stages of “activity” and “severity”. 

To investigate the relationship between changes visible from femoral cribra’s external 

surface and the types of bone processes possibly occurring at the trabecular structural level, this 

research used micro-computed tomography (micro-CT) methods to acquire three-dimensional 

scans from a subsample of femora and derive a series of variables that reflect trabecular bone 

microarchitecture from the region underlying the lesion. This analysis follows a framework 

based on Morgan’s (2014) micro-CT research on cranial porous lesions, and investigates the 

trabecular microarchitecture associated with femoral cribra to explore the types of bone 

processes involved in its formation, while considering proposed etiologies suggested to cause 

this lesion. Part of this analysis seeks to evaluate the relationship between external, macroscopic 

53 

 
signs of unremodeled or remodeling femoral cribra and the internal, microscopic changes 

occurring at the trabecular level that would inform descriptions of lesion “activity” (i.e., whether 

a lesion may be “active” or “healing”). Another facet of this analysis explores morphology of the 

lesion potentially corresponding to “severity” by assessing whether cortical alteration on the 

surface differs with internal trabecular variables. 

This chapter outlines the research questions of the dissertation research and discusses the 

expected outcomes for each question. The overarching aims of this research are to 1) investigate 

the prevalence and patterns of hypothesized malaria at Mis Island during the medieval period 

while considering the interactions between human activities, ecological contexts, and biological 

aspects of malaria, 2) explore the manifestation of the skeletal lesions of interest regarding their 

presence across age groups and co-occurrence, and 3) investigate femoral cribra’s formation and 

progression based on its macroscopic and microscopic expression and relationship with age. 

Research Question 1: Is there skeletal indication of malarial infection at Mis Island 

during the medieval period and if so, how prevalent may malaria have been? 

Expectation 1: Skeletal indication of malarial infection is expected to be present in the 

Mis Island sample and the estimated prevalence of hypothesized malaria is anticipated to be 

relatively high. 

The arable land at Mis Island would have likely provided a conducive environment for 

mosquito vector populations. Additionally, the inferred presence of saqia agriculture practices at 

Mis Island would have provided further sites for mosquito breeding, potentially supporting some 

degree of continuous malarial transmission. Indication of saqia irrigation at Mis Island during 

the medieval period is supported by remnants of the qadus jar uncovered at the medieval 

settlement 3-J-19 on Mis Island (Thomas, 2008). Although the qadus is not exclusive to saqia 

54 

 
irrigation, its primary use is in the cattle-driven water wheels part of this technique. With an 

increase in irrigated areas and water sources to support mosquito breeding, this farming practice 

may have resulted in a larger and/or more supported mosquito population and concomitant 

higher risk of malaria.  

While there were many urban centers in Makuria, uncovered settlements demonstrate a 

steady distribution of villages along the Nile throughout the medieval period (Obłuski, 2020). In 

the Fourth Cataract, remains of medieval sites indicate a settlement pattern of numerous and 

scattered small communities (Obłuski, 2020), several of which are observed in the vicinity of 

Mis Island. Additionally, the Fourth Cataract region saw the establishment of multiple fortified 

sites and churches during the medieval period, which were involved in the region’s organization 

over several phases (Żurawski, 2014). Therefore, while Mis Island was in a region of seemingly 

dispersed population density, the settlement practices were not so isolated as to restrict malarial 

transmission. Rather, the settlement pattern surrounding Mis Island in the Fourth Cataract region 

during the medieval period may have been associated with a regional population of appropriate 

size and proximity to sustain malarial transmission and be susceptible to inter-communal spread.   

Research Question 2: Are there differences in the frequency of hypothesized malaria 

between individuals from cemeteries 3-J-10 and 3-J-11? 

Expectation 2: The hypothesized presence of malaria is expected to be relatively similar 

between these two cemeteries, with potential for a higher frequency of infection in cemetery 3-J-

11. 

While Soler (2012) observed a similar overall prevalence of stress among adults from 

cemeteries 3-J-10 and 3-J-11, Hurst (2013) found significantly higher incidence of cribra 

orbitalia in subadults from 3-J-11 compared to 3-J-10. The current research may find a similar 

55 

 
pattern in greater prevalence of skeletal lesions suggestive of malaria among the cemetery 3-J-11 

subadult individuals. Additionally, there were periods before the late medieval phase during 

which the Nile River was at significantly lower levels than normal (Edwards, 2004). An 

ecological relationship between the seasonal shifts in the Nile River’s levels and An. arabiensis 

populations has been observed in the Northern State of Sudan in which mosquito numbers 

decrease during the river’s flooding as the high waters wash away breeding sites along the river 

(Ageep et al., 2009; Dukeen and Omer, 1986). However, once the river’s flood water diminishes, 

mosquito populations increase as the receding water creates stagnant pools that provide optimal 

breeding sites (Ageep et al., 2009; Dukeen and Omer, 1986). Based on this seasonal pattern 

between the mosquito vector’s environment and population status, abnormally low levels of the 

Nile could potentially provide sustained water sources for breeding rather than high water levels 

disrupting mosquito breeding sites. As such, mosquito vector populations and malarial 

transmission may have continued with mosquito populations possibly further supported by the 

use of artificial breeding sites including those associated with irrigation. 

Research Question 3: Are there differences in the frequency of hypothesized malaria 

between sex and age cohorts at Mis Island? 

Expectation 3: Indication of malarial infection is expected to be present in all sex and 

age cohorts with higher frequencies in younger age groups. 

A relatively similar prevalence of estimated malarial infection between males and 

females is expected with the hypothesis that there was not a considerable difference in exposure 

to mosquito vectors. Variation in exposure to mosquito vectors could result from differences in 

habitual activities, such as higher involvement in agricultural work which can influence risk of 

infection via close proximity to mosquito breeding grounds (Naidoo et al. 2011). However, any 

56 

 
differences in activities between males and females that may have been present at Mis Island 

during the medieval period are not expected to have placed a differential burden of exposure to 

water sources that could have supported mosquito vector breeding and affected malarial spread. 

This research question also considers the degree of endemicity at Mis Island during the 

medieval period which depends on at-risk groups, immunity, and periodicity of contact (Carter 

and Mendis, 2002). Endemic malaria involves some permanency of transmission, although this 

may be further classified as stable or unstable. Stable endemic malaria transmission is associated 

with environmental conditions that are favorable for continuous mosquito life cycles and the 

presence of malaria throughout the year (Carter and Mendis, 2002). Due to a regularity of 

exposure to malaria, some degree of protective immunity may be acquired by individuals in 

these regions. Areas experiencing unstable endemic malaria transmission are associated with a 

generally persistent presence of malaria, but the rate of infection is irregular following shifts in 

the epidemiological environment (Carter and Mendis, 2002). A period of interrupted contact 

between mosquito vectors and human hosts can result in decreased acquired immunity depending 

on the duration. During epidemics of malaria, there is a significant increase in transmission and 

all age groups in the population lack immunity (Carter and Mendis, 2002). Therefore, epidemic 

malaria may be more likely to result in higher mortality before skeletal indications manifest 

compared to endemic malaria, which may be more likely to involve some level of protection and 

chronic anemia (Smith-Guzmán et al., 2016). 

This analysis is expected to find indication of endemic malaria with signs of infection in 

all age groups, with a higher frequency in younger individuals as they are at greatest risk of 

infection (Carter and Mendis, 2002). Mis Island is situated in a region that has been categorized 

as having an unstable endemic malarial risk (Snow et al., 2008), which is likely associated in part 

57 

 
with the relationship between the Nile River’s flooding and the mosquito’s ecological niche. 

Mosquito populations along the Nile in northern Sudan tend to decrease during the flood season 

and increase as the river recedes, leaving stagnant pools of water used by mosquitoes as breeding 

sites (Dukeen and Omer, 1986). The increase in mosquito vector population can then lead to 

heightened risk of exposure for human hosts inhabiting the area. In line with the elevated 

frequencies of hypothesized malaria expected across age groups in this investigation, high 

prevalence of estimated malarial infection in older age groups may be anticipated as indications 

of possible lapses in acquired immunity. 

Research Question 4a: Are the skeletal lesions suggestive of malaria significantly 

associated with age? 

Expectation 4a: The lesions are expected to be associated with age and present more 

frequently in younger individuals, particularly with respect to cribra orbitalia, humeral cribra, 

femoral cribra, and spinal porosity. 

The five lesions of interest—cribra orbitalia, humeral cribra, femoral cribra, spinal 

porosity, and periosteal reaction—and their association with the individuals’ age are examined to 

evaluate age-related patterns in expression and to further consider their onset and chronicity. 

With anemia as a posited cause, this analysis examines the relationship between distribution of 

these lesions and changes in red marrow with age. Porous lesions resulting from anemia are 

thought to manifest only in regions of red or mixed marrow with potential for hyperplastic 

responses and are thus age dependent (Brickley, 2018). Therefore, these lesions are expected to 

be more prevalent in younger individuals.  

Such a pattern for cribra orbitalia, humeral cribra, and femoral cribra were observed by 

Schats (2021) in an investigation of these three lesions among skeletal remains excavated in the 

58 

 
Netherlands from the medieval/early modern period (800–1600 CE). In studying these remains, 

which were part of a research project studying the presence of malaria, Schats observed that 

cribra orbitalia, humeral cribra, and femoral cribra were significantly more prevalent in subadults 

compared to adults. Schats also found that as age increased, the chances of displaying humeral 

cribra and femoral cribra significantly decreased; however, this trend was not true for cribra 

orbitalia. 

Research Question 4b: Are there significant associations among the paleopathological 

lesions suggestive of malaria? 

Expectation 4b: Porous lesions are expected to frequently co-occur and demonstrate 

significant associations, particularly in younger individuals. 

Investigating the degree of association between these lesions further considers their 

formation with regard to the pathogenesis of malaria and the consistency in their patterns of 

expression. The pattern of correlation among lesions observed in this research is predicted to 

parallel that found by Smith-Guzmán (2015a) based on a shared hypothesized etiology of 

malaria, particularly with significant associations between humeral cribra and femoral cribra. 

While cribra orbitalia, humeral cribra, and femoral cribra are likely to frequently co-occur, cribra 

orbitalia may not be significantly associated with the other cribrous lesions as it has been 

associated with multiple etiologies (Wapler et al., 2004). 

A relatively frequent co-occurrence, although not statistically significant, between cribra 

orbitalia, humeral cribra, and femoral cribra was observed by Gomes and colleagues (2022) in a 

sample of subadult remains from Portugal dating to the 19th to 20th centuries. While Schats 

(2021) found that these three cribrous lesions were uncommon to co-occur in the same individual 

within a skeletal sample from the Netherlands (800–1600 CE), their concurrent expression was 

59 

 
seen more in subadult individuals than adults. Furthermore, for the subadult individuals Schats 

found femoral cribra frequently associated with cribra orbitalia and humeral cribra, and a 

statistically significant moderate association between humeral cribra and femoral cribra. Schats 

did not observe significant associations between pairs of the three cribrous lesions, although the 

subadult sample size was small suggesting that potential associations among the lesions of focus 

in the current research may be age-related. 

Research Question 5a: Are the macroscopic variables observed on the external surface 

of femoral cribra associated with age? 

Expectation 5a: Femoral cribra in subadult remains is expected to be larger and appear 

unremodeled compared to femoral cribra in adult remains, which is anticipated to be of more 

varied size and demonstrate signs of remodeling and potentially a raised cortical edge around 

the perimeter. 

This pattern is anticipated based on femoral cribra’s hypothesized etiology including 

anemia and the potential differences in extent of lesion remodeling from surviving health insults 

with age. A pattern of larger and more active-looking femoral cribra is anticipated in subadult 

ages when marrow is more reactive and cortical bone is thinner, thus potentially resulting in 

more severe lesions in response to stress (Lewis, 2018). Additionally, femoral cribra is expected 

to appear more active and severe in younger individuals compared to older individuals who 

survived prior pathological insult and might demonstrate remodeling lesions at different stages of 

healing. 

Mangas-Carrasco and López-Costas (2021) found a clear pattern in the activity of 

femoral cribra with age-at-death in skeletal remains excavated from archaeological sites in NW 

Spain dating to the late medieval period (12th–15th centuries). In examining the lesions’ surface 

60 

 
 
for signs of remodeling and scoring them as active, mixed, or inactive, the authors observed 

more active femoral cribra lesions in subadult individuals and more healing/healed lesions in 

adult individuals. Mangas-Carrasco and López-Costas also investigated potential severity of 

femoral crbira across age groups. While they did not include measurements of the lesion in their 

analyses, they used a 4-grade scale that progresses from small pores that increasingly become 

larger and exhibit trabeculae as the most severe score. The authors did not find significant 

differences in severity of femoral cribra by age; however, the study subsample was small. 

Research Question 5b: Are there patterns between macroscopic variables observed on 

the external surface of femoral cribra potentially related to lesion “activity” and “severity”?  

Expectation 5b: The presence of a raised cortical border around the lesion is expected to 

be associated with signs of remodeling and size of femoral cribra is expected to be associated 

with extent of cortical bone reduction. 

The presence of a raised cortical border surrounding the cribrous area of the lesion (a 

“cortical window”) is expected to be associated with signs of remodeling, as it was suggested as 

indication of possible “healing” processes by Angel (1964). Lesion size is anticipated to vary 

with scores proposed by Radi and colleagues (2013) describing the extent to which the cortical 

surface has been affected. Femoral cribra presented as a cluster of pores on the cortical surface 

(Radi et al. Score 1) is anticipated to be associated with smaller lesions while femoral cribra 

displaying cortical erosion and exposed trabeculae (Radi et al. Score 2) are expected to be larger. 

This prediction is based on a hypothesis that femoral cribra progresses from smaller lesions with 

relatively less cortical modification to comparatively larger lesions in which the cortical surface 

has been further reduced. 

61 

 
Research Question 6a: Are there differences in trabecular microarchitecture at the 

anterior-inferior femoral neck between unaffected individuals and affected individuals 

displaying femoral cribra with varying degrees of skeletal remodeling (i.e., “activity”)? 

Expectation 6a: Compared to unaffected individuals, affected individuals with lesions 

thought to be active are anticipated to show lower values for BV/TV, higher values for BS/BV 

and Tb.Sp., and either lower or higher values for Tb.Th. and Tb.N. Comparisons between 

unaffected individuals and affected individuals with lesions thought to be healing are expected to 

follow a similar patten, although not as distinct as comparing unaffected individuals and affected 

individuals with active lesions. In comparing lesions hypothesized to be active with those that 

appear to be healing, those lesions that demonstrate signs of remodeling on the surface are 

expected to exhibit higher values for BV/TV and Tb.Th., lower values for BS/BV and Tb.Sp., and 

either lower or higher values for Tb.N. 

The expected patterns of trabecular measurement differences for these comparisons are 

based on the types of bone processes (i.e., resorptive, formative, or mixed) hypothesized to be 

involved in the appearance of femoral cribra and possible signs of its activity from the surface. 

Compared to unaffected femora, the trabecular variables of femora with femoral cribra are 

anticipated to follow expected differences indicative of resorptive or mixed processes. These 

processes are hypothesized to be associated with the presence of the lesion and a proposed 

etiology of anemic conditions (Djuric et al., 2008). However, such differences are predicted to be 

more pronounced in comparing unaffected individuals and affected individuals with active 

lesions than in comparing unaffected individuals and affected individuals with healing lesions. 

This outcome is anticipated based on the thought that remodeling lesions would exhibit 

trabecular microarchitecture more similar to bone without the lesion. Finally, affected individuals 

62 

 
 
displaying skeletal remodeling are expected to present microarchitectural differences indicative 

of formative processes compared to those without indication of remodeling from the surface, 

therefore supporting potential interpretations of “healing” lesion activity. 

Research Question 6b: Are there differences in trabecular microarchitecture at the 

anterior-inferior femoral neck between unaffected individuals and affected individuals 

displaying varying extents of femoral cribra (i.e., “severity”)? 

Expectation 6b: Compared to unaffected individuals, affected individuals with lesions 

thought to be active and displaying varying degrees of cortical modification are anticipated to 

show lower values for BV/TV, higher values for BS/BV and Tb.Sp., and either lower or higher 

values for Tb.Th. and Tb.N. These differences are expected to be more pronounced in comparing 

unaffected femora and affected femora with more severe looking lesions than in comparing 

unaffected femora and affected femora with less severe looking lesions. In comparing affected 

individuals with lesions that appear less severe and more severe based on extent of cortical 

surface disruption, lesions that appear more severe are anticipated to demonstrate lower values 

for BV/TV, higher values for BS/BV and Tb.Sp. and either lower or higher values for Tb.Th. and 

Tb.N. 

The expected patterns of trabecular measurement differences for this set of comparisons 

are based on the types of bone processes (i.e., resorptive, formative, or mixed) hypothesized to 

be involved in the appearance of femoral cribra and possible signs of its severity from the 

surface. Compared to unaffected femora, the trabecular variables of the selected femora with 

femoral cribra are anticipated to follow expected differences indicative of osteoclastic activity 

and/or mixed cellular activity that would suggest a skeletal response to anemia, which is 

hypothesized to be associated with the lesion (Djuric et al., 2008). However, such differences are 

63 

 
predicted to be more pronounced in comparing unaffected individuals and affected individuals 

with more severe looking lesions than in comparing unaffected individuals and affected 

individuals with less severe looking lesions. Finally, in comparing affected femora that appear 

less severe and more severe, lesions that appear more severe are expected to present 

microstructural differences suggestive of increased osteoclastic change or marrow hyperplasia.  

64 

 
 
 
 
CHAPTER FIVE: MATERIALS & METHODS 

This chapter describes the skeletal sample studied in this dissertation and the data 

collection and analysis methods used to address the research questions. The first section details 

the demographic composition of the Mis Island skeletal sample of focus in this research. The 

second section describes the methods involved in investigating the presence of malaria at Mis 

Island during the medieval period and in examining the associated suite of skeletal lesions 

implicated in the disease. Finally, the methods used in exploring the external appearance of 

femoral cribra and its internal trabecular microarchitecture are discussed. 

All statistical analyses were performed in R Studio using Version 4.2.2 (Posit team, 2022) 

and statistical significance was set at p ≤ 0.05. Pearson chi-square tests were used to evaluate 

significance of associations between categorical variables except when one or more of the 

expected cell frequencies was less than five, in which case Fisher’s exact test was used instead. 

To assess the effect size of statistically significant outcomes, phi coefficients (φ) were calculated 

in instances of 2x2 contingency tables and Cramer’s V statistics were used when contingency 

tables were larger. These measures were interpreted as follows: values 0 to 0.10 indicating a 

small effect, values between 0.11 and 0.30 suggesting a moderate effect, and values greater than 

0.31 signifying a large effect (Wagner and Gillespie, 2019). 

The Mis Island skeletal sample 

The research sample studied in this dissertation comprises the human skeletal remains 

excavated from cemeteries 3-J-10 and 3-J-11 at the site of Mis Island. Cemetery 3-J-10 was 

situated in an area of alluvial deposits and was surrounded by rock outcrops (Ginns, 2010a, 

2006). This cemetery was potentially where the last Christian burials on Mis Island were interred 

and was in use in the Late Medieval Period between approximately 1100 CE and 1500 CE, 

65 

 
 
 
 
which includes the time period of Nubia converting to Islam (Ginns, 2006). On the southwest 

side of the cemetery, there is a “visible transition of belief” demonstrated by a shift in grave 

monument type with the early Muslim burials positioned near the Christian burials (Ginns, 

2006:17). An estimated 262 box-grave monuments were believed to be present at this site and by 

the end of the excavations, a total of 126 sets of human skeletal remains were recovered from 

across the cemetery (Ginns, 2010a). All of the burials were aligned east to west and there were 

two main burial styles observed in the cemetery—standard box-type monuments with or without 

rocks or mud-bricks covering the head (Ginns 2010a, 2007).  

Cemetery 3-J-11 was located closer to the river near the northern edge of the island in a 

terrain of alluvial formation and pebble deposits (Ginns 2010b). Representing the largest burial 

site on Mis Island, this cemetery spanned 300 CE – 1400 CE and thus was in use from the 

Meroitic Period through the Late Medieval Period (Ginns, 2006; Soler, 2012). A total of 288 sets 

of human remains were excavated from across cemetery 3-J-11 which was about half of the 

individuals likely interred at this site (Ginns, 2010b). The stone box-type monuments varied in 

size and style between graves and based on the variations in burial practices, this cemetery 

contained three primary phases (Ginns, 2010b, 2007, 2006). In most cases at cemetery 3-J-11, 

burials were oriented east to west to varying degrees (Ginns, 2010b). 

The human skeletal remains excavated from cemeteries 3-J-10 and 3-J-11 were 

previously curated by the Department of Anthropology at Michigan State University on behalf of 

the British Museum. During this time, Soler (2012) and Hurst (2013) analyzed the adult and 

subadult remains, respectively, and estimated age-at-death and biological sex as appropriate. 

These demographic data reported by Soler (2012) for adult individuals and Hurst (2013) for 

subadult individuals were used in the current research to group the Mis Island individuals into 

66 

 
age and sex cohorts for analyses. Soler (2012) estimated adult age-at-death with methods 

involving the pubic symphysis, auricular surface of the ilium, sternal rib ends, and cranial suture 

closure. Soler also estimated the biological sex of adult individuals using features of the pelvis 

and cranium and mandible. To estimate age-at-death for subadult individuals, Hurst (2013) used 

methods involving dental development stage, dental eruption, epiphyseal union, and long bone 

length. Hurst (2013) estimated biological sex for adolescent individuals if appropriate and 

possible.  

Individuals whose age-at-death could be estimated had been placed into standard age 

groups by Soler (2012) and Hurst (2013). Table 5.1 lists the subadult and adult age groups and 

associated age ranges used in the current research which are based on Hurst’s (2013) adjustments 

to the Standards for Data Collection from Human Skeletal Remains (Buikstra and Ubelaker, 

1994) and Human Bioarchaeological Database (Redfern, 2010) categories. These systems were 

used in the current research to group individuals into age cohorts as appropriate for each research 

question. Individuals whose biological sex could be estimated had previously been assessed as 

male, probable male, female, probable female, or a category of unknown sex. To increase 

subsample sizes in the current study, probable males were grouped into the male category and 

probable females were combined with the female group.  

Table 5.1: Subadult and adult age groups with associated age ranges based on Hurst’s (2013) 
revised Standards and Human Bioarchaeological Database categories. 

Revised Standards  
(Buikstra & Ubelaker, 1994) 
Infant (<3 years) 
Child (3–11 years) 
Adolescent (12–19 years) 
Young Adult (20–34 years) 
Middle Adult (35–49 years) 
Older Adult (≥50 years) 
Unknown Age Adult (≥20 years) 

Revised Human Bioarchaeological 
Database (HBD) (Redfern, 2010) 
Infant (<3 years) 
Child (3–7 years) 
Juvenile (8–16 years) 
Adolescent (17–19 years) 

67 

 
 
 
 
 
Of the total number of individuals from Mis Island (n = 402), 370 individuals were 

associated with at least one skeletal element to assess the presence of at least one of the five 

lesions of focus for this research. These individuals therefore represent the full sample for this 

dissertation, the demographic composition of which is displayed in Table 5.2. Of the total 370 

individuals included in the study, 121 individuals were from cemetery 3-J-10 and 249 individuals 

were from cemetery 3-J-11. Skeletal remains with age-at-death estimates of less than two-years-

old and associated with an age range including an estimate of less than twelve months were 

excluded from current analyses, as the presence of the lesions could not be reliably assessed. 

Otherwise, individuals from all age groups were included in the research sample. As shown in 

Table 5.2, the study sample included relatively equal proportions of male and female individuals 

from both cemeteries. 

Table 5.2: Mis Island research sample separated by cemetery, age-at-death, and biological sex. 

Age Group 
(Standards) 
Infant 
Child 
Adolescent 
Young 
Adult 
Middle 
Adult 
Older 
Adult 
Unknown 
Age Adult 
Total 

3-J-10 

3-J-11 

Total 

Male  Female  Unknown  Total  Male  Female  Unknown  Total 

– 
– 

12 

23 

2 

1 

– 
– 

4 

15 

11 

5 

19 
23 
5 
– 

1 

– 

– 

19 
23 
5 
16 

39 

13 

6 

– 
– 
3 
14 

30 

12 

– 

– 
– 
2 
20 

26 

21 

1 

31 
68 
16 
2 

1 

– 

2 

31 
68 
21 
36 

57 

33 

3 

50 
91 
26 
52 

96 

46 

9 

38 

35 

48 

121 

59 

70 

120 

249 

370 

68 

 
 
 
 
 
 
 
 
 
Investigating malaria at Mis Island and the associated series of skeletal lesions (Research 

Data collection 

Questions 1, 2, 3, 4a & 4b) 

To investigate malaria at Mis Island, this research used a method developed by Smith-

Guzmán (2015a), who proposed a suite of five skeletal lesions as indicators of malaria. These 

skeletal lesions are cribra orbitalia, humeral cribra, femoral cribra, spinal porosity, and periosteal 

reaction. The first four of these lesions are manifested as porous lesions that completely perforate 

cortical bone on the orbital plate of the frontal bone, the humeral and femoral necks, and the 

lateral and anterior surfaces of the vertebral bodies. Periosteal reaction is observed as the 

deposition of new bone on the outer cortex of long bone shafts as the result of an inflammatory 

response of the periosteum due to infection or trauma. The presence of these lesions was 

evaluated based on Smith-Guzmán’s descriptions and scored as 1=present, 0=absent, or 

unscorable. Lesions that could be expressed bilaterally were scored as present when they 

appeared on the right, left, or both skeletal elements. 

The diagnostic outcome algorithm generated by Smith-Guzmán (2015a:630) was used to 

infer whether an individual may have been infected with malaria, thus providing insight into 

potential prevalence and patterns of malaria across the sample. The algorithm entails looking at 

whether an individual displays a certain combination of lesions, as follows: if an individual 

displays femoral cribra, humeral cribra, or cribra orbitalia and also presents with spinal porosity 

or periosteal reaction, then the individual is associated with a “positive diagnosis” score; 

however, if an individual does not demonstrate one of these combinations of lesions, the 

individual is scored as “negative diagnosis” and interpreted as not associated with potential 

malarial infection. For each individual with the necessary skeletal elements present, this 

69 

 
algorithm was used to score which individuals met the criteria for a “positive” or “negative” 

diagnosis of malaria. 

Criteria for “Positive Diagnosis” (Smith-Guzmán (2015a:630)) = (Cribra orbitalia or humeral 

cribra or femoral cribra are present) and (Spinal porosity or periosteal reaction are present) 

Data analysis 

 In investigating hypothesized malarial infection at Mis Island during the medieval period 

for Research Question 1, frequency data for the five skeletal lesions were compiled and 

compared with those provided by Smith-Guzmán (2015a) taken from two modern samples—one 

endemic for malaria and one that is non-endemic (i.e., malaria-free). Frequencies of each lesion 

were calculated by dividing the number of individuals exhibiting a given lesion in relation to the 

total number of skeletal elements observable for that lesion. This comparison was carried out to 

evaluate whether the prevalence of lesions observed in the Mis Island sample were more similar 

to a population associated with a high prevalence of malaria or one considered to not have been 

affected by malaria. 

After performing Smith-Guzmán’s (2015a) algorithm for each individual to infer a 

possible “positive diagnosis” or “negative diagnosis” of malaria, the frequency of individuals 

scored as “positive diagnosis” was calculated in relation to the total number of individuals with 

the appropriate skeletal completeness. For individuals with observable lesions that did not result 

in a “positive diagnosis” but were missing lesions that could have produced a positive result, no 

score was given. Because this could introduce bias in assessing the frequencies of estimated 

malaria, a conservative overall prevalence was calculated based only on individuals scorable for 

all five lesions (Buzon and Sanders, 2016; Smith, 2015). The regular and conservative estimates 

for the prevalence of malaria at Mis Island during the medieval period were compared to 

70 

 
 
published frequencies that had used the same method from the sites of Tombos in Upper Nubia 

(Buzon and Sanders, 2016) and the South Tombs Cemetery at Amarna, Egypt (Smith-Guzmán et 

al., 2016). The site of Tombos is located at the Third Cataract of the Nile River and was 

inhabited between the mid-18th Dynasty of the Egyptian New Kingdom and the Napatan Period 

(~1400–650 BCE) (Buzon and Sanders, 2016). The skeletal remains excavated from the South 

Tombs Cemetery in Amarna, Egypt date to the Amarna Period (1349–1332 BCE) (Smith-

Guzmán et al., 2016). Of the total Mis Island subsample that was sufficiently preserved for the 

algorithm to be performed, the results were then parsed according to intra-site comparisons. 

For Research Question 2, the prevalence of individuals with hypothesized malarial 

infection and the associated five skeletal lesions were compared between cemetery 3-J-10 and 

cemetery 3-J-11. Hypothesized malaria was compared between both cemeteries overall and by 

age group (e.g., young adults from 3-J-10 vs. young adults from 3-J-11) using chi-square or 

Fisher’s exact tests as appropriate. Each lesion was also compared between cemeteries with chi-

square tests and significant results were further assessed by calculating associated phi 

coefficients. Potential differences in indications of malaria between cemeteries were then 

evaluated between males from cemetery 3-J-10 and males from cemetery 3-J-11, and between 

females from cemetery 3-J-10 and females from cemetery 3-J-11. For these analyses, only 

individual remains that could be assessed for biological sex estimation were included. Chi-square 

tests were used to examine hypothesized malaria infection and each skeletal lesion between 

males and females from both cemeteries with subsequent calculation of phi coefficients for 

statistically significant outcomes.  

The prevalence of hypothesized malaria between age and sex cohorts was examined in 

Research Question 3. Hypothesized presence of malaria was compared between males and 

71 

 
 
females separated by age groups using chi-square and Fisher’s exact tests. The prevalence of 

each lesion between males and females was also compared with chi-square tests and phi 

coefficients for statistically significant results. Chi-square test and Cramer’s V statistic were 

utilized to investigate the inferred malarial status across age groups. Further examination 

between age of the individual and the skeletal lesions was not included at this stage as that 

relationship was a focus for the subsequent research question.  

The relationship between age and the skeletal lesions suggestive of malaria was examined 

in Research Question 4a using chi-square tests and Cramer’s V statistics. To better visualize the 

relationship between these variables, a multiple correspondence analysis (MCA) was performed 

and presented in a biplot. The test was conducted with the presence and absence of the five 

skeletal lesions and the Standards age groups of infant, child, adolescent, young adult, middle 

adult, and older adult. The association among the five lesions was then investigated with chi-

square and Fisher’s exact tests and phi coefficients to address Research Question 4b. These 

analyses were performed on all individuals and then separated by subadult individuals (infant, 

child, and adolescent age groups) and adult individuals (young, middle, and older adults and 

adults of unknown age). The co-occurrence specifically between cribra orbtialia, humeral cribra, 

and femoral cribra was also examined to contribute to previous discussions of their correlation as 

part of a “cribrous syndrome”, originally defined by Miquel-Feucht and colleagues (1999). The 

prevalence of all three lesions together and as pairs were assessed across the entire sample, and 

separately among subadult and adult individuals. These findings were discussed in relation to 

observations made in other studies. 

72 

 
 
 
 
Exploring the expression of femoral cribra (Research Questions 5a & 5b and 6a & 6b) 

External features of femoral cribra—Data collection 

The investigation on femoral cribra’s external expression potentially associated with 

“severity” and “activity”, and the pattern of these variables across age groups, was conducted on 

all proximal femora displaying the lesion in the Mis Island cemetery 3-J-10 and 3-J-11 groups. 

For this research component, individuals in the “Unknown Age Adult (≥20 years)” age group 

were not included in the subsample as this age-at-death estimate was not narrow enough for 

more nuanced interpretation of the results. Additionally, individuals whose age-at-death 

estimates were less than two-years-old and associated with age ranges involving an estimate of 

less than twelve months were excluded from the subsample, as the femoral neck was not 

morphologically distinct enough to assess the presence or absence of femoral cribra.  

A series of macroscopic observations were collected from the external surface of each 

femoral cribra lesion to investigate variation in these feature’s relationship with age and their 

potential association with lesion “activity” and “severity”. As many variables that could be 

observed were collected for each femoral cribra lesion but were considered not recordable if the 

lesion was too damaged or incomplete. To evaluate possible changes related to “activity”, each 

femoral cribra lesion was visually examined for indication of unremodeled and remodeling bone. 

Following Mensforth and colleagues’ (1978) descriptions for porotic hyperostosis, “active” 

femoral cribra may be presented as sharp-edged lesions, whereas femoral cribra that has 

experienced some degree of remodeling, or “healing”, may display smooth-edged bone in the 

affected area. Observations noting these sharp- vs. smooth-edged differences were made for each 

case, with those presenting exclusively sharp-edged margins recorded as possible “active” 

lesions and those exhibiting rounded bony margins within the lesion recorded as possible 

73 

 
“healing” lesions. Additionally, the presence or absence of a raised, thickened border which 

sometimes surrounds the porous area was evaluated (i.e., a “cortical window”). This border was 

considered present when there was a distinct margin immediately adjacent and encircling the 

lesion’s pores or trabeculae (Figure 5.1).  

Figure 5.1: Example of bilateral “cortical windows” encircling femoral cribra (Cemetery 3-J-10, 
Skeleton 1005). 

To assess possible differences in “severity”, femoral cribra was scored using Radi and 

colleagues’ (2013) categories that reflect the degree of cortical disruption. If femoral cribra was 

present on a femur, the lesion’s appearance was assessed according to this system and scored as 

either 1) a cluster of pores on the cortical surface, or 2) an area of cortical erosion exposing 

trabeculae that may be depressed. This scoring system was selected for the current study given 

the relatively unambiguous distinction between the two scores, as well as previous findings of 

low intraobserver and interobserver error (Radi et al., 2013). In addition to these scores, several 

measurements were taken using a tape measure to investigate variation in lesion size. These 

measurements included the perimeter of the lesion and lengths of the lesion parallel and 

perpendicular to the femoral neck. 

74 

 
 
 
 
 
External features of femoral cribra—Data analysis 

Analyses of the variation in femoral cribra’s external surface were conducted separately 

for left and right femora as both skeletal elements were not present for every individual. 

Individuals were separated and compared according to the Standards age group categories (Table 

5.1), except in analyses involving measurements of the lesion. For analyses that examined the 

three measurements of femoral cribra, the subadult age ranges used in the Human 

Bioarchaeological Database (Table 5.1) were used as these involve narrower age ranges that 

could better account for growth of the femoral neck with increasing age.  

Towards addressing Research Question 5a, Fisher’s exact test and Cramer’s V statistic 

were used to examine the relationship between age groups and femoral cribra’s Radi et al. Score, 

estimated activity, and presence of a cortical window. Multiple correspondence analysis (MCA) 

was used to explore the relationship between age group, Radi et al. Score, estimated activity, and 

status of a cortical window. Scatter plots were used to investigate variation in size, represented as 

perimeter of the lesion, and estimated activity of the lesion across age groups. In examining 

Research Question 5b, Fisher’s exact tests were used to assess the association between 

hypothesized lesion activity and the presence of a cortical window for left and right femora, and 

phi coefficients were used to evaluate the strength of these associations. Lesion size was 

compared between Radi et al. Scores 1 and 2 by calculating the mean and standard deviation of 

each score for the three femoral cribra measurements taken perpendicular to the femoral neck, 

parallel to the femoral neck, and around the lesion’s perimeter. These calculations were 

conducted separately by age group using the subadult age ranges of the Human 

Bioarchaeological Database. Boxplots were also used to compare these three measurements 

between lesions with Radi et al. Scores 1 and 2 within each age group. 

75 

 
 
Internal features of femoral cribra—Data collection 

A total of 27 femora were examined with micro-CT methods to explore trabecular 

microarchitecture (Table 5.3). The maximum length of a femur to fit in the sample sled was 

405mm, which was one criterion for including proximal femora in the study sample. Preference 

was given towards femora belonging to individuals whose skeletal remains were excavated from 

cemetery 3-J-11 at Mis Island; however, femora were selected from the cemetery 3-J-10 

subsample if needed. For individuals presenting bilateral expressions of femoral cribra, the left 

femur was selected for the sample and the right femur was included if the left was either absent 

or unsuitable for micro-CT analysis. While asymmetrical expressions of femoral cribra is a 

subject to evaluate in future research, the current study focused on the lesion’s variation across 

individuals. Additionally, previous work by Radi and colleagues (2013) did not find significant 

differences in the frequency and appearance of femoral cribra between sides. To gain a better 

understanding of femoral cribra’s variation and development, subadult and adult individuals from 

different age cohorts were included. In addition to age group, adult individuals were separated by 

sex for comparison. 

76 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 5.3: Number of femora per cohort examined using micro-CT methods, separated by 
affected/unaffected status, cemetery, and side (YA: Young Adult, MA: Middle Adult). 

Affected 

Unaffected 

Cohort 

3-J-11 

3-J-10 

3-J-11 

3-J-10 

Total 

Left  Right  Left  Right  Left  Right  Left  Right 

Subadult  
2–3 yrs 
Subadult  
3–5 yrs 
Subadult  
4–7 yrs 
Subadult  
10–15 yrs 
YA Female 
20–34 yrs 
YA Male  
20–34 yrs 
MA Female  
35–49 yrs 
MA Male  
35–49 yrs 
Total 

1 

3 

2 

1 

1 

1 

1 

1 

– 

– 

– 

– 

– 

1 

– 

– 

1 

– 

1 

1 

– 

– 

– 

– 

– 

– 

– 

– 

1 

– 

1 

– 

2 

– 

– 

– 

– 

– 

– 

1 

– 

2 

1 

– 

1 

– 

1 

– 

– 

– 

– 

– 

– 

1 

1 

– 

– 

– 

– 

– 

– 

– 

– 

– 

4 

5 

4 

2 

3 

3 

4 

2 

17 

10 

27 

The selected sample included proximal femora with and without femoral cribra to 

compare trabecular microarchitecture between affected and unaffected individuals. The 

subsample of femora with femoral cribra displayed varying expressions of the lesion based on 

macroscopic observations investigated for their potential association with “activity” and 

“severity”. These characteristics primarily included the appearance of sharp- vs. smooth-edged 

lesion margins and the extent of porosity/exposed trabeculae. The subsample of femora without 

femoral cribra demonstrated either no visible signs of porosity or, in the case of subadult 

individuals, displayed porosity associated with remodeling during growth. This normal porosity 

seen on the surface of subadult metaphyses is due to resorption under the periosteum, revealing 

haversian canals in the cortical bone (Ortner, 2003). Although this porosity at juvenile 

metaphyses can overlap and be confused with pathological change (Buikstra and Ubelaker, 1994; 

Ortner, 2003), this study sought to distinguish pathological porosity as perforations through the 

77 

 
 
cortex and exposure of trabecular bone (Brickley and Ives, 2006) displayed on the anteroinferior 

femoral neck. The subsample without femoral cribra was selected to parallel the age and sex 

composition of the subsample with femoral cribra to allow for more controlled comparisons.  

Micro-CT scan and data acquisition was conducted with the facilities at the MSU 

Institute for Quantitative Health Science & Engineering (IQ) and with the aid of Jeremy Hix of 

the MSU IQ Imaging Core Facility. The Quantum GX2 micro-CT imaging system manufactured 

by PerkinElmer was used for scanning and the Bone Phantom QRM-MicroCT-HA-D10 was used 

for calibrating hydroxyapatite density. Image acquisition settings included: 90kV, 88µA, voxel 

size 120µm, FOV acquisition of 70 and recon of 60, and scan time of approximately 14 minutes. 

After the micro-CT images were acquired, the Analyze 14.0 (AnalyzeDirect, Overland Park, KS) 

software application Bone Microarchitecture Analysis (BMA) was used for processing micro-CT 

scans, collecting quantitative data on trabecular microarchitecture, and visually assessing cortical 

and trabecular differences. 

Generating volumes of interest (VOIs) 

Before obtaining quantitative measurements from the micro-CT scans, a volume of 

interest (VOI) was selected from each scanned specimen. The measurements were then 

calculated from the trabecular bone captured within the VOI’s boundaries. The approach for VOI 

selection must be carefully considered, as the dimension, size, and location of a VOI will impact 

the interpretation of the results. A VOI that is inappropriately small or large relative to the region 

of interest can introduce unintended biases into the study by under- or over-sampling trabecular 

microarchitecture (Fajardo and Müller 2001; Lazenby et al. 2011). Improperly fitted VOIs across 

a study sample can thus increase the potential of Type I and Type II errors (Lazenby et al., 2011; 

Morgan, 2014). This consideration also applies to comparing individuals of varying body size, as 

78 

 
a uniform VOI across the sample can over- or under-sample (Fajardo and Müller, 2001, Lazenby 

et al., 2011). Some researchers, therefore, suggest scaling VOIs for each case rather than 

establishing a fixed sized VOI to decrease the potential for error (Lazenby et al., 2011). Previous 

research suggests that manually-adapted VOIs produce more pertinent and comprehensive results 

compared to fixed geometric VOIs (Djuric et al., 2013; Marchand et al., 2010). In a micro-CT 

study on porous lesions of the orbit and cranial vault, Morgan (2014) tested the reliability and 

reproducibility of uniform 2D and 3D VOIs and custom sized 3D VOIs in cranial bones. 

Compared to 2D and 3D standard fixed size VOIs, custom sized 3D VOIs with varied 

dimensions depending on each individual specimen were considered the most appropriate 

method for quantifying bone microarchitecture.  

The location of the sampled bone must also be considered, as trabecular architecture can 

change over relatively short distances (Whitehouse and Dyson 1974; Whitehouse, 1975). 

Therefore, anatomical and biomechanically relevant variables specific to the location should be 

taken into account when determining a VOI. For example, the proximal femur is a 

biomechanically complex structure with varying trabecular groups and morphological features 

across its anatomical regions. The trabeculae underlying the medial femoral neck, where femoral 

cribra manifests, belong to the principal compressive group (Whitehouse and Dyson, 1974). This 

vertical system extends from the medial proximal diaphysis to the articular surface of the head 

and is characterized by dense trabeculae (Whitehouse and Dyson, 1974). 

In the current research, custom sized 3D VOIs were established for each specimen which 

resulted in approximately cube-shaped subregions. For femora displaying femoral cribra, the 

VOIs were generated to encompass the visibly affected surface and the trabeculae underneath. 

For femora without the lesion, the VOIs were created to mirror the size and location of the VOIs 

79 

 
 
of their affected counterparts. Scans were rotated as needed to ensure proper VOI selection in 

affected and unaffected femora. Using the Transform application, the VOI dimensions were 

adjusted within x, y, and z planes to encompass the region of focus. A 3D rendering of the 

scanned specimen was used to visualize the medial-lateral and superior-inferior boundaries of the 

VOI from the bone’s surface. After these edges were established, the depth of the VOI was 

adjusted to one quarter of the anterior-posterior distance from the surface. This depth was chosen 

to isolate and emphasize the trabeculae directly underlying the area of interest while also 

attempting to avoid potential biomechanical influences of the femoral neck’s trabecular structure. 

Prior to establishing this protocol of VOI selection, the researcher assessed the scanned 

trabecular regions to confirm its suitability and considered biases in varying the morphology and 

location of the subregions (Kivell et al., 2011). To limit potential confounding effects in 

comparing data from VOIs at different areas on the anteromedial femoral neck, subregions of the 

unaffected femora were made specific for comparing to each of the affected femora in a cohort 

group that displayed lesions in slightly different locations. Therefore, for some comparisons 

multiple VOIs were created from one unaffected femur to be compared with its affected 

counterparts depending on whether lesions varied in location. Additionally, data from affected 

femora with lesions at slightly different areas on the femoral neck were only compared to other 

affected femora in their cohort group if they were similar in location or if the VOIs could be 

adjusted appropriately for comparison. 

Image thresholding and segmentation 

After isolating the VOI from a scanned femur, the process of segmentation was 

conducted to identify objects within that subregion. Image segmentation involves establishing 

threshold values for voxels of varying gray levels that accurately associate with the given object 

80 

 
 
in that volume of interest (e.g., bone tissue vs. space occupied by air). The Segment module 

allows automatic and manual approaches to distinguish bone from non-bone objects and to 

isolate cortical bone from trabecular bone (AnalyzeDirect, Overland Park, KS). The outcome of 

this procedure was an object map with the following features defined and separated from which 

quantitative measurements were automatically calculated: air surrounding the specimen, cortex, 

intra-trabecular space, and trabecular bone. 

To better segment cortex from trabecular bone in the VOIs with similar grayscale values, 

this research used the Walls tool with the Spline option to define curved boundaries interpolated 

across each slice of the subregion separating cortical and trabecular bone. After a Wall was 

established across the subregion, the objects were separated through threshold segmentation. 

This process entails assigning a range of voxels in the volume to a certain object such that all 

voxels falling between the threshold’s minimum and maximum values are ascribed to that object. 

In this research, thresholds were derived using the Segment module which calculated the volume 

histogram and provided the option to produce up to eight threshold values. These automatically 

calculated threshold values were used in segmenting bone from air in the subregion and the most 

appropriate threshold option was determined through visual assessment. The threshold range that 

distinguished bone was assigned for the cortex and trabecular bone objects, which were then 

separated using the Object Separator tool in conjunction with the pre-defined Wall drawn 

between them. Intra-trabecular space and the air surrounding the bone were then assigned to 

threshold values less than those assigned to bone. These two objects were also then separated 

using the Object Separator tool which was made more effective with the already established 

Wall. 

81 

 
 
Data acquisition 

After establishing the VOIs, quantitative measures were automatically calculated for the 

trabecular bone within each femora’s VOI using the BMA (Bone Microarchitecture Analysis) 

application. Data for the five variables described as follows were derived from each VOI 

(AnalyzeDirect, Overland Park, KS; Hildebrand et al., 1999; Morgan, 2014): 

•  Bone volume density (BV/TV, %): the ratio of trabecular bone volume (BV) to the total 

volume of interest (TV) 

•  Specific bone surface (BS/BV, mm2/mm3): the ratio of the bone surface area (BS) to the 

volume of trabecular bone (BV) 

•  Trabecular thickness (Tb.Th., mm): the mean thickness of trabeculae within a volume 

of interest 

•  Trabecular number (Tb.N., 1/mm): the average number of trabeculae per unit length 

and reported as a 2D variable for each axial slice 

•  Trabecular spacing (Tb.Sp., mm): the mean distance between trabeculae within a 

volume of interest 

Standard deviation values were calculated for trabecular thickness, number, and spacing. As a 

two-dimensional measurement, trabecular number was originally reported for each axial slice in 

a given VOI. In order to arrive at a single value representative of Tb.N. for the entire VOI, the 

mean across slices was calculated.  

Internal features of femoral cribra—Data analysis 

These variables were understood in the following terms (Morgan, 2014). Bone volume 

density and specific bone surface were examined to evaluate how the amount of bone was 

affected by an underlying cause. Trabecular thickness represented the relative amount of bone 

82 

 
 
and differences resulting from trabecular resorption or formation. Trabecular number reflected 

the general structure and organization of the trabecular bone with differences indicating bone 

resorption or formation. Trabecular spacing corresponded to organization and amount of marrow 

space with changes suggesting trabecular resorption or formation.  

For Research Questions 6a and 6b the data for the five quantitative, microarchitectural 

measures were compared among individuals within their respective demographic cohorts. The 

first set of comparisons (Research Question 6a) were between unaffected individuals and 

affected individuals displaying femoral cribra hypothesized to reflect either “active” or “healing” 

lesions based on macroscopic signs of skeletal remodeling on the surface. All affected lesions 

included in this stage of analysis shared a Radi et al. 2 Score. Lesions that appeared “active” and 

“healing” were also compared with each other within their respective cohorts. The second set of 

comparisons (Research Question 6b) were between unaffected individuals and affected 

individuals with “active” femoral cribra of varying expressions possibly reflecting lesion 

“severity”, which was described as either “less severe” or “more severe” based on indications 

from the surface. A subgroup of lesions inferred to be “less severe” or “more severe” were then 

compared with each other in the same cohort. Comparisons were conducted between pairs of 

individuals from the same cohort with equivalent VOIs and were made across the five variables. 

For each comparison pair, the measures for one individual were evaluated to be either greater 

than or less than the other individual for each variable. This process was done across pairs and 

tabulated to investigate potential patterns. Differences in the variables were interpreted in terms 

of inferred types of bone processes as presented by Morgan (2014) and contextualized as 

potential insight into the pathogenesis of femoral cribra. The anticipated changes to the 

83 

 
 
trabecular microarchitectural variables for these processes described by Morgan (2014:149) are 

as follows: 

•  Resorptive processes (osteoclastic activity): decrease in BV/TV and Tb.Th., increase in 

BS/BV and Tb.Sp., and either decrease or increase in Tb.N. 

•  Formative processes (osteoblastic activity): increase in BV/TV and Tb.Th., decrease in 

BS/BV and Tb.Sp., and either decrease or increase in Tb.N. 

•  Mixed processes (mixed cellular activity): decrease in BV/TV, increase in BS/BV and 

Tb.Sp., and either decrease or increase in Tb.Th. and Tb.N. 

84 

 
 
 
CHAPTER SIX: RESULTS 

Research Question 1 

Research Question 1: Is there skeletal indication of malarial infection at Mis Island 

during the medieval period and if so, how prevalent may malaria have been? 

The frequencies of each lesion in the Mis Island sample were calculated (Table 6.1) and 

compared to those reported by Smith-Guzmán (2015a) for malaria-endemic and malaria-free 

samples (Table 6.2 and Figure 6.1). Within the Mis Island sample, femoral cribra was the most 

prevalent of the five lesions (77%), followed by cribra orbitalia and periosteal reaction (both 

59%), spinal porosity (52%), and humeral cribra (37%). When compared to Smith-Guzmán’s 

(2015a) endemic and non-endemic samples, the Mis Island sample demonstrated higher 

frequencies of femoral cribra and humeral cribra. The frequencies of cribra orbitalia and 

periosteal reaction in the Mis Island sample were similar to those of the malaria-endemic sample, 

although slightly higher in the Mis Island sample. The prevalence of spinal porosity in the Mis 

Island sample was intermediate between the endemic and non-endemic samples. 

Table 6.1: Prevalence of lesions observed in the total Mis Island sample. 

Lesion 
Cribra orbitalia 
Humeral cribra 
Femoral cribra 
Spinal porosity 
Periosteal reaction 

Present  Absent  Total  Prevalence (%) 
292 
255 
292 
316 
334 

59% 
37% 
77% 
52% 
59% 

119 
161 
66 
152 
136 

173 
94 
226 
164 
198 

Table 6.2: Frequencies of lesions in the Mis Island sample compared to Smith-Guzmán’s (2015a) 
endemic and non-endemic samples. 

Lesion 
Cribra orbitalia 
Humeral cribra 
Femoral cribra 
Spinal porosity 
Periosteal reaction 

Mis Island  Endemic  Non-endemic 

59% 
37% 
77% 
52% 
59% 

57% 
24% 
39% 
89% 
52% 

2% 
5% 
0% 
18% 
17% 

85 

 
 
 
 
Figure 6.1: Grouped bar plot of lesion prevalence in the Mis Island sample and Smith-Guzmán’s 
(2015a) endemic and non-endemic samples. 

Of the total sample (n = 370), 296 individuals could be assessed for the potential presence 

of malaria using a subset of the lesions in the algorithm presented by Smith-Guzmán (2015a). 

The demographic data of this subsample are presented in Table 6.3. Within this subsample, 223 

individuals (75%) met the algorithmic criteria for a “positive diagnosis” and thus demonstrated 

signs of possible infection (Table 6.4). Two hundred individuals were scorable for all five lesions 

and 155 met the “positive diagnosis” criteria (Table 6.4). Therefore, a conservative estimate of 

78% displayed indications of hypothesized malarial infection. When separated by cemetery, 71% 

and 77% of individuals from cemeteries 3-J-10 and 3-J-11 met the criteria for a “positive 

diagnosis”, respectively. When only the individuals scorable for all five lesions were included for 

the algorithm from each cemetery, conservative estimates of 80% and 76% indicated signs of 

possible infection in cemeteries 3-J-10 and 3-J-11, respectively.  

86 

 
 
 
Table 6.3: Individuals with the appropriate skeletal elements present to estimate either a 
“positive or negative diagnosis” for malaria using Smith-Guzmán’s (2015a) algorithm (M: Male, 
F: Female, U: Unknown). 

3-J-10 

U  Total  M 
– 
19 
19 

M 
– 

– 

0 

9 

F 
– 

– 

0 

4 

15 

11 

1 

– 

8 

1 

19 

19 

5 

– 

1 

– 

– 

5 

13 

27 

9 

1 

F 
– 

– 

2 

3-J-11 
U 
23 

Total 
23 

60 

16 

1 

1 

– 

1 

60 

20 

31 

45 

23 

1 

Total 

42 

79 

25 

44 

72 

32 

2 

– 

2 

13 

17 

22 

22 

17 

6 

– 

25 

24 

44 

93 

43 

58 

102 

203 

296 

Infant 
(<3 yrs) 
Child 
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 
Unknown 
Age Adult 
(≥20 yrs) 
Total 

Table 6.4: Prevalence of “positive diagnoses” and “negative diagnoses” for Mis Island 
cemeteries 3-J-10 and 3-J-11 following a regular estimate approach and a conservative estimate 
approach. 

“Positive 
Diagnosis” 

“Negative 
Diagnosis” 

Total 

Frequency of 
“Positive Diagnosis” 

  3-J-10 

e
t
a
m

i
t
s
e

e
t
a
m

i
t
s
e

r
a
l
u
g
e
R

e
v
i
t
a
v
r
e
s
n
o
C

  3-J-10 

3-J-11 

Total 

3-J-11 

Total 

66 

157 

223 

48 

107 

155 

27 

46 

73 

12 

33 

45 

93 

203 

296 

60 

140 

200 

71% 

77% 

75% 

80% 

76% 

78% 

The frequencies of hypothesized malarial infection for the total Mis Island sample were 

compared to those reported from studies on the Upper Nubian site Tombos (Buzon and Sanders, 

2016) and the South Tombs Cemetery in Amarna, Egypt (Smith-Guzmán et al., 2016) which used 

87 

 
 
 
 
 
 
 
 
 
Smith-Guzmán’s method to estimate the potential prevalence of malaria (Table 6.5). Of the total 

number of individuals analyzed from Tombos dating between the New Kingdom and Napatan 

periods (~1400–650 BCE, n = 60), 37% individuals met the algorithmic criteria for a “positive 

diagnosis”. Of the individuals analyzed from the South Tombs Cemetery dating to the Amarna 

Period (1349–1332 BCE, n = 417 although it is unclear if all individuals were included), an 

estimated 50% of individuals displayed indication of potential infection. Both the regular and 

conservative estimates for the prevalence of hypothesized malaria at Mis Island (~300–1500 

CE)—75% and 78%, respectively—were higher than those found in the skeletal remains 

analyzed from Tombos and the South Tombs cemetery. 

Of the estimates reported from the other sites, the closest frequency to those from Mis 

Island was the estimate based only on complete skeletal remains (i.e., a conservative estimate) 

from Tombos dating to the Third Intermediate and Napatan Periods (Table 6.5). The conservative 

estimate from this subsample was 56% (10/18) and was associated with a regular estimate of 

48% (11/23) (Buzon and Sanders, 2016). The next closest frequency was that reported from the 

South Tombs Cemetery which was an estimate of 50% (Smith-Guzmán et al., 2016). The 

estimated prevalence of malaria in the Tombos New Kingdom subsample was the most dissimilar 

compared to Mis Island and was reported at 30% (11/37) and conservatively at 35% (7/20) 

(Buzon and Sanders, 2016). 

88 

 
 
 
 
 
Table 6.5: Regular and conservative estimates for the prevalence of malaria at Mis Island, 
Tombos (Buzon and Sanders, 2016), and the South Tombs Cemetery (Smith-Guzmán et al., 
2016). 

Site 

Mis Island, Fourth Cataract, Upper Nubia 
(Medieval Period) 
Tombos, Third Cataract, Upper Nubia 
(All time periods: New Kingdom – 
Napatan Periods) 
Tombos, Third Cataract, Upper Nubia 
(Third Intermediate & Napatan Periods) 
Tombos, Third Cataract, Upper Nubia 
(New Kingdom Period) 
South Tombs Cemetery, Amarna, Egypt  
(Amarna Period) 

Estimated Prevalence of Malaria 

Regular  
estimate 

Conservative  
estimate 

75% 

78% 

48% 

30% 

37% 

50% 

56% 

35% 

Research Question 2 

Research Question 2: Are there differences in the frequency of hypothesized malaria 

between individuals from cemeteries 3-J-10 and 3-J-11? 

The prevalence of hypothesized malaria was first compared between all individuals from 

cemeteries 3-J-10 and 3-J-11. A chi-square test between both cemeteries revealed no significant 

difference in the prevalence of inferred malarial infection (Table 6.6). Furthermore, a similar 

frequency of “positive diagnoses” for malaria was found between both cemeteries, as 71% of 

individuals from cemetery 3-J-10 and 77% of individuals from cemetery 3-J-11 displayed signs 

of possible infection (Table 6.6). When separated by age, Fisher’s exact and chi-square tests 

yielded no significant differences between cemeteries 3-J-10 and 3-J-11 for each age cohort 

(Table 6.6). The frequencies of “positive diagnoses” for malaria were also similar between the 

same age groups from both cemeteries (Table 6.6). 

89 

 
 
 
 
 
 
Table 6.6: Prevalence of “positive diagnoses” (P.D.) and “negative diagnoses” (N.D.) between 
cemeteries 3-J-10 and 3-J-11, separated by age group, as well as p-values from Fisher’s exact 
(FET) tests and chi-square tests (df = 1). 

3-J-10 

3-J-11 

P.D.  N.D.  Total  P.D.%  P.D.  N.D.  Total  P.D.% 
74% 
14 

74% 

17 

23 

19 

6 

5 

16 

5 

11 

3 

0 

2 

19 

84% 

48 

12 

60 

80% 

5 

100% 

19 

13 

85% 

28 

1 

3 

20 

95% 

31 

90% 

15 

12 

27 

56% 

30 

15 

45 

67% 

4 

5 

9 

44% 

14 

9 

23 

61% 

65 

27 

92 

71% 

156 

46 

202 

77% 

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent  
(12–19 yrs) 
Young Adult  
(20–34 yrs) 
Middle Adult  
(35–49 yrs) 
Older Adult  
(≥50 yrs) 
Total 

3-J-10 vs. 
3-J-11 
FET 
p = 1.00 
FET 
p = 1.00 
FET 
p = 1.00 
FET 
p = 0.62 
χ2 = 0.48 
p = 0.49 
FET 
p = 0.45 
χ2 = 1.46 
p = 0.23 

In comparing the prevalence of each lesion between individuals from cemeteries 3-J-10 

and 3-J-11, significant differences were found between the cemeteries for cribra orbitalia and 

spinal porosity (Table 6.7). Cribra orbitalia was observed at a higher frequency among 

individuals from cemetery 3-J-11 (64%) than individuals from cemetery 3-J-10 (49%), which 

was a pattern found across all age groups (Table 6.8). Spinal porosity was also more prevalent 

among individuals from cemetery 3-J-11 (59%) than individuals from cemetery 3-J-10 (37%); 

however, only the adult age groups from cemetery 3-J-11 demonstrated higher frequencies in 

spinal porosity compared to individuals from cemetery 3-J-10 (Table 6.11). Although these 

comparisons yielded statistically significant results, the associated phi coefficients suggest that 

the effects are relatively small for cribra orbitalia and small-medium for spinal porosity (Table 

6.8 and Table 6. 11, respectively). The results for humeral cribra, femoral cribra, and periosteal 

reaction reflected no significant differences in the prevalence of these lesions between 

individuals from cemetery 3-J-10 and individuals from cemetery 3-J-11 (Table 6.7). In 

90 

 
 
 
 
 
comparing their prevalence between cemeteries, these three lesions were observed at similar 

frequencies in cemetery 3-J-10 and 3-J-11 (Tables 6.9, 6.10, and 6.12). 

Table 6.7: Results of chi-square tests (df = 1) between cemeteries 3-J-10 and 3-J-11 for each 
lesion and phi coefficients for significant results (bolded p-values). 

Cribra orbitalia 

Humeral cribra 

Femoral cribra 

Spinal porosity 

Periosteal reaction 

χ2 = 6.13 
p = 0.013 
φ = -0.15 
χ2 = 0.24 
p = 0.62 
χ2 = 0.018 
p = 0.89 
χ2 = 12.18 
p = 0.00048 
φ = -0.2 
χ2 = 0.70 
p = 0.40 

Table 6.8: Prevalence of cribra orbitalia (CO) in cemeteries 3-J-10 and 3-J-11, separated by age.  

Total  % 

Total  % 

Table 6.9: Prevalence of humeral cribra (HC) in cemeteries 3-J-10 and 3-J-11, separated by age. 

Total  % 

Total  % 

CO 
Present 
14 
14 
1 
6 
6 
4 
45 

Infant 
Child 
Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Total 

HC 
Present 
9 
11 
2 
3 
1 
1 
27 

Infant 
Child 
Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Total 

3-J-10 
CO 
Absent 
5 
4 
3 
8 
21 
6 
47 

3-J-10 
HC 
Absent 
9 
10 
2 
6 
20 
4 
51 

CO 
Present 
16 
39 
17 
19 
20 
15 
126 

HC 
Present 
6 
37 
9 
12 
2 
1 
67 

3-J-11 
CO 
Absent 
2 
8 
1 
14 
32 
13 
70 

3-J-11 
HC 
Absent 
18 
22 
6 
11 
36 
17 
110 

89% 
18 
83% 
47 
94% 
18 
58% 
33 
38% 
52 
54% 
28 
196  64% 

25% 
24 
63% 
59 
60% 
15 
52% 
23 
5% 
38 
18 
6% 
177  38% 

19 
18 
4 
14 
27 
10 
92 

74% 
78% 
25% 
43% 
22% 
40% 
49% 

18 
21 
4 
9 
21 
5 
78 

50% 
52% 
50% 
33% 
5% 
20% 
35% 

91 

 
 
 
 
 
 
 
 
 
 
Table 6.10: Prevalence of femoral cribra (FC) in cemeteries 3-J-10 and 3-J-11, separated by age. 

3-J-10 

3-J-11 

FC 
Present 
13 
16 
5 
12 
18 
5 
69 

FC 
Absent 
6 
6 
0 
2 
6 
1 
21 

Total  % 

19 
22 
5 
14 
24 
6 
90 

68% 
73% 
100% 
86% 
75% 
83% 
77% 

FC 
Present 
16 
60 
17 
25 
27 
9 
154 

FC 
Absent 
9 
6 
0 
7 
11 
12 
45 

Total  % 

25 
66 
17 
32 
38 
21 
199 

64% 
91% 
100% 
78% 
71% 
43% 
77% 

Infant 
Child 
Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Total 

Table 6.11: Prevalence of spinal porosity (SP) in cemeteries 3-J-10 and 3-J-11, separated by age. 

SP 
Present 
6 
13 
5 
9 
4 
0 
37 

Infant 
Child 
Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Total 

PR 
Present 
14 
12 
2 
9 
24 
5 
66 

Infant 
Child 
Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Total 

3-J-10 
SP 
Absent 
12 
8 
0 
4 
29 
9 
62 

3-J-10 
PR 
Absent 
5 
10 
3 
4 
11 
7 
40 

Total  % 

Total  % 

SP 
Present 
9 
33 
18 
29 
28 
10 
127 

3-J-11 
SP 
Absent 
17 
27 
1 
4 
21 
20 
90 

26 
35% 
60 
55% 
19 
95% 
33 
88% 
49 
57% 
33% 
30 
217  59% 

18 
21 
5 
13 
33 
9 
99 

33% 
62% 
100% 
69% 
12% 
0% 
37% 

Total  % 

Total  % 

PR 
Present 
18 
28 
10 
20 
31 
21 
128 

3-J-11 
PR 
Absent 
10 
36 
9 
12 
18 
10 
95 

64% 
28 
44% 
64 
53% 
19 
63% 
32 
63% 
49 
68% 
31 
223  57% 

74% 
19 
55% 
22 
40% 
5 
69% 
13 
69% 
35 
42% 
12 
106  62% 

Table 6.12: Prevalence of periosteal reaction (PR) in cemeteries 3-J-10 and 3-J-11, separated by 
age. 

To examine possible differences of hypothesized malaria between males and females 

from cemeteries 3-J-10 and 3-J-11, only individuals whose remains could be estimated for 

biological sex were included (n = 49 for 3-J-10, n = 101 for 3-J-11). A chi-square test between 

males from cemetery 3-J-10 and males from cemetery 3-J-11 yielded no significant difference 

92 

 
 
 
 
 
 
 
 
 
 
 
for inferred malaria (Table 6.13). Additionally, similar frequencies of “positive diagnoses” were 

found in both subgroups (Table 6.13). While a Fisher’s exact test between female individuals 

from 3-J-10 and 3-J-11 did not result in a significant difference, the p-value approached 

statistical significance (Table 6.13). In comparing the frequencies of “positive diagnoses” for 

malaria between females from both cemeteries, a higher frequency of 79% was found among 

female individuals from 3-J-11 compared to the frequency of 58% observed among females from 

3-J-10 (Table 6.13). The outcome of a chi-square test between both males and females from 3-J-

10 and males and females from 3-J-11 did not result in a significant difference (Table 6.13).  

Table 6.13: Prevalence of “positive diagnoses” (P.D.) and “negative diagnoses” (N.D.) between 
cemeteries, separated by sex, and results of chi-square (df = 1) and Fisher’s exact tests. 

3-J-10 

3-J-11 

Male 

P.D.  N.D.  Total 
9 
16 

25 

P.D.% 
64% 

P.D.  N.D.  Total 
15 
28 

43 

P.D.% 
65% 

Female 

14 

10 

24 

58% 

46 

12 

58 

79% 

Total 

30 

19 

49 

61% 

74 

27 

101 

73% 

3-J-10 vs. 
3-J-11 
χ2 = 0.0086 
p = 0.93  
FET 
p = 0.061 
φ = -0.22 
χ2 = 2.25 
p = 0.13 

The prevalence of the five lesions between males from cemeteries 3-J-10 and 3-J-11, and 

between females from both cemeteries, was assessed using chi-square or Fisher’s exact tests as 

appropriate. Of the five lesions, spinal porosity was found to be significantly different between 

males from cemetery 3-J-10 and males from 3-J-11, and between females from cemetery 3-J-10 

and 3-J-11 (Table 6.17). In examining the frequency of spinal porosity in each group, males from 

cemetery 3-J-11 displayed a higher prevalence of the lesion (68%) than males from cemetery 3-

J-10 (29%), and a higher frequency of the lesion was observed among females from cemetery 3-

J-11 (55%) than females from cemetery 3-J-10 (19%) (Table 6.17). A significantly higher 

prevalence of spinal porosity among individuals from cemetery 3-J-11 than cemetery 3-J-10 was 

93 

 
 
 
 
 
also found when males and females were combined within their respective cemeteries and 

compared (Table 6.17). The associated phi coefficients for these comparisons indicate moderate-

strong effects (Table 6.17).  

No significant differences were found in comparing males from cemetery 3-J-10 and 

males from 3-J-11, and between females from cemetery 3-J-10 and 3-J-11 for cribra orbitalia, 

humeral cribra, femoral cribra, and periosteal reaction (Tables 6.14, 6.15, 6.16, and 6.18, 

respectively). When males and females from cemetery 3-J-10 and males and females from 

cemetery 3-J-11 were compared, a significantly higher prevalence of cribra orbitalia was found 

among individuals from cemetery 3-J-11 compared to individuals from cemetery 3-J-10 (Table 

6.14). However, the phi coefficient associated with this result suggests a weak relationship. 

Table 6.14: Prevalence of cribra orbitalia (CO) between cemeteries 3-J-10 and 3-J-11, separated 
by sex, and results of chi-square tests (df = 1). Phi coefficient (φ) provided for significant 
outcome (bolded p-value). 

CO 
Present 

3-J-10 
CO 
Absent 

Total  % 

CO 
Present 

3-J-11 
CO 
Absent 

Total  % 

Male 

Female 

8 

9 

20 

16 

28 

29% 

22 

25 

36% 

33 

31 

28 

53 

42% 

61 

54% 

Total 

17 

36 

53 

32% 

55 

59 

114  48% 

3-J-10 vs. 
3-J-11 

χ2 = 1.32 
p = 0.25 
χ2 = 2.32 
p = 0.13 
χ2 = 3.86 
p = 0.050 
φ = -0.15 

Table 6.15: Prevalence of humeral cribra (HC) between cemeteries 3-J-10 and 3-J-11, separated 
by sex, and results of Fisher’s exact and chi-square tests (df = 1).  

HC 
Present 

3-J-10 
HC 
Absent 

Total  % 

HC 
Present 

3-J-11 
HC 
Absent 

Total  % 

Male 

Female 

Total 

4 

1 

5 

17 

13 

30 

21 

19% 

14 

7% 

7 

9 

35 

14% 

16 

28 

37 

65 

35 

20% 

46 

20% 

81 

20% 

3-J-10 vs. 
3-J-11 

FET 
p = 1 
FET 
p = 0.43 
χ2 = 0.49 
p = 0.48 

94 

 
 
 
 
 
 
 
Table 6.16: Prevalence of femoral cribra (FC) between cemeteries 3-J-10 and 3-J-11 separated by 
sex and results of Fisher’s exact and chi-square tests (df = 1).  
3-J-10 
FC 
Absent 

3-J-11 
FC 
Absent 

3-J-10 vs. 
3-J-11 

FC 
Present 

FC 
Present 

Total  % 

Total  % 

Male 

Female 

Total 

18 

18 

36 

6 

3 

9 

24 

75% 

20 

21 

86% 

44 

45 

80% 

64 

16 

14 

30 

36 

56% 

58 

76% 

94 

68% 

χ2 = 2.34 
p = 0.13 
FET 
p = 0.54 
χ2 = 2.14 
p = 0.14 

Table 6.17: Prevalence of spinal porosity (SP) between cemeteries 3-J-10 and 3-J-11 separated 
by sex and results of chi-square tests (df = 1). Phi coefficients (φ) provided for significant 
outcomes (bolded p-values). 

SP 
Present 

3-J-10 
SP 
Absent 

Total  % 

SP 
Present 

3-J-11 
SP 
Absent 

Total  % 

Male 

Female 

8 

5 

20 

28 

29% 

36 

17 

53 

68% 

21 

26 

19% 

34 

28 

62 

55% 

Total 

13 

41 

54 

24% 

70 

45 

115  61% 

3-J-10 vs. 
3-J-11 

χ2 = 11.44 
p = 0.00072 
φ = -0.38 
χ2 = 9.41 
p = 0.0022 
φ = -0.33 
χ2 = 19.91 
p = 8.13e-06 
φ = -0.34 

Table 6.18: Prevalence of periosteal reaction (PR) between cemeteries 3-J-10 and 3-J-11 
separated by sex and results of chi-square tests (df = 1).  

PR 
Present 

3-J-10 
PR 
Absent 

Total  % 

PR 
Present 

3-J-11 
PR 
Absent 

Total  % 

Male 

Female 

Total 

22 

18 

40 

10 

13 

23 

32 

69% 

35 

31 

58% 

39 

63 

63% 

74 

18 

23 

41 

53 

66% 

62 

63% 

115  64% 

3-J-10 vs. 
3-J-11 

χ2 = 0.066 
p = 0.80 
χ2 = 0.20 
p = 0.65 
χ2 = 0.013 
p = 0.91 

95 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
Research Question 3 

Research Question 3: Are there differences in the frequency of hypothesized malaria 

between sex and age cohorts at Mis Island? 

As there were no significant differences between males from cemeteries 3-J-10 and 3-J-

11, and no significant differences between females from both cemeteries, males and females 

from the two cemeteries were combined for subsequent analyses. A chi-square test between 

males and females at Mis Island for the hypothesized presence of malaria yielded no significant 

difference (Table 6.19). Comparisons between males and females separated by age group also 

did not result in significant differences (Table 6.19). While the frequencies of “positive 

diagnoses” tended to be higher among females overall and across age groups, the frequencies 

between both males and females were relatively similar except for the older adult age group 

which involved a small subsample for males. 

Table 6.19: Prevalence of “positive diagnoses” (P.D.) and “negative diagnoses” (N.D.) between 
males and females, separated by age group, and results of Fisher’s exact (FET) and chi-square 
tests (df = 1).  

Male 

Female 

M vs. F 

P.D.  N.D.  Total  P.D.%  P.D.  N.D.  Total  P.D.% 

2 

19 

21 

2 

0 

0 

3 

16 

5 

0 

2 

100% 

2 

22 

37 

7 

– 

86% 

57% 

29% 

19 

22 

16 

– 

1 

0 

2 

11 

9 

0 

2 

100% 

21 

33 

25 

90% 

67% 

64% 

FET 
p = 1 
FET 
p = 1 
χ2 = 0.72 
p = 0.40 
FET 
p = 0.19 

– 

– 

– 

Age Group 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 
Unknown 
Age Adult  
(≥20 yrs) 

Total 

44 

24 

68 

65% 

60 

22 

82 

73% 

χ2 = 1.25 
p = 0.26 

In comparing the prevalence of the five lesions between males and females, chi-square 

tests yielded no significant differences except for femoral cribra (Table 6.20). The overall 

96 

 
 
 
 
prevalence of femoral cribra observed among female individuals (78%) was greater than that 

found among male individuals (63%), which was a pattern also observed across the adult age 

groups (Table 6.23). However, the phi coefficient associated with this result suggests a relatively 

weak relationship between males and females for femoral cribra (Table 6.20). The frequencies of 

cribra orbitalia, humeral cribra, spinal porosity, and periosteal reaction were relatively similar 

between males and females (Tables 6.21, 6.22, 6.24, and 6.25). 

Table 6.20: Results of chi-square tests (df = 1) between males and females for each lesion and 
phi coefficients for significant results (bolded p-values). 

Cribra orbitalia 

Humeral cribra 

Femoral cribra 

Spinal porosity 

Periosteal reaction 

χ2 = 2.3682 
p = 0.1238 
χ2 = 0.17305 
p = 0.6774 
χ2 = 3.8763 
p = 0.04897 
φ = -0.17 
χ2 = 1.6886 
p = 0.1938 
χ2 = 0.64175 
p = 0.4231 

Table 6.21: Prevalence of cribra orbitalia (CO) between males and females separated by age. 

Male 

Female 

CO 
Present 
2 
11 
13 
4 

CO 
Absent 
0 
13 
30 
7 

0 

30 

1 

51 

Total 

% 

2 
24 
43 
11 

1 

81 

100% 
46% 
30% 
36% 

0% 

37% 

CO 
Present 
2 
12 
12 
15 

CO 
Absent 
0 
9 
23 
12 

1 

42 

0 

44 

Total 

% 

2 
21 
35 
27 

1 

86 

100% 
57% 
34% 
56% 

100% 

49% 

Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Unknown 
Age Adult 
Total 

97 

 
 
 
 
 
 
 
 
 
 
 
Table 6.22: Prevalence of humeral cribra (HC) between males and females separated by age. 

Male 

Female 

HC 
Present 
1 
8 
1 
1 

HC 
Absent 
0 
10 
31 
4 

0 

11 

0 

45 

Total 

% 

1 
18 
32 
5 

0 

56 

100% 
44% 
3% 
20% 

0% 

20% 

HC 
Present 
0 
7 
2 
1 

HC 
Absent 
1 
7 
25 
17 

0 

10 

0 

50 

Total 

% 

1 
14 
27 
18 

0 

60 

0% 
50% 
7% 
6% 

0% 

17% 

Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Unknown 
Age Adult 
Total 

Table 6.23: Prevalence of femoral cribra (FC) between males and females separated by age. 

Male 

Female 

FC 
Present 
1 
17 
19 
1 

FC 
Absent 
0 
6 
10 
6 

0 

38 

0 

22 

Total 

% 

1 
23 
29 
7 

0 

60 

100% 
74% 
66% 
14% 

0% 

63% 

FC 
Present 
2 
20 
25 
13 

FC 
Absent 
0 
3 
7 
7 

2 

62 

0 

17 

Total 

% 

2 
23 
32 
20 

2 

79 

100% 
87% 
78% 
65% 

100% 

78% 

Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Unknown 
Age Adult 
Total 

Table 6.24: Prevalence of spinal porosity (SP) between males and females separated by age. 

Male 

Female 

SP 
Present 
3 
18 
17 
6 

SP 
Absent 
0 
5 
26 
6 

0 

44 

0 

37 

Total 

% 

3 
23 
43 
12 

0 

81 

100% 
78% 
40% 
50% 

0% 

54% 

SP 
Present 
2 
19 
14 
4 

SP 
Absent 
0 
3 
23 
23 

0 

39 

0 

49 

Total 

% 

2 
22 
37 
27 

0 

88 

100% 
86% 
38% 
15% 

0% 

44% 

Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Unknown 
Age Adult 
Total 

98 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 6.25: Prevalence of periosteal reaction (PR) between males and females separated by age. 

Male 

Female 

PR 
Present 
1 
17 
31 
8 

PR 
Absent 
1 
7 
15 
5 

0 

57 

0 

28 

Total 

% 

2 
24 
46 
13 

0 

85 

50% 
71% 
67% 
62% 

0% 

67% 

PR 
Present 
1 
12 
23 
18 

PR 
Absent 
1 
9 
13 
12 

3 

57 

1 

36 

Total 

% 

2 
21 
36 
30 

4 

93 

50% 
57% 
64% 
60% 

75% 

61% 

Adolescent 
Young Adult 
Middle Adult 
Older Adult 
Unknown 
Age Adult 
Total 

A significant association was found between estimated prevalence of malaria and age 

group using a chi-square test χ2 (5, n = 294) = 23.903, p-value = 0.00023, with a relatively 

moderate strength of association (V = 0.29). The highest frequency of “positive diagnoses” for 

malaria was observed in the adolescent age group (96%), followed by the young adult (89%) and 

child (81%) cohorts, and the infant age group (74%). The middle adult and older adult cohorts 

presented the lowest frequencies of hypothesized malarial infection (63% and 56%, 

respectively); however, these prevalences were more than half of the individuals in each group. 

Based on these results, a general pattern was observed in the prevalence of inferred malaria that 

increased with age to the adolescent cohort, followed by a high prevalence in the young adult 

group, and decreased with age in the middle and older adult groups. 

Table 6.26: Prevalence of “positive diagnoses” (P.D.) and “negative diagnoses” (N.D.)  between 
age groups. 

Age 
Infant (<3 years) 
Child (3–11 years) 
Adolescent (12–19 years) 
Young Adult (20–34 years) 
Middle Adult (35–49 years) 
Older Adult (≥50 years) 
Total 

P.D.  N.D.  Total 
11 
31 
15 
64 
1 
24 
39 
5 
27 
45 
14 
18 
73 
221 

42 
79 
25 
44 
72 
32 
294 

P.D.% 
74% 
81% 
96% 
89% 
63% 
56% 
75% 

99 

 
 
 
 
 
 
Research Question 4a & 4b 

Research Question 4a: Are the skeletal lesions suggestive of malaria significantly 

associated with age? 

The results presented in this section focus on the relationship between the skeletal lesions 

suggestive of malaria and age of the individual. The prevalence of cribra orbitalia, humeral 

cribra, femoral cribra, spinal porosity, and periosteal reaction by age group is shown Table 6.27, 

as well as the results of chi-square tests and Cramer’s V statistics. In testing the association 

between each lesion and age group, significant associations were indicated for cribra orbitalia, 

humeral cribra, femoral cribra, and spinal porosity. Additionally, the related Cramer’s V 

measures suggest that the associations between each of these four lesions and age are relatively 

strong. In examining the prevalence of the four lesions across age groups (Table 6.27 and Figure 

6.2), humeral cribra and spinal porosity demonstrated their highest frequencies in the child, 

adolescent, and young adult age groups and lower frequencies for the infant, middle adult, and 

older adult age groups. Femoral cribra also presented at higher frequencies in the child and 

adolescent age groups, followed by young and middle adult groups, and exhibited its lowest 

frequencies in the infant and older adult groups. Cribra orbitalia showed similarly higher 

frequencies in the infant, child, and adolescent age groups, followed by young adult and older 

adult groups, and its lowest frequency in the middle adult age group. No significant association 

was found between periosteal reaction and age of the individual. The prevalence of periosteal 

reaction across age groups demonstrated a pattern generally opposite of that observed for the 

other lesions, with higher frequencies in the infant and adult age groups and lower frequencies in 

the child and adolescent age groups. 

100 

 
 
 
 
Table 6.27: Lesion prevalence by age group and results of chi-square tests (df = 5) and Cramer’s V statistic (bolded p-values represent 
statistically significant results). 

CO 

HC 

FC 

SP 

PR 

.
t
c
e
f
f

n  %  A
37  81%  15 
65  82%  48 
22  82%  11 
47  53%  15 
79  33%  3 
38  50%  2 

Age 
Group  A
30 
Infant 
53 
Child 
18 
Adol. 
25 
YA 
26 
MA 
19 
OA 
171  288  — 
Total 
χ2 = 50.12 
Chi-
square 
p = 1.31e-09 
Cramer’s 
V 

V = 0.42 

.
t
c
e
f
f

.
t
c
e
f
f

.
t
c
e
f
f

.
t
c
e
f
f

%  A
n  %  A
n 
15 
66% 
44 
42  36%  29 
88 
46 
86% 
80  60%  76 
37  100%  23 
19  58%  37 
38 
71% 
31 
32  47%  22 
32  82 
73% 
62 
45 
5% 
59 
10 
52% 
27 
14 
9% 
23 
164  316  — 
255  289  — 
94  255  — 
χ2 = 58.52 
χ2 = 29.57 
χ2 = 56.85 
p = 2.46e-11 
p = 1.79e-05 
p = 5.42e-11 

n  %  A
44  34%  32 
81  57%  40 
24  96%  12 
46  83%  29 
39%  55 
39  26%  26 

n  % 
47  68% 
86  47% 
24  50% 
45  64% 
84  65% 
43  60% 
194  329  — 
χ2 = 10.00 
p = 0.075 

V = 0.47 

V = 0.32 

V = 0.43 

V = 0.17 

101 

 
 
 
 
 
 
 
Figure 6.2: Plot of lesion frequency by age group for cribra orbitalia (CO), humeral cribra (HC), 
femoral cribra (FC), spinal porosity (SP), and periosteal reaction (PR). 

The results of the multiple correspondence analysis involving the presence and absence 

of lesions and age groups are presented in the biplot below (Figure 6.3), which accounts for a 

total of 26.79% of variation. The first and second dimensions presented have eigenvalues of 

0.367 and 0.302, respectively, and are representative of 14.7% and 12.1% of variation, 

respectively. The presence of humeral cribra, spinal porosity, femoral cribra, and cribra orbitalia, 

as well as the absence of periosteal reaction, clustered closer to the child, adolescent, and young 

adult age groups. Conversely, the presence of periosteal reaction and absence of humeral cribra, 

spinal porosity, femoral cribra, and cribra orbitalia clustered closer to the infant, middle adult, 

and older adult age groups. Regarding the relationship among the lesions, the presence of 

humeral cribra, spinal porosity, femoral cribra, and cribra orbitalia clustered closer together, 

whereas the presence of periosteal reaction was separated from these lesions by the second axis. 

102 

 
 
 
Figure 6.3: Multiple correspondence analysis biplot for age groups and the presence/absence of 
cribra orbitalia, humeral cribra, femoral cribra, spinal porosity, and periosteal reaction. 

Research Question 4b: Are there significant associations among the paleopathological 

lesions suggestive of malaria? 

The results of this section present findings on the relationship among the skeletal lesions 

suggestive of malaria. Table 6.28 displays the outcomes of testing the association between pairs 

of lesions for all individuals, regardless of age, with the appropriate skeletal elements present. 

Significant associations were found between the following pairs: cribra orbitalia and humeral 

cribra, cribra orbitalia and femoral cribra, humeral cribra and femoral cribra, humeral cribra and 

spinal porosity, femoral cribra and spinal porosity, and humeral cribra and periosteal reaction. All 

significant associations were positive except between humeral cribra and periosteal reaction 

which were inversely related. The association between humeral cribra and femoral cribra was 

relatively strong, and the association between humeral cribra and cribra orbitalia was generally 

103 

 
 
 
 
moderate. The relationship between femoral cribra and cribra orbitalia, humeral cribra and spinal 

porosity, femoral cribra and spinal porosity, and humeral cribra and periosteal reaction were 

relatively weaker and reflected a small to medium level of association.  

Table 6.28: Association between pairs of lesions for all individuals with the appropriate skeletal 
elements present; the upper values are the results of chi-square tests (df = 1) and the phi 
coefficients, and the lower values are the subsample sizes. 

CO 

CO 

– 

HC 
χ2 = 11.18 
p = 0.00083 
φ = 0.23 

HC 

n = 209 

– 

FC 
χ2 = 5.88 
p = 0.015 
φ = 0.16 
χ2 = 24.47 
p = 7.53e-07 
φ = 0.32 

FC 

n = 234 

n = 236 

– 

SP 
χ2 = 3.15 
p = 0.076 
φ = 0.11 
χ2 = 8.67 
p = 0.0032 
φ = 0.19 
χ2 = 6.77 
p = 0.0093 
φ = 0.16 

SP 

n = 258 

n = 241 

n = 266 

– 

PR 
χ2 = 0.88 
p = 0.35 
φ = 0.06 
χ2 = 5.07 
p = 0.024 
φ = -0.14 
χ2 = 2.79 
p = 0.095 
φ = -0.10 
χ2 = 1.29 
p = 0.26 
φ = -0.07 

PR 

n = 268 

n = 244 

n = 274 

n = 295 

– 

The co-occurrence between pairs of lesions was then examined separately in subadult and 

adult subsamples. In assessing these relationships in the subadult age groups, the following pairs 

were significantly associated (Table 6.29): cribra orbitalia and femoral cribra, humeral cribra and 

femoral cribra, humeral cribra and spinal porosity, femoral cribra and spinal porosity, femoral 

cribra and periosteal reaction, and spinal porosity and periosteal reaction. These significant 

associations were all positive except the associations between femoral cribra and periosteal 

reaction, and spinal porosity and periosteal reaction which were inversely related. Relatively 

large measures of association were found between humeral cribra and femoral cribra, and 

between femoral cribra and spinal porosity. The relationship between femoral cribra and cribra 

orbitalia, humeral cribra and spinal porosity, femoral cribra and periosteal reaction, and spinal 

104 

 
 
 
 
 
porosity and periosteal reaction displayed generally medium levels of association. When the 

lesions’ co-occurrence was examined among adult individuals only, no significant associations 

between the pairs of lesions were found (Table 6.30). 

Table 6.29: Association between pairs of lesions for subadult individuals with the appropriate 
skeletal elements present; the upper values are the results of the Fisher’s exact tests (FET) and 
chi-square tests (df = 1) and the phi coefficients, and the lower values are the subsample sizes. 

CO 

CO 

– 

HC 
FET 
p = 0.33 
φ = 0.11 

HC 

n = 104 

– 

FC 
FET 
p = 0.035 
φ = 0.21 
χ2 = 24.35 
p = 8.05e-07 
φ = 0.42 

FC 

n = 114 

n = 137 

– 

SP 
FET 
p = 0.46 
φ = 0.07 
χ2 = 4.93 
p = 0.026 
φ = 0.19 
χ2 = 13.67 
p = 0.00022 
φ = 0.31 

SP 

n = 114 

n = 132 

n = 140 

– 

PR 
FET 
p = 0.81 
φ = 0.02 
χ2 = 3.19 
p = 0.074 
φ = -0.15 
χ2 = 4.99 
p = 0.026 
φ = -0.18 
χ2 = 4.10 
p = 0.043 
φ = -0.17 

PR 

n = 117 

n = 136 

n = 148 

n = 144 

– 

Table 6.30: Association between pairs of lesions for adult individuals with the appropriate 
skeletal elements present; the upper values are the results of the Fisher’s exact tests (FET) and 
chi-square tests (df = 1) and the phi coefficients, and the lower values are the subsample sizes. 

CO 

CO 

– 

HC 
FET 
p = 0.44 
φ = 0.09 

HC 

n = 105 

– 

FC 
χ2 = 0.60 
p = 0.44 
φ = 0.07 
FET 
p = 0.26 
φ = 0.14 

FC 

n = 120 

n = 99 

– 

SP 

n = 144 

n = 109 

n = 126 

SP 
χ2 = 1.44 
p = 0.23 
φ = 0.10 
FET 
p = 0.069 
φ = 0.19 
χ2 = 0.012 
p = 0.91 
φ = 0.01 

– 

PR 
χ2 = 3.68 
p = 0.055 
φ = 0.16 
FET 
p = 1 
φ = -0.02 
χ2 = 0.043 
p = 0.84 
φ = 0.02 
χ2 = 0.57 
p = 0.45 
φ = 0.06 

PR 

n = 151 

n = 108 

n = 126 

n = 151 

– 

105 

 
 
 
 
 
 
Table 6.31 displays the co-occurrence between the three cribrous lesions that have been 

considered part of a “cribrous syndrome” (Miquel-Feucht et al., 1999) in all individuals, as well 

as separated by subadults and adults. The combined presence of cribra orbitalia, humeral cribra, 

and femoral cribra was observed in 27% of all individuals with the skeletal elements present to 

assess the three lesions. When this co-occurrence was examined among subadults and adults 

separately, subadult individuals demonstrated a substantially higher frequency (44%) compared 

to adult individuals (8%). The prevalence of pairings between the three cribrous lesions revealed 

a similar pattern, in which pairs of lesions co-occurred at much higher frequencies in the 

subadult individuals than the adult individuals. Of the three pairings, the most commonly 

observed co-occurrence in both subadults and adults was cribra orbitalia and femoral cribra, 

followed by humeral cribra and femoral cribra, and lastly cribra orbitalia and humeral cribra. 

Table 6.31: Co-occurrence between cribra orbitalia (CO), humeral cribra (HC), and femoral 
cribra (FC) among all individuals, and separated by subadult and adult individuals. 
All Individuals 

Subadults 

Adults 

Combination of  
cribrous lesions 
CO + HC 
CO + FC 
HC + FC 
CO + HC + FC 

26% (54/209) 
49% (115/234) 
36% (86/236) 
27% (51/192) 

43% (45/104) 
68% (78/114) 
51% (70/137) 
44% (44/101) 

9% (9/105) 
31% (37/120) 
16% (16/99) 
8% (7/91) 

Research Question 5a & 5b 

Research Question 5a: Are the macroscopic variables observed on the external surface 

of femoral cribra associated with age? 

The results of this section examine the relationship between the age of the individual and 

variables reflecting the variation in appearance of femoral cribra from the lesion surface. As 

shown in Table 6.32, significant associations were found between age and the lesion variables 

Radi et al. score, estimated activity, and presence of a “cortical window” for both left and right 

femora. Furthermore, all associations for both left and right femora were relatively strong, 

106 

 
 
 
particularly those between age and estimated activity and age and the presence of a cortical 

window. 

Table 6.32: Results of Fisher’s exact tests and Cramer’s V statistics in examining the association 
between age and Radi et al. score, estimated activity, and cortical window (note: p-values 
between age and estimated activity are simulated based on 2,000 replicates). 

Comparison 

Age and Radi et al. Score 

Age and Estimated Activity 

Age and Cortical Window 

Left femora  Right femora 
n = 203 
p = 0.0010 
V = 0.33 
n = 168 
p = 0.00050 
V = 0.55 
n = 182 
p = 3.74e-08 
V = 0.53 

n = 175 
p = 0.00025 
V = 0.40 
n = 168 
p = 0.00050 
V = 0.54 
n = 159 
p = 1.637e-08 
V = 0.53 

Tables 6.33 and 6.34 display the prevalence of Radi et al. scores for femoral cribra across 

age groups for left and right femora, respectively. In examining the relationship between age and 

Radi et al. scores, the majority of femora in all age groups demonstrated a Score 2 for both left 

and right femora. This majority reached 77–100% for all age groups except the infant cohort, in 

which a Score 2 accounted for 58% and 57% of the observed left and right femoral cribra. The 

infant age group consequently demonstrated the most cases of Radi et al. Score 1 which was 

observed in 42% and 43% of left and right femora within the age cohort.  

107 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 6.33: Prevalence of femoral cribra lesions evaluated as Radi et al. Scores 1 and 2 separated 
by age group for left femora. 

a
r
o
m
e
f

t
f
e
L

Age group 

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

n   Number of 
Radi et al. 
Score 1 
11 

26 

Number of 
Radi et al. 
Score 2 
15 

73 

21 

31 

41 

11 

10 

0 

7 

3 

1 

63 

21 

24 

38 

10 

% of Radi et 
al. Score 1 

% of Radi et 
al. Score 2  

42% 

14% 

0% 

23% 

7% 

9% 

58% 

86% 

100% 

77% 

93% 

91% 

Table 6.34: Prevalence of femoral cribra lesions evaluated as Radi et al. Scores 1 and 2 separated 
by age group for right femora. 

a
r
o
m
e
f

t
h
g
i
R

Age group 

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

n   Number of 
Radi et al. 
Score 1 
9 

21 

Number of 
Radi et al. 
Score 2 
12 

67 

22 

24 

30 

11 

2 

1 

3 

3 

1 

65 

21 

21 

27 

10 

% of Radi et 
al. Score 1 

% of Radi et 
al. Score 2  

43% 

3% 

5% 

13% 

10% 

9% 

57% 

97% 

95% 

88% 

90% 

91% 

As shown in Tables 6.35 and 6.36, lesions estimated to be active or healing demonstrated 

a consistent trend with the age of the individual. Overall, for both left and right femora, the 

frequency of lesions believed to be active decreased with age, whereas the frequency of lesions 

that demonstrated signs of possible healing increased with age. In the infant, child, and 

adolescent age groups, the majority of femoral cribra lesions were assessed as active.  In the 

108 

 
 
 
 
 
 
 
 
young, middle, and older adult groups, the majority of femoral cribra lesions displayed signs of 

healing. 

Table 6.35: Prevalence of femoral cribra lesions scored as active and healing separated by age 
group for left femora. 
Age group 

% of estimated 
active 

% of estimated 
healing 

n  Number of 
estimated 
active 
21 

26 

Number of 
estimated 
healing 
5 

71 

21 

29 

36 

9 

60 

14 

11 

7 

2 

11 

7 

18 

29 

7 

n  Number of 
estimated 
active 
18 

21 

Number of 
estimated 
healing 
3 

66 

22 

22 

28 

9 

56 

17 

10 

8 

1 

10 

5 

12 

20 

8 

a
r
o
m
e
f

t
f
e
L

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

a
r
o
m
e
f

t
h
g
i
R

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

81% 

85% 

67% 

38% 

19% 

22% 

19% 

15% 

33% 

62% 

81% 

78% 

86% 

85% 

77% 

45% 

29% 

11% 

14% 

15% 

23% 

55% 

71% 

89% 

Table 6.36: Prevalence of femoral cribra lesions scored as active and healing separated by age 
group for right femora. 
Age group 

% of estimated 
active 

% of estimated 
healing 

The presence of a cortical window demonstrated a distinct pattern between subadult and 

adult age groups (Tables 6.37 and 6.38). This feature was virtually absent in the infant, child, and 

adolescent cohorts, with only one femur displaying a cortical window from those groups. In the 

109 

 
 
 
 
 
 
 
young, middle, and older adult age groups however, cortical windows were present and increased 

in frequency with age. The only age group in which cortical windows were observed in more 

than half of the affected femora was in the older adult cohort, although the subsample for this age 

group was small. 

Table 6.37: Prevalence of femoral cribra lesions with and without a cortical window separated by 
age group for left femora. 
Age group 

n  Window 
present 
0 

26 

Window 
absent 
26 

% of window 
 present 
0% 

% of window  
absent 
100% 

a
r
o
m
e
f

t
f
e
L

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

71 

21 

28 

29 

7 

1 

0 

4 

8 

5 

70 

21 

24 

21 

2 

1% 

0% 

14% 

28% 

71% 

99% 

100% 

86% 

72% 

29% 

Table 6.38: Prevalence of femoral cribra lesions with and without a cortical window separated by 
age group for right femora. 

a
r
o
m
e
f

t
h
g
i
R

Age group 

Infant  
(<3 yrs) 
Child  
(3–11 yrs) 
Adolescent 
(12–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

n  Window 
present 
0 

21 

Window 
absent 
21 

% of window 
 present 
0% 

% of window  
absent 
100% 

66 

22 

19 

24 

7 

0 

0 

5 

7 

4 

66 

22 

14 

17 

3 

0% 

0% 

26% 

29% 

57% 

100% 

100% 

74% 

71% 

43% 

The results of the multiple correspondence analysis (MCA) involving age of the 

individual, estimated lesion activity, presence/absence of a cortical window, and Radi et al. Score 

110 

 
 
 
 
 
 
 
 
for left femora and right femora with femoral cribra are presented in Figures 6.4 and 6.5, 

respectively. The MCA for affected left femora accounts for a total of 33.59% of variation. The 

first and second dimensions presented have eigenvalues of 0.542 and 0.382 and are 

representative of 19.7% and 13.9% of variation, respectively. The MCA for affected right femora 

accounts for a total of 35.25% of variation. The first and second dimensions presented have 

eigenvalues of 0.561 and 0.408 and are representative of 20.4% and 14.8% of variation, 

respectively. In the biplots for both left and right femora, Radi et al. Score 2 is centered near the 

origin, which reflects the relative ubiquity of this variable’s expression in the sample and 

indicates that it is not a differentiating feature. Radi et al. Score 1 tended to cluster closer to the 

subadult age groups, particularly for left femora, and further away from adult age groups. 

Additionally, Radi et al. Score 1 was more distant from the lesion variables of estimated healing 

and presence of a cortical window. Active lesions and the absence of a cortical window clustered 

together and were closest to the infant, child, and adolescent age groups and further away from 

the adult cohorts. Conversely, lesions demonstrating signs of healing and the presence of a 

cortical window clustered closest to the middle and older adult groups, with the young adult 

cohort near these variables for right femora.  

111 

 
 
 
Figure 6.4: MCA biplot for age group and the hypothesized activity, presence/absence of a 
cortical window, and Radi et al. Score for left femora with femoral cribra. 

112 

 
 
 
Figure 6.5: MCA biplot for age group and the hypothesized activity, presence/absence of a 
cortical window, and Radi et al. Score for right femora with femoral cribra. 

Figures 6.6 and 6.7 display the measurements of lesion perimeter for femoral cribra 

estimated to be active and healing within each age group for left and right femora. As reflected in 

the previous analyses, the frequency of healing lesions generally increased with age. The 

adolescent age group demonstrated clustering of relatively larger lesions compared to other age 

groups. Additionally, the juvenile and adolescent cohorts had the largest lesion perimeters at the 

lowest ends of their ranges compared to the other age groups. In general, the range of lesion size 

was greater in the adult groups compared to the subadult age cohorts. Within each age group, 

there is overlap in the perimeter of femoral cribra between lesions estimated to be active and 

healing. However, in the adult age groups, the smallest lesions appear to be healing and in the 

113 

 
 
 
 
subadult age groups the smallest lesions include both active and healing lesions with an overall 

greater proportion of active lesions. 

Figure 6.6: Scatter plot of lesion perimeter between lesions hypothesized to be active or healing 
for left femora, separated by age group. 

114 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 6.7: Scatter plot of lesion perimeter between lesions hypothesized to be active or healing 
for right femora, separated by age group. 

Research Question 5b: Are there patterns between macroscopic variables observed on 

the external surface of femoral cribra potentially related to lesion “severity” and “activity”?  

The association between the presence of a cortical window and whether a femoral cribra 

lesion displayed signs of remodeling was examined to explore a possible connection regarding 

hypothesized lesion activity. Fisher’s exact tests between estimated activity of the lesion and the 

presence of a cortical window demonstrated a significant association for both left and right 

femora (p = 0.00041 and p = 2.88e-06, respectively). Phi coefficients indicate that both 

associations are medium to strong with the right side having a stronger association (φ = 0.28 for 

left femora and φ = 0.39 for right femora). Table 6.39 demonstrates relatively similar patterns 

between both variables for left and right femora, with the right femora having fewer cases of 

lesions displaying both signs of remodeling and an absent cortical window. 

115 

 
 
 
Table 6.39: Contingency tables between estimated activity and the presence of a cortical window 
for left and right femora. 

Window Absent  Window Present 

  Active 
t
f
e
L

Healing 

  Active 

t
h
g
i
R

Healing 

109 

53 

105 

38 

4 

14 

2 

14 

The relationship between femoral cribra’s Radi et al. Score and measurements of lesion 

length perpendicular to the femoral neck, length parallel to the femoral neck, and perimeter was 

evaluated in explore these variables’ possible connection to lesion severity. Across age groups, 

lesions with a Radi et al. Score 2 tended to demonstrate greater lengths perpendicular to the 

femoral neck compared to those with a Radi et al. Score 1 for both left and right femora (Table 

6.40 and Figures 6.8 and 6.9). Lesions displaying a Radi et al. Score 2 also demonstrated a 

greater range of measurements perpendicular to the femoral neck.  

Table 6.40: Summary statistics of lesion length perpendicular to the femoral neck between 
lesions categorized as Radi et al. Scores 1 and 2 for left and right femora separated by age group. 

Perpendicular to Neck Measurements (mm) 
Left Femora 

Right Femora 

Score 1  
Mean ± SD 
2.89 (±1.69) 
n = 9 
4.63 (±2.67) 
n = 8 
4 (±0.00) 
n = 1 
— 
n = 0 
5.67 (±3.50) 
n = 6 
6.67 (±2.08) 
n = 3 
4 (±0.00) 
n = 1 

Score 2  
Mean ± SD 
7.79 (±2.64) 
n = 14 
6.12 (±1.98) 
n = 46 
14.58 (±6.06) 
n = 26 
20.80 (±6.38) 
n = 5 
11.91 (±6.32) 
n = 22 
14.67 (±6.74) 
n = 27 
11 (±7.62) 
n = 7 

Score 1  
Mean ± SD 
3.63 (±1.85) 
n = 8 
5 (±0.00) 
n = 1 
7 (±1.41) 
n = 2 
— 
n = 0 
6 (±3.61) 
n = 3 
6 (±2.65) 
n = 3 
5 (±0.00) 
n = 1 

Score 2  
Mean ± SD 
7.75 (±2.05) 
n = 12 
8.32 (±3.70) 
n = 47 
12.04 (±5.85) 
n = 28 
18.83 (±4.96) 
n = 6 
15.63 (±6.48) 
n = 16 
14.16 (±6.04) 
n = 22 
11.63 (±7.19) 
n = 8 

Age group 

Infant  
(<3 yrs) 
Child  
(3–7 yrs) 
Juvenile 
(8–16 yrs) 
Adolescent 
(17–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

116 

 
 
 
 
Figure 6.8: Boxplot of lesion length perpendicular to the femoral neck between lesions 
categorized as Radi et al. Score 1 or 2 for left femora, separated by age group. 

Figure 6.9: Boxplot of lesion length perpendicular to the femoral neck between lesions 
categorized as Radi et al. Score 1 or 2 for right femora, separated by age group. 

117 

 
 
 
 
Lesions with a Radi et al. Score 2 also tended to be larger in the dimension parallel to the 

femoral neck than lesions with a Radi et al. Score 1 in the same age group for both left and right 

femora (Table 6.41 and Figures 6.10 and 6.11). However, these differences were not as 

pronounced as those observed among the lesion measurements perpendicular to the femoral 

neck. Compared to the measurements perpendicular to the femoral neck between Radi et al. 

scores, the measurements parallel to the neck tended to be relatively closer between Radi et al. 

Scores 1 and 2. Additionally, the lesion measurements parallel to the femoral neck generally 

demonstrated smaller ranges for both Radi et al. scores.  

Table 6.41: Summary statistics of lesion length parallel to the femoral neck between lesions 
categorized as Radi et al. Scores 1 and 2 for left and right femora, separated by age group. 

Parallel to Neck Measurements (mm) 

Left femora 

Right femora 

Age group 

Infant  
(<3 yrs) 
Child  
(3–7 yrs) 
Juvenile 
(8–16 yrs) 
Adolescent  
(17–19 yrs) 
Young Adult  
(20–34 yrs) 
Middle Adult  
(35–49 yrs) 
Older Adult  
(≥50 yrs) 

Score 1  
Mean ± SD 
2.44 (±1.01) 
n = 9 
3.25 (±1.16) 
n = 8 
4.00 (±0.00) 
n = 1 
— 
n = 0 
4.92 (±3.14) 
n = 6 
5.67 (±1.53) 
n = 3 
2.00 (±0.00) 
n = 1 

Score 2  
Mean ± SD 
5.00 (±1.36) 
n = 15 
6.11 (±1.98) 
n = 46 
9.19 (±4.37) 
n = 26 
11 (±2.65) 
n = 5 
7.62 (±3.41) 
n = 21 
6.62 (±3.80) 
n = 26 
6.00 (±4.60) 
n = 6 

Score 1  
Mean ± SD 
2.38 (±0.74) 
n = 8 
5 (±0.00) 
n = 1 
5 (±0.00) 
n = 2 
— 
n = 0 
7.00 (±5.20) 
n = 3 
3.33 (±0.58) 
n = 3 
4.5 (±0.00) 
n = 1 

Score 2  
Mean ± SD 
4.92 (±0.79) 
n = 12 
6.53 (±1.93) 
n = 47 
9.57 (±5.53) 
n = 28 
10.6 (±5.37) 
n = 5 
8.65 (±3.66) 
n = 17 
7.85 (±3.91) 
n = 20 
6.67 (±3.67) 
n = 6 

118 

 
 
 
 
 
 
 
 
 
 
Figure 6.10: Boxplot of lesion length parallel to the femoral neck between lesions categorized as 
Radi et al. Score 1 or 2 for left femora, separated by age group. 

Figure 6.11: Boxplot of lesion length parallel to the femoral neck between lesions categorized as 
Radi et al. Score 1 or 2 for right femora, separated by age group. 

119 

 
 
 
 
Lesions with a Radi et al. Score 2 had larger measurements of femoral cribra perimeter 

compared to lesions with a Radi et al. Score 1 across age groups for both left and right femora 

(Table 6.42 and Figures 6.12 and 6.13). Additionally, the perimeter measurements for lesions 

displaying a Radi et al. Score 2 generally demonstrated larger ranges than the perimeter of 

lesions appearing as a Radi et al. Score 1. 

Table 6.42: Summary statistics of lesion perimeter measurements between lesions categorized as 
Radi et al. Scores 1 and 2 for left and right femora, separated by age group. 

Perimeter Measurements (mm) 

Left femora 

Right femora 

Age group  

Score 1  
Mean ± SD  

Score 2  
Mean ± SD  

Score 1  
Mean ± SD 

Score 2  
Mean ± SD 

Infant  
(<3 yrs) 
Child  
(3–7 yrs) 
Juvenile 
(8–16 yrs) 
Adolescent 
(17–19 yrs) 
Young Adult 
(20–34 yrs) 
Middle Adult 
(35–49 yrs) 
Older Adult 
(≥50 yrs) 

8.00 (±3.84) 
n = 9 
12.50 (±3.63) 
n = 8 
14.00 (±0.00) 
n = 1 
— 
n = 0 
16.83 (±7.17) 
n = 6 
21.67 (±2.08) 
n = 3 
10.00 (±0.00) 
n = 1 

19.86 (±5.52) 
n = 14 
23.89 (±7.56) 
n = 46 
40.19 (±15.91) 
n = 26 
53.80 (±16.81) 
n = 5 
34.14 (±15.61) 
n = 21 
36.00 (±16.43) 
n = 25 
26.50 (±19.49) 
n = 6 

8.88 (±3.98) 
n = 8 
14.00 (±0.00) 
n = 1 
27.50 (±10.61) 
n = 2 
— 
n = 0 
20.67 (±15.89) 
n = 3 
12.00 (±4.00) 
n = 3 
13.00 (±0.00) 
n = 1 

19.67 (±5.12) 
n = 12 
23.19 (±7.50) 
n = 47 
35.22 (±14.58) 
n = 27 
46.00 (±14.16) 
n = 5 
43.56 (±15.52) 
n = 16 
38.70 (±15.46) 
n = 20 
34.17 (±17.41) 
n = 6 

120 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 6.12: Boxplot of lesion perimeter measurements between lesions categorized as Radi et 
al. Score 1 or 2 for left femora, separated by age group. 

Figure 6.13: Boxplot of lesion perimeter between lesions scored as Radi et al. Score 1 or 2 for 
right femora, separated age group. 

121 

 
 
 
 
Research Question 6a & 6b 

Research Question 6a: Are there differences in trabecular microarchitecture at the 

anterior-inferior femoral neck between unaffected individuals and affected individuals 

displaying femoral cribra with varying degrees of skeletal remodeling (i.e., “activity”)? 

Unaffected femora vs. affected femora with active or healing lesions 

Figures 6.14, 6.15, 6.16, 6.17, and 6.18 represent grouped bar graphs for the five 

trabecular variables organized by comparative pairs between unaffected individuals and affected 

individuals displaying lesions hypothesized to be either active or healing. Of the 17 total 

comparison pairs, nine were between unaffected individuals and affected individuals with a 

lesion appearing active on the external surface, and eight were between unaffected individuals 

and affected individuals with a lesion demonstrating signs of healing. Tables 6.43, 6.44, 6.45, 

6.46, and 6.47 summarize the number comparisons in which the trabecular measures for the 

affected femora, with either active or healing lesions, were greater or less than those obtained 

from unaffected femora.  

Figure 6.14 displays the values of bone volume density (BV/TV) for each of the 17 pairs. 

In all nine comparisons between unaffected individuals and affected individuals with an 

estimated active lesion, BV/TV in active lesions were less than the unaffected counterparts. Of 

the eight comparisons between unaffected individuals and affected individuals with a lesion 

thought to be healing, five cases demonstrated lower BV/TV values in the healing lesions. The 

other three comparisons between unaffected individuals and individuals with a healing lesion 

showed higher BV/TV values in healing lesions compared to the unaffected counterparts; 

however, the values were very similar in one of those pairs (Pair 1). 

122 

 
 
 
Table 6.43: Number of comparison pairs in which BV/TV values of affected femora, 
hypothesized to have active or healing lesions, were greater or less than unaffected femora. 

BV/TV 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

0/9 

3/8* 

9/9 

5/8 

Unaffected vs. 
Affected Active 
Unaffected vs. 
Affected Healing 

*BV/TV values for Pair 1 femora are similar 

Figure 6.14: Group bar plot for BV/TV by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be active or healing. 

The values for specific bone surface (BS/BV) were compared between unaffected 

individuals and affected individuals with hypothesized active and healing lesions in Figure 6.15. 

Of the nine pairs in which individuals with estimated active lesions were compared to unaffected 

individuals, BS/BV values were greater in the active lesions compared to unaffected femora in 

seven cases. The remaining two pairs demonstrated lower values of BS/BV in active lesions 

123 

 
 
 
 
 
 
compared to femora without lesions. In comparing BS/BV values between unaffected individuals 

and affected individuals with estimated healing lesions, four of the eight pairs demonstrated 

lower values in healing lesions compared to femora without a lesion. Of the remaining pairs, 

BS/BV values were greater in individuals with healing lesions compared to unaffected 

individuals in three cases, and BS/BV values were identical in one comparison pair (Pair 16). 

Table 6.44: Number of comparison pairs in which BS/BV values of affected femora, 
hypothesized to have active or healing lesions, were greater or less than unaffected femora. 

BS/BV 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

7/9 

3/8 

2/9 

4/8 

Unaffected vs. 
Affected Active 
Unaffected vs. 
Affected Healing* 

*BS/BV values for Pair 16 femora are the same 

Figure 6.15: Group bar plot for BS/BV by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be active or healing. 

124 

 
 
 
 
Figure 6.16 presents measures of trabecular thickness (Tb.Th.) for the pairs comparing 

unaffected individuals and affected individuals with active or healing lesions. Six of the nine 

pairs comparing femora with hypothesized active lesions to unaffected femora revealed higher 

values of Tb.Th. in the individuals with active lesions. The remaining three pairs showed lower 

values of Tb.Th. in active lesions compared to unaffected femora. Of the eight pairs comparing 

individuals with estimated healing lesions and their unaffected counterparts, Tb.Th. was greater 

in the affected femora with healing lesions in seven comparisons. Consequently, trabecular 

thickness in individuals with healing lesions was less than unaffected individuals in only one of 

the eight comparison pairs. 

Table 6.45: Number of comparison pairs in which Tb.Th. values of affected femora, 
hypothesized to have active or healing lesions, were greater or less than unaffected femora. 

Tb.Th. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. 
Affected Active 
Unaffected vs. 
Affected Healing 

6/9 

7/8 

3/9 

1/8 

125 

 
 
 
Figure 6.16: Group bar plot for Tb.Th. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be active or healing. 

The mean values of trabecular number (Tb.N.) were compared between unaffected 

individuals and individuals with either active or healing femoral cribra in Figure 6.17. In seven 

of nine comparisons, Tb.N. in affected femora with hypothesized active lesions was less than that 

in femora without the lesion. The remaining two comparisons between unaffected individuals 

and individuals with active lesions revealed higher Tb.N. values in the active lesions. Of the 

eight comparisons between unaffected individuals and individuals with estimated healing 

lesions, half of the pairs showed higher values of Tb.N. in the healing lesions although three 

pairs had relatively similar values. The other half of the eight pairs showed lower values of Tb.N. 

in healing lesions compared to the unaffected femora. 

126 

 
 
 
 
 
Table 6.46: Number of comparison pairs in which Tb.N. values of affected femora, hypothesized 
to have active or healing lesions, were greater or less than unaffected femora. 

Tb.N. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. 
Affected Active 
Unaffected vs. 
Affected Healing 

2/9 

4/8 

7/9 

4/8 

Figure 6.17: Group bar plot for Tb.N. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be active or healing. 

In Figure 6.18, measures of trabecular spacing (Tb.Sp.) are compared between 

individuals without femoral cribra and individuals with hypothesized active or healing femoral 

cribra. Of the nine pairs comparing individuals with active lesions to the corresponding 

unaffected individuals, seven comparisons demonstrated higher Tb.Sp. values in the active 

lesions. In the remaining two comparisons, individuals with active lesions presented lower values 

127 

 
 
 
 
 
 
of Tb.Sp. compared to their unaffected counterparts.  Five of the eight pairs comparing 

individuals with healing femoral cribra to individuals without the lesion showed lower values of 

Tb.Sp. in the healing lesions. In the other three cases, Tb.Sp. in affected femora with healing 

lesions was greater than femora without the lesion. 

Table 6.47: Number of comparison pairs in which Tb.Sp values of affected femora, hypothesized 
to have active or healing lesions, were greater or less than unaffected femora. 

Tb.Sp. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. 
Affected Active 
Unaffected vs. 
Affected Healing 

7/9 

3/8 

2/9 

5/8 

Figure 6.18: Group bar plot for Tp.Sp. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be active or healing. 

128 

 
 
 
 
 
 
 
Affected femora with active lesions vs. affected femora with healing lesions 

Values for each of the five trabecular microarchitecture variables were compared between 

individuals with lesions hypothesized to be active and individuals with lesions thought to be 

healing, the results of which are summarized in Table 6.48. The comparisons across the five 

variables were conducted between seven pairs of individuals with each pair sharing the same 

estimated demographics (i.e., age-at-death and sex). In 6/7 of the comparison pairs, bone volume 

density (BV/TV) was greater in healing lesions compared to active lesions (Figure 6.19). The 

remaining comparison pair (Pair 5) revealed a lower BV/TV for the individual with a healing 

lesion compared to the individual with an active lesion; however, the values were very similar. In 

comparing specific bone surface (BS/BV), 6/7 comparisons showed lower values in healing 

lesions compared to active lesions (Figure 6.20). The other comparison pair demonstrated a 

higher BS/BV value in the healing lesion compared to the active lesion. All seven comparisons 

for trabecular thickness (Tb.Th.) showed higher values in healing lesions compared to active 

lesions (Figure 6.21). Healing lesions had greater mean values of trabecular number (Tb.N.) 

compared to active lesions in 4/7 comparisons, while lesions showing signs of healing had lower 

values of Tb.N. compared to lesions that appeared active in 3/7 comparisons (Figure 6.22). In the 

comparisons for trabecular spacing (Tb.Sp.), 6/7 pairs revealed lower values of Tb.Sp. in healing 

lesions compared to active lesions and only 1/7 pair demonstrated a higher value of Tb.Sp. in the 

healing lesion (Figure 6.23). 

129 

 
 
 
 
 
Table 6.48: Number of comparison pairs in which the trabecular measures of affected femora, 
hypothesized to have active or healing lesions, were greater or less than unaffected femora. 

Healing is greater 
than active 
6/7 
1/7 
7/7 
4/7 
1/7 

Healing is less 
than active 
1/7* 
6/7 
0/7 
3/7 
6/7 

BV/TV 
BS/BV 
Tb.Th. 
Tb.N. 
Tb.Sp. 

*BS/BV values for Pair 5 femora are similar 

Figure 6.19: Group bar plot for BV/TV by comparison pairs of affected femora with lesions 
hypothesized to be active and healing. 

130 

 
 
 
 
 
 
Figure 6.20: Group bar plot for BS/BV by comparison pairs of affected femora with lesions 
hypothesized to be active and healing. 

Figure 6.21: Group bar plot for Tb.Th. by comparison pairs of affected femora with lesions 
hypothesized to be active and healing. 

131 

 
 
 
 
 
Figure 6.22: Group bar plot for Tb.N. by comparison pairs of affected femora with lesions 
hypothesized to be active and healing. 

Figure 6.23: Group bar plot for Tb.Sp. by comparison pairs of affected femora with lesions 
hypothesized to be active and healing. 

132 

 
 
 
 
 
Research Question 6b: Are there differences in trabecular microarchitecture at the 

anterior-inferior femoral neck between unaffected individuals and affected individuals 

displaying varying extents of femoral cribra (i.e., “severity”)? 

Unaffected femora vs. affected femora with “less severe” or “more severe” lesions 

The five trabecular variables were compared between pairs of unaffected individuals and 

affected individuals displaying femoral cribra hypothesized to be either less severe or more 

severe. Each comparison pair was between individuals from the same age cohort. Of the ten 

comparison pairs, five pairs were between unaffected individuals and affected individuals with a 

lesion appearing less severe from the surface, and five pairs were between unaffected individuals 

and affected individuals with a lesion displaying a more severe external surface.  

Figure 6.24 shows the values of bone volume density (BV/TV) compared between 

unaffected individuals and affected individuals with lesions hypothesized to be less or more 

severe, the results of which are summarized in Table 6.49. In 3/5 comparisons between 

unaffected individuals and affected individuals with a less severe lesion, BV/TV values were 

lower in the affected femora with less severe lesions. The other two comparisons, which were 

between pairs of individuals in the youngest age group (2–3 years), demonstrated higher values 

of BV/TV in the affected femora with less severe lesions compared to their unaffected 

counterparts. In comparing BV/TV between unaffected femora and affected femora with more 

severe lesions, 3/5 pairs showed lower values in the femora with the lesion. The other two 

comparisons between unaffected femora and femora with more severe lesions were between 

individuals of the youngest age cohort (2–3 years). One of these pairs demonstrated a higher 

BV/TV in the affected individual compared to the unaffected individual, and the other pair had 

similar values. 

133 

 
 
 
Table 6.49: Number of comparison pairs in which BV/TV values of affected femora, with lesions 
hypothesized to be less severe or more severe, were greater or less than unaffected femora. 

BV/TV 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. Affected  
“less severe” lesion 

Unaffected vs. Affected 
“more severe” lesion* 

2/5 

1/5 

3/5 

3/5 

*BS/BV values for Pair 4 femora are similar 

Figure 6.24: Group bar plot for BV/TV by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be less severe or more severe. 

The values for specific bone surface (BS/BV) were compared between unaffected 

individuals and affected individuals with lesions thought to be less severe or more severe in 

Figure 6.25, the outcomes of which are summarized in Table 6.50. Of the five comparisons 

between unaffected femora and femora with a lesion estimated to be less severe, four pairs 

demonstrated slightly lower values of BS/BV in the affected femora. The other pair, which was 

134 

 
 
 
 
 
 
between individuals from the 4–7 years cohort, displayed a higher value of BS/BV in the femur 

with femoral cribra. Of the five comparisons between unaffected femora and femora with a 

lesion thought to be more severe, the affected femora had lower values of BS/BV in 3/5 pairs. In 

the remaining two comparisons, specific bone surface was greater in the affected femora with 

more severe lesions than in the unaffected femora. 

Table 6.50: Number of comparison pairs in which BS/BV values of affected femora, with lesions 
hypothesized to be less severe or more severe, were greater or less than unaffected femora. 

BS/BV 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. Affected  
“less severe” lesion 

Unaffected vs. Affected 
“more severe” lesion 

1/5 

2/5 

4/5 

3/5 

135 

 
 
 
 
 
Figure 6.25: Group bar plot for BS/BV by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be less severe or more severe. 

The values for trabecular thickness (Tb.Th.) were compared between unaffected 

individuals and individuals with femoral cribra hypothesized to be either less severe or more 

severe in Figure 6.26. As summarized in Table 6.51, 3/5 comparisons between unaffected femora 

and femora with lesions that appeared less severe demonstrated higher values of trabecular 

thickness in the affected femora. Of the remaining two comparisons, one pair had similar values 

(Pair 2, 2–3 years) and the other pair (Pair 9, 4–7 years) showed a lower value for Tb.Th. in the 

affected femur with a less severe lesion. A similar pattern was observed in comparing unaffected 

femora and affected femora with lesions that appeared more severe (Table 6.51). Three of the 

five comparisons between femora without lesions and femora with more severe lesions 

demonstrated greater Tb.Th. values in the affected femora. Of the other two comparisons, one 

136 

 
 
 
 
pair had similar values (Pair 4, 2–3 years) and the other pair showed a lower measure of Tb.Th. 

in the femur with a more severe lesion compared to the unaffected femur. 

Table 6.51: Number of comparison pairs in which Tb.Th. values of affected femora, with lesions 
hypothesized to be less severe or more severe, were greater or less than unaffected femora. 

Tb.Th. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. Affected  
“less severe” lesion 

Unaffected vs. Affected 
“more severe” lesion* 

4/5* 

3/5 

1/5 

1/5 

*Tb.Th. values for Pair 2 femora and for Pair 4 femora are similar 

Figure 6.26: Group bar plot for Tb.Th. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be less severe or more severe. 

Figure 6.27 presents mean values of trabecular number (Tb.N.) for the pairs comparing 

unaffected individuals and affected individuals with lesions thought to be more severe or less 

137 

 
 
 
 
 
 
severe. Of the five comparisons between femora without lesions and femora with less severe 

lesions, three resulted in higher values of trabecular number for the affected femora whereas two 

demonstrated lower values for the affected femora (Table 6.52). In all five comparisons between 

unaffected femora and affected femora with more severe lesions, the affected femora 

demonstrated lower values for trabecular number than unaffected femora, although one of these 

displayed only slight difference (Pair 3, 2–3 years). 

Table 6.52: Number of comparison pairs in which Tb.N. values of affected femora, with lesions 
hypothesized to be less severe or more severe, were greater or less than unaffected femora. 

Tb.N. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. Affected  
“less severe” lesion 

Unaffected vs. Affected 
“more severe” lesion 

3/5 

0/5 

2/5 

5/5 

138 

 
 
 
Figure 6.27: Group bar plot for Tb.N. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be less severe or more severe. 

The values for trabecular spacing (Tb.Sp.) are compared between individuals without 

femoral cribra and individuals with the lesion that appeared less severe or more severe in Figure 

6.28. In 3/5 comparisons between unaffected femora and affected femora hypothesized to be less 

severe, the values for trabecular spacing were lower among the femora with less severe lesions 

(Table 6.53). The remaining two pairs demonstrated higher values of trabecular spacing for the 

affected femora compared to the unaffected femora. When femora without the lesion were 

compared to femora displaying more severe lesions, 3/5 pairs demonstrated higher measures of 

Tb.Sp. in the femora with more severe lesions, although the values for Pair 7 (3–5 years) were 

close (Table 6.53). The other pair in this group (Pair 3, 2–3 years) had similar values of 

trabecular spacing between the unaffected femur and the affected femur with a lesion thought to 

be more severe. 

139 

 
 
 
 
Table 6.53: Number of comparison pairs in which Tb.Sp. values of affected femora, with lesions 
hypothesized to be less severe or more severe, were greater or less than unaffected femora. 

Tb.Sp. 

Affected is greater 
than unaffected 

Affected is less 
than unaffected 

Unaffected vs. Affected  
“less severe” lesion 

Unaffected vs. Affected 
“more severe” lesion* 

2/5 

3/5* 

3/5 

1/5 

*Tb.Sp. values for Pair 3 femora and Pair 7 femora are similar 

Figure 6.28: Group bar plot for Tb.Sp. by comparison pairs of unaffected femora and affected 
femora with lesions hypothesized to be less severe or more severe. 

Affected femora with “less severe” lesions vs. affected femora with “more severe” lesions 

The five trabecular variables were compared between femoral cribra lesions hypothesized 

to be less severe and more severe in three pairs of individuals from the same estimated age 

groups. The outcomes of these comparisons are summarized in Table 6.54. In all three 

140 

 
 
 
 
 
 
comparison pairs (3/3), bone volume density (BV/TV) was lower in the lesions thought to be 

more severe than those thought to be less severe (Figure 6.29). Comparisons of specific bone 

surface (BS/BV) demonstrated higher values for more severe lesions in 2/3 comparisons, 

whereas the youngest pair (Pair 1, 2–3 years) demonstrated a lower BS/BV value for the more 

severe lesion (Figure 6.30). Measures of trabecular thickness (Tb.Th.) were lower in the more 

severe lesion for Pair 3, slightly higher in the more severe lesion of Pair 2, and similar between 

the less severe and more severe lesions in Pair 1 (Figure 6.31). In 2/3 comparisons trabecular 

number (Tb.N.) was greater in the more severe lesions than the less severe lesions (Pairs 2 & 3), 

whereas in the other comparison (Pair 1) trabecular number was lower in the more severe lesion 

(Figure 6.32). Values for trabecular spacing (Tb.Sp.) were lower in the more severe lesions in 2/3 

comparisons (Pairs 2 & 3) and higher in the more severe lesion of the other pair (Pair 1) (Figure 

6.33).  

Table 6.54: Number of comparison pairs in which the trabecular measures of affected femora, 
with lesions hypothesized to be less severe or more severe, were greater or less than unaffected 
femora. 

BV/TV 
BS/BV 
Tb.Th.* 
Tb.N. 
Tb.Sp. 

“More severe” is less than 
“less severe” 
3/3 (Pair 1, Pair 2, Pair 3) 
1/3 (Pair 1) 
1/3 (Pair 3) 
1/3 (Pair 1) 
2/3 (Pair 2 & Pair 3) 

“More severe” is greater than 
“less severe” 
0/3 
2/3 (Pair 2 & Pair 3) 
1/3 (Pair 2) 
2/3 (Pair 2 & Pair 3) 
1/3 (Pair 1) 

*Tb.Th. values for Pair 1 femora are similar 

141 

 
 
 
 
Figure 6.29: Group bar plot for BV/TV by comparison pair of affected femora with lesions 
hypothesized to be less severe and more severe. 

Figure 6.30: Group bar plot for BS/BV by comparison pairs of affected femora with lesions 
hypothesized to be less severe and more severe. 

142 

 
 
 
 
 
Figure 6.31: Group bar plot for Tb.Th. by comparison pairs of affected femora with lesions 
hypothesized to be less severe and more severe (error bars represent SD).  

Figure 6.32: Group bar plot for Tb.N. by comparison pairs of affected femora with lesions 
hypothesized to be less severe and more severe (error bars represent SD). 

143 

 
 
 
 
Figure 6.33: Group bar plot for Tb.Sp. by comparison pairs of affected femora with lesions 
hypothesized to be less severe and more severe (error bars represent SD). 

144 

 
 
 
CHAPTER SEVEN: DISCUSSION AND CONCLUSIONS 

In this chapter, the results of the data analyses presented in the previous chapter are 

interpreted and concluding thoughts are provided at the end. The first section discusses the 

results regarding the hypothesized prevalence of malaria at Mis Island within the larger context 

of the medieval period. The estimated dynamics of malaria at Mis Island during the medieval 

period are then discussed from comparing cemeteries 3-J-10 and 3-J-11, and comparing sex and 

age cohorts. The second section interprets the results on the skeletal lesions of interest in this 

dissertation. The suite of five skeletal lesions used to study the presence of malaria are first 

discussed, followed by the macroscopic and microscopic findings on femoral cribra. 

Hypothesized malaria at Mis Island 

The results of the five skeletal lesions’ frequencies and the application of Smith-

Guzmán’s (2015a) algorithm support the presence of malaria at Mis Island during the medieval 

period and suggest a relatively high prevalence of the disease. In comparing the frequencies of 

skeletal lesions between Mis Island and those reported by Smith-Guzmán (2015a) from the other 

two samples—one endemic for malaria and the other malaria-free—the Mis Island sample was 

closer to the sample endemic for malaria. The frequencies of these lesions in the Mis Island 

sample were not only more similar to those of the sample endemic for malaria (Smith-Guzmán, 

2015a), but rather exceeded the endemic sample for all lesions except spinal porosity, which was 

intermediate between the other two samples. Of particular note was the comparison between the 

samples for femoral cribra, in which the frequency observed at Mis Island was almost twice as 

much as that reported for the sample endemic for malaria. This finding supports the potential 

presence of malaria at Mis Island, especially as there were no cases of femoral cribra reported by 

Smith-Guzmán (2015a) in the malaria-free sample. 

145 

 
 
 
In applying Smith-Guzmán’s (2015a) algorithm to investigate the potential presence of 

malaria for each individual, 75%, or conservatively 78%, of the scorable individuals met the 

algorithmic criteria for malaria and therefore had possibly been infected with the disease. This 

high estimate for potential malarial infection in the sample supports its presence at Mis Island 

during the medieval period and suggests that these individuals interred in the site’s cemeteries 

may have been substantially impacted by the disease. The frequency of individuals 

demonstrating signs of potential infection are especially prominent when compared to the 

frequencies reported for the sites of Tombos (~1400–650 BCE) located at the Third Cataract in 

Upper Nubia (Buzon and Sanders, 2016) and the South Tombs Cemetery (1349–1332 BCE) at 

Amarna, Egypt (Smith-Guzmán et al., 2016). When compared with Mis Island, the closest 

reported frequency from the other sites was associated with the skeletal remains dating to the 

Third Intermediate and Napatan periods at Tombos (48% or 56%), followed by the South Tombs 

Cemetery at Amarna (50%), and lastly the skeletal remains dating to the New Kingdom from 

Tombos (30% or 35%). The comparatively lower frequencies of individuals demonstrating signs 

of possible malarial infection from Tombos and the South Tombs Cemetery could indicate that 

these communities experienced a lower prevalence of malaria. Alternatively, these results could 

reflect a higher prevalence of malaria, but more individuals who did not survive long enough for 

skeletal lesions to develop as compared to Mis Island. The higher frequency of individuals 

displaying skeletal lesions hypothesized to be associated with malarial infection at Mis Island, 

however, could be interpreted as a high prevalence of malaria with more continuous transmission 

that could result in elevated cases of estimated chronic malaria. This difference therefore may 

reflect the conditions believed to have been particularly conducive to mosquito vector 

146 

 
 
populations in the region during the medieval period associated with the introduction and 

expansion of saqia irrigation. 

The timing of saqia irrigation’s introduction in ancient Nubia has been a topic of some 

disagreement, with propositions estimating its presence in Nubia around the 3rd century CE 

during the Meroitic period (Fuller and Lucas, 2020) and others suggesting its arrival later during 

the early post-Meroitic period (Edwards, 2004). Despite this discrepancy, the spread and 

subsequent intensification of this agricultural technology in Nubia is considered a primary factor 

in the restructuring of subsistence systems and population patterns in the post-Meroitic period 

(Fuller and Lucas, 2020). Indication of saqia waterwheel irrigation at Mis Island during the 

medieval period is supported by remnants of the qadus jar uncovered at the medieval settlement 

3-J-19 on Mis Island (Thomas, 2008). Additional evidence of saqia irrigation in the area is 

associated with a potentially earlier date from the large Meroitic site adjacent to Mis Island at 

Umm Muri, where numerous fragments of qadus were excavated (Ahmed, 2014). With a relative 

lack of seluka land present along the Nile, the adoption of saqia irrigation fostered agricultural 

shifts with year-round cultivation, new crops, and more reliable food resources (Edwards, 2004). 

These changes allowed for a growth in population and territorial settlement, as well as an 

expansion of arable land (Edwards, 2004; Fuller and Lucas, 2020; Welsby, 2002); however, this 

anthropogenic influence may have affected dynamics with mosquitoes as malarial vectors. 

The intensification of saqia irrigation leading into and during the medieval period is cited 

as a primary factor in settlement growth around the Fourth Cataract region which in turn, is 

suggested to have increased mosquito vector populations and concomitantly malarial risk. In the 

Fourth Cataract region, the totality of the uncovered Kushite material culture indicates a 

relatively isolated rural society during the Napatan and Meroitic periods (Ahmed, 2014). 

147 

 
However, the Middle Nile Valley witnessed a substantial expansion of settlements and 

population between 375–1500 CE.  Settlement patterns intensified in the Fourth Cataract region 

and adjacent western Abu Hamed Reach in the post-Meroitic and medieval periods (Malcolm et 

al., 2007) and rural settlements increased along the Dongola Reach during the 6th century CE 

(Obłuski, 2020). This growth in population and settlement density is largely attributed to the 

more extensive implementation of saqia irrigation, which allowed for expansion of agricultural 

activity in areas previously less inhabited and during extended periods of time (Malcolm et al., 

2007; Obłuski, 2020). However, these settlement patterns may have contributed to a context in 

which mosquito populations and malarial transmission could be more supported (Malcolm et al., 

2007), as compared to areas of particularly low population density in which the lack of people 

may not support or sustain transmission (Tatem et al., 2008). 

In addition to changes in settlement patterns and human population density, the 

implementation of saqia irrigation may have further supported mosquito vector populations by 

providing more artificial breeding sites. The use of saqia irrigation is thought to have resulted in 

an agricultural revolution in Nubia during the post-Meroitic and early medieval periods, in which 

the amount of arable land increased and year-round cultivation could be maintained (Edwards, 

2004; Welsby, 2002). The saqia waterwheels enabled irrigation of water from the Nile into more 

elevated areas inland, whereas previously most of the cultivatable land would have likely 

included the flooded seluka areas potentially accompanied by hand-watering methods and/or the 

shaduf tool that lifts water from the source (Dalton et al., 2023). However, the increase in the 

amount and duration of irrigated land may have provided additional breeding sites for mosquito 

vectors beyond the river’s edge and the naturally available seluka land. Irrigation canals and 

associated artificial sources of water involved in agriculture provide additional breeding sites for 

148 

 
 
mosquito vectors (Coates and Redding-Coates, 1981), which can support their population and 

promote their proliferation (Fuller et al., 2012). Observations of such artificial water sources in 

regions around the Fourth and Third cataracts have revealed that An. arabiensis vectors use these 

breeding sites especially during the months when the river levels are high, and therefore provide 

suitable breeding opportunities located further from the river’s disruption (Ageep et al., 2009; 

Dukeen and Omer, 1986). 

The communities associated with the more urban sites of Tombos and the South Tombs 

Cemetery at Amarna predate the estimated implementation of saqia irrigation in ancient Nubia 

and Egypt, and thus may not have experienced the extent of influences for malarial risk 

hypothesized to have accompanied this agricultural technology. Occupation of Tombos at the 

Third Cataract occurred before the estimated introduction of saqia irrigation in Nubia and 

Napatan agriculture was likely similar to Middle Kingdom and New Kingdom Egyptian 

occupations in Nubia (Fuller and Lucas, 2020). Additionally, while settlement during the 

Napatan period was focused along the Dongola Reach with centers around Tabo and Kawa south 

of the Third Cataract (Edwards, 2004), the Napatan settlement patterns below the Third Cataract 

in the Dongola Reach were relatively isolated and likely were not as conducive to mosquito 

populations (Malcolm et al., 2007). In ancient Egypt, the saqia waterwheel is believed to have 

been introduced during the second century BCE and the hydraulic noria is thought to have 

appeared by the fourth century BCE (Elmesery, 2020), both of which occurred after the period 

associated with the South Tombs Cemetery.  

The closer frequency of estimated malarial infection from Tombos during the Third 

Intermediate/Napatan periods and the South Tombs Cemetery with the Mis Island sample 

compared that of the skeletal remains from Tombos during the New Kingdom may be associated 

149 

 
with differences in habitual activities and inferred social status. Recent interpretations of Tombos 

during the New Kingdom suggest its establishment as an Egyptian fortress during the New 

Kingdom occupation with both Egyptian and local Nubian inhabitants (Gibbon and Buzon, 

2018). Analyses of inferred activity levels at Tombos have been interpreted as indication that the 

New Kingdom inhabitants may have included more administrative workers, whereas 

comparatively higher activity levels hypothesized for the local inhabitants during the Third 

Intermediate/Napatan period are attributed in part to heightened agricultural activity (Gibbon and 

Buzon, 2018). Interpretations of the skeletal remains from the South Tombs Cemetery have 

suggested the individuals interred in this non-elite cemetery experienced relatively high levels of 

physical activity (Kemp et al., 2013). These heightened amounts of activity in the two samples 

may have been associated with more exposure to mosquito breeding sites and thus, potentially 

influenced the higher frequencies of estimated malaria that were closer to that observed from the 

Mis Island sample. 

Comparing cemeteries 3-J-10 and 3-J-11 

The comparable frequencies of individuals demonstrating signs of potential malarial 

infection from cemeteries 3-J-10 and 3-J-11 suggest similar exposure to the disease between 

these cemetery groups at Mis Island. Overall, the estimates of hypothesized malaria were 

similarly elevated between cemeteries 3-J-10 and 3-J-11 at 71% and 77%, respectively. As 

cemetery 3-J-10 reflects burials from the late medieval period (~1100–1500 CE) and cemetery 3-

J-11 represents burials spanning pre-medieval to late medieval years (~300–1400 CE), these 

results suggest that there may not have been substantial differences in risk of infection between 

the duration of the medieval period and the late medieval period. Upon closer examination, 

equivalent frequencies were observed when each age cohort from one cemetery was compared to 

150 

 
 
the corresponding age group in the other cemetery. When individuals associated with estimated 

biological sex were compared between cemeteries, no significant difference was found in 

comparing males from cemetery 3-J-10 and males from cemetery 3-J-11 and the frequencies of 

hypothesized malaria were nearly identical. The comparison between females from cemetery 3-J-

10 and females from cemetery 3-J-11 revealed a higher frequency observed among females from 

cemetery 3-J-11; however, this difference was not statistically significant despite approaching 

significance. One factor that could have influenced this outcome was the difference in subsample 

sizes, as there were fewer individuals from cemetery 3-J-10 available for analysis and therefore 

could be underrepresented in the comparisons between cemeteries.  

In comparing the five skeletal lesions between cemeteries 3-J-10 and 3-J-11, significantly 

higher frequencies of cribra orbitalia and spinal porosity were observed in cemetery 3-J-11 which 

suggests potentially higher stress and/or more chronic pathological insult experienced by these 

individuals. For cribra orbitalia, higher frequencies were observed across age groups of 

individuals from cemetery 3-J-11 compared to 3-J-10, which parallels Hurst’s (2013) finding of 

significantly higher prevalence of cribra orbitalia among subadult individuals from cemetery 3-J-

11. While the skeletal lesions studied by Hurst (2013) other than cribra orbitalia were not 

statistically significant between subadults from both cemeteries, Hurst also observed an overall 

greater prevalences of other skeletal stress indicators in cemetery 3-J-11. In the current research, 

differences in the prevalence of spinal porosity between the cemeteries were most pronounced in 

the adult age groups. When individuals with a biological sex estimate were compared, both 

males and females from cemetery 3-J-11 demonstrated significantly higher frequencies of spinal 

porosity than males and females from cemetery 3-J-10, respectively. As was found in a 

subsequent analysis for Research Question 4, spinal porosity was observed at peak frequencies in 

151 

 
the adolescent and young adult age groups which may suggest a relatively longer formation 

compared to the other porous lesions, or a later onset of formation in response to the health 

insult. The higher frequency of spinal porosity among the adults in particular from cemetery 3-J-

11 may indicate greater pathological insult experienced either during younger ages that 

manifested in the skeleton over time in those individuals who survived, or during more recent 

adolescent or adult ages. 

While changes in the Nile River’s levels during the medieval period may not have 

affected estimated prevalence of malarial infection visible in the skeletal remains as 

hypothesized, these environmental changes may have contributed to the higher prevalence of 

these lesions in cemetery 3-J-11. During the late medieval period, levels of the Nile River are 

thought to have been relatively high, with particularly prominent floods between approximately 

1300 and 1522 CE (Edwards, 2004; Malcolm et al., 2007). However, prior to this time, 

particularly low levels of the Nile are thought to have been present between 828–837 CE and 

again between 939–948 CE (Edwards, 2004). In noting how low levels of the Nile River and 

drought could have deleterious effects on cultivation and subsistence, Hurst suggests that some 

portion of cemetery 3-J-11 may represent elevated amounts of morbidity and mortality from such 

times of hardship. Given malaria’s synergistic relationship with other pathologies (White, 2018), 

the combined results of the inter-cemetery comparisons may reflect the effects of comorbidities 

and their interactions on an individual’s health. Such co-occurrence of harmful health factors 

may have contributed to the higher frequency of estimated malaria among females from 

cemetery 3-J-11, although not statistically significant, as malaria poses a particular risk to 

pregnant women who are in an immunosuppressed state (Carter and Mendis, 2002). 

152 

 
 
Comparing sex and age cohorts 

The results of comparing the prevalence of hypothesized malaria between males and 

females, overall and by age group, indicate that there may not have been a substantial difference 

in exposure to mosquito vectors based on sex at Mis Island during the medieval period. While 

the frequency of estimated malarial infection was slightly higher in females compared to males, 

the prevalences were similarly elevated and no statistically significant difference was found. 

Therefore, these findings suggest a lack of differential proximity to mosquito vectors and their 

habitat between males and females at Mis Island that could have markedly influenced malarial 

risk.  

A similar exposure to malaria between males and females is further supported by the 

absence of statistically significant differences for the majority of the skeletal lesions examined; 

however, a significantly higher frequency of femoral cribra was observed in the female group, 

albeit a weak relationship. The prevalence of femoral cribra was greater across age groups for 

females compared to males, but the larger subsample size of observable femora among female 

individuals may have influenced this result. While some previous studies investigating this lesion 

have not found this relationship (Schats, 2021; Smith-Guzmán, 2015a), Radi and colleagues 

(2013) also observed a significantly higher prevalence of femoral cribra among females 

compared to males in their 20th century skeletal sample from a cemetery in Bologna, Italy. The 

higher amount of femoral cribra in females may indicate differential stress experienced earlier in 

life, although Soler (2012) observed comparably elevated levels of non-specific skeletal 

indicators between males and females. Interpretation of this difference between males and 

females for the single lesion therefore is difficult beyond acknowledged sample size differences. 

However, this pattern for femoral cribra could be related to the immunosuppression experienced 

153 

 
 
by women during pregnancy (Carter and Mendis, 2002) or possibly differences were present that 

could have affected the remodeling process of the lesion leading to a more sustained expression. 

The statistically significant age-related pattern of prevalence of hypothesized malaria 

indicates an elevated level of chronic condition rather than acute status. An overall trend was 

found in the frequency of estimated malaria that increased with age to the adolescent age group, 

followed by a high prevalence in the young adult cohort, and decreased with age in the middle 

and older adult age groups. Additionally, the frequencies across age groups were high with the 

child, adolescent, and young adult cohorts over 80% and the infant, middle adult, and older adult 

groups over 50%. These results may reflect a greater prevalence of malarial infection in younger 

age groups with comparatively less prevalence in older age groups and/or some degree of lesion 

remodeling in those individuals who survived into adulthood.  

This pattern across the age cohorts is potentially suggestive of a high prevalence of more 

chronic malaria in which some degree of immunity was likely acquired with age or duration of 

contact. Visible skeletal lesions implicated in malarial infection are likely more indicative of 

chronic rather than acute conditions, as some time would presumably be involved for skeletal 

signs to manifest (Wood et al., 1992) and chronic malaria is associated with increased marrow 

cellularity and erythroid hyperplasia (Wickramasinghe and Abdalla, 2000). Therefore, 

individuals with these lesions may represent those who survived for a variable amount of time 

and thus likely developed some degree of immunity. Repeated exposure to malaria can induce 

partial immunity that can decrease symptomatic expressions, although still involve carrying of 

the parasite (Ratti and Wallace, 2020; Tumwiine et al., 2007). This acquired immunity, however, 

can fluctuate and disappear without recurrent exposure (Ratti and Wallace, 2020; Tumwiine et 

al., 2007) and thus, is associated with age of the individual. Newborn infants whose mothers 

154 

 
 
obtained some acquired immunity inherit a degree of protection for approximately six months, 

after which time they become susceptible to infection and build their own immune defense to 

infections (Ratti and Wallace, 2020). While acquired antimalarial immunity is less likely during 

the youngest ages, older children and adults may be more able to develop such immunity (Carter 

and Mendis, 2002). With repeated exposure, adult individuals can also develop a protection that 

suppresses the parasite, which can take about one or two decades to build and could be slowly 

lost with breaks in exposure (Ratti and Wallace, 2020). Achieving and maintaining some degree 

of acquired antimalarial protection, however, depends on the endemicity of malaria and the 

regularity of exposure. 

 The pattern and frequencies of estimated malaria across age groups suggests a primarily 

endemic type of malaria at Mis Island during the medieval period rather than strictly epidemic 

malaria. Epidemic malaria takes place in non-endemic areas that typically have no or minimal 

transmission and involve an increase in transmission with “low or absent” protective immunity 

in all age groups (Carter and Mendis, 2002:567). Individuals experiencing epidemic malaria may 

therefore be less likely to survive before skeletal indication of malaria could manifest, whereas 

individuals in areas of endemic malaria would be more likely to experience chronic anemia that 

might be visible in the skeleton (Smith-Guzmán et al., 2016). Given the high frequency of 

individuals displaying signs of possible infection and skeletal lesions across the cohorts at Mis 

Island, the primary type of endemicity is unlikely to be epidemic.  

Endemic malaria can further be classified as stable or unstable. In stable malaria, 

immunity is high in older individuals as the continuous effects of malaria may lead to sustained 

immunity, whereas in unstable malaria immunity is inconstant in older individuals as fluctuations 

in transmission can interrupt acquired antimalarial protection (Carter and Mendis, 2002). A lack 

155 

 
 
of insight into the severity and activity of all five skeletal lesions at this time limits conclusive 

considerations on the possible onset and chronic status that might be associated with relative 

immunity. However, the high frequency of middle and older adult individuals displaying signs of 

possible malarial infection may reflect periods of unreliable immunity status associated with 

unstable endemic malaria. 

Skeletal lesions implicated in hypothesized malaria 

Association between skeletal lesions and age 

In the Mis Island skeletal sample, cribra orbitalia, humeral cribra, femoral cribra, and 

spinal porosity were significantly associated with age of the individual with generally greater 

prevalences in the younger age groups. The highest frequencies of cribra orbitalia were observed 

in the infant, child, and adolescent cohorts, with similar prevalences across the three groups, and 

the lowest frequencies were seen in the adult age groups. Rather than demonstrating a heightened 

prevalence in the infant age group, frequencies of humeral and femoral cribra increased to peak 

prevalences in the child and adolescent groups and were then lower in the adult age groups. This 

distribution of lesion prevalence across age groups is similar to that observed by Schats (2021) in 

skeletal remains from the medieval/early modern Netherlands (800–1600 CE). For cribra 

orbitalia, humeral cribra, and femoral cribra, Schats (2021) reports frequencies reaching a peak 

in the child age group (4–12 years) with slightly lower prevalence in the adolescent (13–19 

years) age group. The similarities in age distribution for these lesions between the samples may 

reflect a common tendency for them to form during childhood and subsequently persist into 

adulthood. In the Mis Island sample, the age distribution of spinal porosity demonstrated a 

different pattern with increasing frequencies reaching peak prevalence in the adolescent age 

cohort, followed by the young adult group, and decreasing frequencies in the middle and older 

156 

 
adult groups. Compared to the other porous lesions, this finding may tentatively indicate a longer 

period of formation for spinal porosity or a later onset.  

The association between the porous lesions and age of the individual, and the distribution 

of the lesions across age groups, are consistent with the hypothesis that their formation may be 

related to anemia. In anemia, skeletal lesions are thought to primarily result from red bone 

marrow expansion in order to promote red blood cell production (Brickley, 2018). Given the 

changes in red marrow distribution with age, the presence of skeletal lesions resulting from 

anemia are considered to represent formation during younger ages when red marrow is present, 

and only in regions where red or mixed marrow remains. Marrow throughout the skeleton is red 

in infancy but progressively converts to yellow marrow, such that the adult pattern of distribution 

is obtained by 25 years old and red marrow generally remains in the skull, vertebrae, sternum 

and ribs, pelvis, and proximal humeri and femora (Kricun, 1985). Based on the pattern observed 

in the Mis Island skeletal sample, the porous lesions may have formed during younger ages when 

red marrow was still present and thus could have resulted from anemia. The age distribution of 

spinal porosity peaking in adolescent and young adult ages, rather than the younger age groups 

as demonstrated by the cribrous lesions, may be associated with the relative extents of red 

marrow with increasing age. Larger amounts of red marrow persist in the vertebrae into 

adulthood and are therefore sites in which hyperplasia may be more likely to occur in adult 

individuals with anemia (Brickley et al., 2020).  

If these porous lesions tend to form during younger ages, the presence of the lesions 

observed in adult individuals could represent their persistence over time from continued chronic 

insult or incomplete resorption. The relatively lower prevalence of these lesions in adult 

individuals may be due to remodeling of previous lesions such that they are obscured in older 

157 

 
 
ages. Additionally, the lower prevalence of the lesions in adult age groups could be associated 

with nonsurvival of individuals during childhood. 

While the four porous lesions were significantly associated with age, periosteal reaction 

did not demonstrate a significant relationship with age of the individual. Instead, this lesion was 

observed at a relatively consistent prevalence across age groups with the highest frequency in the 

infant cohort. Hurst (2013) observed varying patterns of lesion activity for periosteal reaction 

between age groups at Mis Island and suggests that in this population, periosteal reaction may be 

more related to individual circumstances rather than an overarching experience. Therefore, rather 

than a close association with biological variables tied to age (e.g., red marrow distribution), the 

results reflect that periosteal reaction can be the result of multiple factors that result in 

disturbance of the periosteum (e.g., infection and trauma). 

Association among skeletal lesions 

In examining the association between pairs of the five lesions of interest in all individuals 

regardless of age, significant associations were observed primarily among the porous lesions; 

however, these patterns were influenced by age. When the relationships between lesions were 

examined separately in subadult and adult groups, significant associations were found between 

pairs of lesions among subadults, but no significant associations were observed among adults. 

This finding supports the previous discussion on the connection between age of the individual 

and lesion expression.  

For the three cribrous lesions implicated in the “cribrous syndrome”, significant 

associations were observed between each pair of cribrous lesion (CO-HC, CO-FC, HC-FC) when 

all individuals regardless of age were examined. The investigation into patterns among these 

lesions in subadult individuals also revealed significant associations between the pairs of 

158 

 
cribrous lesion, except between cribra orbitalia and humeral cribra. The frequency of the three 

cribrous lesions co-occurring in the Mis Island sample is higher compared to other studies. While 

27% of the individuals scorable for all three lesions at Mis Island displayed the combination of 

lesions, a frequency of 4.9% was observed by Schats (2021) in a sample of subadults and adults 

from the medieval/early modern Netherlands (800–1600 CE). When the simultaneous presence 

of the three lesions is examined in subadults only, Mis Island also demonstrates elevated 

prevalence with 44% compared to 33.3% observed by Djuric and colleagues (2008) and 13.2% 

reported by Schats (2021). Additionally, while significant associations were found between 

cribra orbitalia and the postcranial cribrous lesions in the Mis Island sample, no significant 

association between cribra orbitalia and humeral and femoral cribra was found by Smith-

Guzmán (2015a) in the endemic malaria sample, or by Schats (2021) in the sample from the 

medieval/early modern Netherlands. 

The significant associations found between cribra orbitalia and the postcranial cribrous 

lesions in the Mis Island sample, not observed in previous studies, could be due to various 

factors. Given the high frequencies of cribra orbitalia and femoral cribra in the Mis Island 

sample, the significant associations could be an artifact of the greater likelihood for co-

occurrence when the prevalence of one lesion is high. Regarding possible etiology, the 

associations may reflect shared pathological processes and origins including anemia. 

Alternatively, some proportion of cribra orbitalia may be due to different processes that were 

simultaneous with the those producing the postcranial cribrous lesions. This possibility is based 

on previous studies that did not find significant associations between the lesions and the non-

specific characterization of cribra orbitalia which has been attributed to multiple conditions. In 

the Mis Island sample, Hurst (2013:230) observed a “relatively common occurrence” of scurvy 

159 

 
 
in subadults and a significant association between skeletal indication of scurvy and cribra 

orbitalia. In an effort to avoid inclusion of cribra orbitalia more likely associated with scurvy in 

the current study, cribra orbitalia was scored as present when there was perforation through the 

cortical bone and a lack of subperiosteal new bone formation (Smith-Guzmán et al., 2018). 

However, cribra orbitalia matching that description and possibly resulting from processes other 

than marrow hyperplasia may account for some cases. With malaria and its concomitant anemia 

as a posited etiology for the porous lesions observed at Mis Island, comorbidities may be 

possible as malarial infection has a synergistic relationship with other conditions and pathologies 

(White, 2018). 

In addition to the current research, previous studies have found significant associations 

between humeral cribra and femoral cribra, including in Smith-Guzmán’s (2015a) malaria 

endemic sample and Schats’s (2021) sample of subadult remains from the medieval/early modern 

Netherlands. Furthermore, Schats (2021:87) observes that “cribra humeri does not occur without 

cribra femora being present”, which was a pattern also found by Garcia and colleagues (2002) 

who observed 100% association between humeral cribra and femoral cribra in skeletal remains 

from a Late Roman (3rd–5th centuries CE) necropolis in Tarragona, Spain. A similar relationship 

between humeral cribra and femoral cribra was observed in the Mis Island sample, in that of the 

individuals with humeral cribra who could also be scored for femoral cribra (n=91), 95% 

displayed both lesions. These findings suggest that humeral cribra may be unlikely to form in 

isolation without the simultaneous presence of femoral cribra. Therefore, humeral cribra may 

form during a shared etiology and process in the formation of femoral cribra. The appearance of 

femoral cribra without humeral cribra could be related to the potential for thinner cortical bone 

thickness in the femoral neck as observed in adults (Schats, 2021). 

160 

 
The significant associations found between the two postcranial cribrous lesions and 

spinal porosity (HC-SP and FC-SP) suggest that they are involved in a similar etiological 

response. This result was observed when all individuals were analyzed and when subadults were 

assessed separately but was not observed in the adult cohort. Therefore, a possible common 

etiology is likely age-related. These findings support a shared process of formation during 

younger ages which is consistent with the hypothesis that anemia might be a cause. 

Periosteal reaction demonstrated a significant association with humeral cribra (PR-HC) 

when examined in all individuals combined, and with femoral cribra and spinal porosity (PR-FC 

and PR-SP) when these pairings were investigated in subadults only. However, the association 

between periosteal reaction and these porous lesions was an inverse relationship. This result is 

contrary to Smith-Guzmán’s findings from the malaria endemic sample, in which the presence of 

periosteal reaction was found to be associated with femoral cribra and spinal porosity (2015a). 

Smith-Guzmán interpreted these results as suggestion that the three lesions could be part of “the 

same inflammatory response” (2015a:629). Therefore, instances of periosteal reaction in the Mis 

Island sample may represent various etiological responses in addition to the hypothesized 

malarial infection. 

Femoral cribra 

Indicators suggestive of lesion activity 

Observations of estimated lesion activity from the external surface of femoral cribra 

resulted in an age-related pattern consistent with the expected outcomes for “active” and 

“healed” lesion status. To investigate signs of lesion activity, this research used an approach 

based on Mensforth and colleagues’ (1978) method for porotic hyperostosis in which femoral 

cribra displaying sharp edges was estimated as “active” and femoral cribra presenting smooth-

161 

 
 
edged bone was hypothesized to be “healing”. The significant association found between age of 

the individual and estimated activity of femoral cribra, in which signs of “healing” increased 

with age, supports the proposition that these features of femoral cribra are indicative of the 

lesion’s activity. The results from the current research are similar to those reported by Mangas-

Carrasco and López-Costas (2021) who also observed a healing pattern of femoral cribra with 

age. Using a scoring system of active, mixed, or inactive based on the lesion’s degree of 

expression, Mangas-Carrasco and López-Costas found that signs attributed to healing increased 

with age in remains excavated from late medieval archaeological sites in NW Spain. 

Additionally, the results of lesion size for “active” and “healing” femoral cribra observed in each 

age group revealed clustering of relatively larger lesions in the juvenile and adolescent cohorts 

with primarily active-looking lesions compared to the more varied sizes of primarily healing 

lesions in the adult cohorts. These findings support the anticipated age-related pattern of the 

lesion’s formation and progression. 

With origins considered to be primarily pathological in nature, the higher prevalence of 

more active-looking femoral cribra in younger individuals could reflect the expectation that 

subadults represent more “frail” individuals who did not survive health insults. Conversely, the 

greater frequency of femoral cribra that appears healing in the adult age groups could represent 

those individuals who survived prior stress and experienced some degree of recovery. This 

pattern is also consistent with the proposed etiology of anemia, in that femoral cribra may tend to 

form during younger ages and therefore appear active in subadult individuals and demonstrate 

signs of healing over time with increasing age and survivorship. However, the cases of femoral 

cribra in adult individuals that appear active could reflect the retainment of red bone marrow in 

the proximal femur into adulthood (Kricun, 1985). Additionally, active-looking femoral cribra in 

162 

 
older adult cohorts could represent the potential for reconversion of yellow marrow into red 

marrow during conditions such as chronic anemia—especially hemolytic anemia—in response to 

a need for heightened hematopoiesis (Kricun, 1985). While the results of lesion activity 

indicators from femoral cribra’s external surface are consistent with expectations of “active” and 

“healing” states, an examination of the quantitative measures reflecting the underlying bony 

structure contributes to these findings. 

The results of the trabecular microarchitecture variables derived from micro-CT scans of 

unaffected individuals and affected individuals with varying degrees of activity support 

indications of the lesion being active or healing observed on the surface. In comparing unaffected 

femora with femora thought to have “active” lesions, the “active” lesions within the comparison 

pairs demonstrated the following pattern: all had lower values for bone volume density (BV/TV), 

most had higher values for specific bone surface (BS/BV), most had higher values for trabecular 

thickness (Tb.Th.), most had lower values for trabecular number (Tb.N.), and most had higher 

values for trabecular spacing (Tb.Sp.). Based on the expected changes to these trabecular 

variables involved in resorptive, formative, and mixed processes (Morgan, 2014), these active 

femoral cribra lesions are most consistent with mixed cellular activity. However, when 

unaffected femora and affected femora with lesions hypothesized to be “healing” based on the 

external appearance were compared, this pattern was not as consistent. Compared to the 

unaffected femora, the affected femora with “healing” lesions demonstrated generally similar 

occurrences of lower and higher values of all trabecular variables except for trabecular thickness 

(Tb.Th.), which was greater in the “healing” femora compared to the unaffected femora in all but 

one comparison pair. This lack of an overall pattern could reflect various stages of healing in 

these affected femora, as the differences in the variables for some pairs were suggestive of more 

163 

 
mixed or resorptive processes in the affected femora, while other comparison pairs were 

indicative of more formative processes in the affected femora. The latter pairs were observed 

particularly in the adult cohorts and may suggest more advanced healing with age. 

The comparisons between the trabecular variables of affected femora with lesions 

hypothesized to be “active” or “healing” further support the correlation between indications of 

activity from the external surface with underlying bone processes. A relatively consistent pattern 

was observed across the comparison pairs in which the “healing” lesions demonstrated the 

following relative to the “active” lesions: almost all had higher values for bone volume density 

(BV/TV), all but one had lower values for specific bone surface (BS/BV), all had higher values 

for trabecular thickness (Tb.Th.), and almost all had lower values for trabecular spacing (Tb.Sp.). 

Trabecular number (Tb.N.) was the only variable in which the results were divided relatively 

evenly with the “healing” lesions displaying higher or lower values. Based on the anticipated 

changes in trabecular variables for the types of bone processes (Morgan, 2014), this trend in the 

relative differences of trabecular microarchitectural variables suggests that the lesions that appear 

“healing” on the surface may have undergone more formative processes compared to the lesions 

that appear “active” on the surface. Qualitative observations of the micro-CT scans between 

these comparison pairs support the quantitative findings. As reflected in Figure 7.1, the lesion 

that appears “healing” displays a relatively greater amount of trabecular bone volume, the 

trabeculae are thicker and appear more numerous in this case, and there is less spacing between 

the trabeculae.  

164 

 
 
Figure 7.1: External and internal features of a lesion hypothesized to be active (1089) and a 
lesion hypothesized to be healing (3277) from Comparison Pair 4 (10–15 yrs). 

In further investigation of femoral cribra’s activity status, the presence of a raised margin 

adjacent to the area of porosity or exposed trabeculae, referred to as a “cortical window” in the 

current research, may be consistent with lesion “healing” in the Mis Island skeletal sample. This 

cortical window was significantly associated with signs of healing in the affected porous area 

and when present, was observed primarily with lesions estimated to be healing. Additionally, the 

cortical window was virtually absent in subadult individuals and was observed with increasing 

frequency with age in the adult individuals. These results support the hypothesis that the cortical 

window is indicative of a healing process that can occur over time.  

While this feature was significantly associated with lesions estimated to be healing, the 

occurrence of its expression was relatively low. The absence of a cortical window in lesions 

thought to be healing could be attributed to a potentially faster rate of remodeling and healing 

processes in trabecular bone compared to cortical bone (Oftadeh et al., 2015; Sandberg and 

Aspenberg, 2016), such that signs of “healing” in trabecular bone may precede a cortical window 

in femoral cribra. Although mainly observed in combination with signs of healing from the 

porous bone or trabeculae, there were several instances in which it was observed with lesions 

estimated to be active. These cases could be a result of a process impeding porous or trabecular 

remodeling, reactivity of a previously healing lesion, or a different origin of the cortical window 

from the proposed cause. 

165 

 
 
 
The comparisons of trabecular microarchitecture variables obtained from micro-CT scans 

involving femoral cribra with cortical windows and estimated to be healing generally support 

this feature’s association with indication of healing activity observed from the surface. When 

compared to affected femora with active-looking lesions, and to some extent with unaffected 

femora, most of the femora with cortical windows demonstrated differences in the trabecular 

measures that were consistent with formative processes. However, one femur with a cortical 

window did not follow this trend and when compared to its unaffected counterpart (Pair 10), 

seemed to reflect more resorptive processes. Additionally, when compared to an affected femur 

with a lesion estimated to be active (Pair 5), the femur with the cortical window estimated to be 

healing had relatively similar trabecular measures except lower trabecular number (Tb.N.) and 

greater trabecular spacing (Tb.Sp.). While possible factors behind the results for this particular 

lesion with the cortical window are currently unclear, this individual may have experienced an 

especially severe condition, or some contributing factor may have been influencing trabecular 

bone processes. 

In discussing variation in skeletal features of the anterior femoral neck, Göhring (2021) 

contends that convex porous lesions lacking cortical borders should be considered separate from 

concave lesions of exposed trabeculae with evident cortical margins. Göhring suggests that the 

former is “cribra femoris” and is pathological in nature, whereas the latter should be identified as 

Allen’s fossa which Göhring describes as likely caused by “friction of the zona orbicularis” of 

the hip joint capsule during activity (2021:4). In the current research, instances in which cortical 

windows were scored as present likely overlap with Göhring’s definition of Allen’s fossa as a 

depressed nonmetric trait. Additionally, all cortical windows observed in the Mis Island sample 

were associated with Radi et al. 2 scores, and therefore displayed trabecular exposure. While the 

166 

 
effects of the zona orbicularis on features of the femoral neck are not dismissed in this 

discussion, this structure’s involvement in the expression of cortical windows observed in the 

studied skeletal remains is currently questioned. The zona orbicularis is a set of curved fibers 

associated with the hip joint capsule perpendicular to the joint axis at the narrowest part of 

femoral neck which aid in joint stability and do not attach directly to bone (Tsutsumi et al., 2021; 

Wagner et al., 2012). The morphology of this structure, however, has been a subject of 

investigation and uncertainty with some studies finding that it is absent or less prominent 

anteriorly (Malagelada et al., 2015; Wagner et al., 2012) and may not be part of the “local collar” 

(Tsutsumi et al., 2021:1163). Therefore, a connection between the zona orbicularis and formation 

of cortical windows around femoral cribra as defined in the current research is currently unclear. 

This reservation especially pertains to observed cases of femoral cribra with small, focal 

windows on the anterior neck, as well as areas of exposed trabeculae with windows that are not 

isolated to the narrowest section of the femoral neck. 

Indicators suggestive of lesion severity 

In examining the patterns of femoral cribra’s cortical disruption using Radi and 

colleagues’ (2013) scoring system, the results suggest that the lesion’s expression as a grouping 

of pores on the surface (Radi et al. Score 1) may represent a less severe and earlier stage of 

formation whereas the lesion’s appearance with trabecular exposure (Radi et al. Score 2) may 

reflect a more severe stage. The Radi et al. Score 1 expression of femoral cribra was most 

frequently observed in the infant cohort and represented almost half of the cases in that age 

group, whereas this score was seen at lower frequencies in the other age groups. This score also 

clustered closer to the active lesion status than the healing lesion status in the MCA for left 

femora. Conversely, the Radi et al. Score 2 expression accounted for most of the lesions 

167 

 
observed in the child, adolescent, and adult age groups. Radi and colleagues (2013) did not 

examine subadult remains in their study and therefore the subadult frequencies observed in the 

current research cannot be compared. However, a greater prevalence of the Radi et al. Score 1 

expression was reported among the adult individuals in Radi and colleagues’ research compared 

to the adult individuals of Mis Island. Therefore, the Radi et al. Score 1 does not seem to be a 

strictly age-related expression of femoral cribra, but rather may reflect a less severe lesion state.   

These indications of femoral cribra’s potential severity are further supported by the 

results of lesion size. Comparing measurements of the lesion—perpendicular and parallel to the 

femoral neck and perimeter—between the two scores generally revealed smaller dimensions for 

the Radi et al. Score 1 lesions and larger dimensions for the Radi et al. Score 2 lesions. These 

differences were most pronounced for the size of lesion perpendicular to the femoral neck and 

consequently perimeter of the lesion. The trend of Radi et al. Score 2 lesions displaying 

particularly larger dimensions perpendicular to the neck compared to Radi et al Score 1 lesions, 

but relatively more similar dimensions parallel to the neck, may tentatively indicate a propensity 

of femoral cribra to expand perpendicular to the neck in its progression. However, the small 

subsample size of Radi et al. Score 1 lesions in the current research may not represent potential 

variation in this measurement.  

The comparisons of trabecular microarchitecture variables obtained from micro-CT scans 

between unaffected femora and affected femora thought to be “less severe” or “more severe” 

support this finding for the Radi et al. scores. The comparison pairs involving subadult 

individuals with age-at-death estimates between 2–3 years included the only available example 

of a lesion assessed as a Radi et al. Score 1 that could fit in the sample sled and be scanned. As a 

result, the “less severe” lesion in the comparison pairs from this age cohort (subadult 2–3 yrs) 

168 

 
represents a lesion assessed as a Radi et al. Score 1 (1352) and the “more severe” lesion was 

evaluated as a Radi et al. Score 2 (5013) (Figure 7.2). In comparing these affected femora with 

the unaffected femora from this age group, both the “less severe” (Radi et al. Score 1) and “more 

severe” (Radi et al. Score 2) lesions demonstrated differences in the trabecular variables that are 

suggestive of more formative processes. While the “less severe” and “more severe” lesions 

displayed similar trends compared to the femora without the lesion in this cohort, the differences 

in the trabecular variables demonstrated by the “less severe” lesion are potentially suggestive of 

more formative activity compared to the “more severe” lesion (Morgan, 2014). Of these, larger 

differences in the unaffected vs. “less severe” femoral cribra comparison pairs were noted in the 

higher bone volume density (BV/TV) and trabecular number (Tb.N.) measures, and the lower 

trabecular spacing (Tb.Sp.) values.  

Furthermore, in comparing these “less severe” (Radi et al. Score 1) and “more severe” 

(Radi et al. Score 2) lesions, the “more severe” lesion with a Radi et al. Score 2 showed lower 

bone volume density (BV/TV), specific bone surface (BS/BV), and trabecular number (Tb.N.), 

similar trabecular thickness (Tb.Th.), and higher trabecular spacing (Tb.Sp.). If femoral cribra 

lesions displaying a Radi et al. Score 1 reflect a less severe and earlier stage of formation, and 

lesions with a Radi et al. Score 2 represent greater severity in femoral cribra’s progression, the 

findings from this series of micro-CT scans suggest that at this age formative processes may be 

involved in the lesion’s appearance; however, more resorptive processes may potentially be part 

of femoral cribra’s subsequent expression with exposed trabeculae. This proposed model of 

processes is visualized in the micro-CT scans (Figure 7.2), in that compared to the lesion with 

the Radi et al. Score 1 (1352), the lesion with the Radi et al. Score 2 (5013) displays less bone 

volume density, specific bone surface, trabecular number, and higher trabecular spacing in the 

169 

 
area of the lesion. As these comparisons between Radi et al. scores were only attainable for 

subadult individuals with age-at-death estimates of 2-3 years, further research is required to 

better understand potential differences. 

Figure 7.2: External and internal features of a lesion hypothesized to be “less severe” (1352) and 
a lesion hypothesized to be “more severe” (5013) from Comparison Pair 1 (2–3 yrs).  

The comparisons between unaffected femora and femora with lesions that appeared “less 

severe” or “more severe” for individuals with age-at-death estimates of 3–5 years or 4–7 years 

included lesions that were all assessed as a Radi et al. Score 2, but displayed different extents of 

trabecular exposure on the surface. Regardless of estimated severity, the affected femora in these 

cohorts demonstrated the following compared to the unaffected counterparts: all displayed lower 

bone volume density (BV/TV), and all but one femur had lower values for trabecular number 

(Tb.N.) and higher measures of trabecular spacing (Tb.Sp.). The values for specific bone surface 

(BS/BV) and trabecular thickness (Tb.Th.) varied with no apparent association with the 

hypothesized severity. The results of these comparisons suggest that the presence of the lesions 

reflect resorptive or mixed processes; however, no clear pattern between the different estimated 

lesion severities for these femora compared to the unaffected femora was observed in this 

subsample.  

In comparing these Radi et al. Score 2 lesions hypothesized to be “less severe” or “more 

severe” based on the degree of trabecular exposure on the surface (Comparison Pairs 2 and 3), 

the “more severe” lesions displayed lower values of bone volume density (BV/TV) and 

170 

 
 
 
trabecular spacing (Tb.Sp.), and higher values of specific bone surface (BS/BV) and trabecular 

number (Tb.N.). The “more severe” lesion in Comparison Pair 3 had a lower trabecular thickness 

(Tb.Th.) value than the “less severe” lesion in that pair and while the average for the “more 

severe” lesion in Comparison Pair 2 was slightly higher than the “less severe” lesion, the values 

were similar. These trabecular differences in the “more severe” looking lesions that demonstrate 

greater exposure of the trabeculae may represent the simultaneous processes of bone resorption 

and formation during marrow hyperplasia. As discussed for lesions of the cranium, this process 

may involve thinning of trabeculae while osteoblastic activity is delegated to restoring the 

resorbed outer cortical bone with increased number of trabeculae, which can create the “hair-on-

end” appearance (Morgan, 2014; Stuart-Macadam, 1987). This interpretation may be supported 

by visual observations of the “more severe” lesions’ micro-CT images (Figure 7.3) which display 

some radial trabeculation towards the surface lacking cortical bone. These results tentatively 

suggest that more pronounced differences in the degree of trabecular exposure observed on the 

surface might reflect various stages of marrow hyperplasia in the lesion’s progression of severity. 

However, further research is required as these interpretations are based on a small number of 

femora and potential responses during these particular subadult ages. 

171 

 
 
Figure 7.3: External and internal features of lesions hypothesized to be “less severe” (1303 and 
1221), and lesions hypothesized to be “more severe” (1342 and 3032) from Comparison Pair 2 
(3–5 yrs) and Comparison Pair 3 (4–7 yrs).  

Anemia as a possible etiology of femoral cribra 

The investigation on the trabecular microarchitecture of femoral cribra yielded insights 

that support a possible etiology of chronic anemia associated with the lesion. The majority of 

active-looking femoral cribra lesions examined in this research demonstrated trabecular features 

that may indicate mixed cellular processes possibly associated with marrow hyperplasia, and 

resorptive processes. In the cohorts older than the subadult 2–3 years age group, “active” femoral 

cribra lesions tended to demonstrate lower bone volume density, fewer number of trabeculae, and 

increased trabecular spacing. The observations from the current study are consistent with Djuric 

and colleagues’ (2008) qualitative findings from histological analyses of subadult femoral cribra 

lesions. Djuric et al. observed thinned and porous cortical bone, expansion of the intertrabecular 

space, and extension of trabeculae perpendicular to the surface, which led the researchers to 

propose chronic anemia as a potential cause of femoral cribra due to marrow expansion. While 

172 

 
 
 
 
qualitative observations on the orientation of the trabeculae were not made for every lesion in the 

current study, such indications are visible in the micro-CT images including those in this chapter. 

Conclusions and future research directions 

The relatively high estimated prevalence of malaria at Mis Island, particularly in 

comparison to the sites of Tombos and the South Tombs Cemetery at Amarna, is consistent with 

the hypothesis of heightened malaria associated with the effects of settlement growth and 

expansion of irrigated land that accompanied saqia irrigation intensification in the medieval 

period. While a significant difference in estimated malarial infection between cemeteries 3-J-10 

and 3-J-11 was not observed, higher prevalences of cribra orbitalia and spinal porosity in 

cemetery 3-J-11 in the current research parallel Hurst’s (2013) findings and contribute to the 

suggestion of heightened stress prior to the late medieval period possibly associated with lower 

levels of the Nile River. The investigation into potential dynamics of malaria at Mis Island 

during the medieval period suggest a seemingly pervasive exposure to malaria shared among the 

site’s inhabitants in the context of endemic malaria. 

While the method used in this dissertation research suggests the presence of malarial 

infection in the skeletal remains, one limitation of the study is that the disease does not produce 

“pathognomonic lesions in the skeleton” (Schats, 2023:34) and its presence at the Fourth 

Cataract during the medieval period has not been confirmed. Therefore, future inquiry into 

malaria in ancient Nubia during the medieval period would benefit from confirmatory testing for 

the disease in skeletal remains using methods such as aDNA or immunological testing. 

Additionally, insight into the potential effects of anthropogenic influences on malaria in ancient 

Nubia would be supplemented by studying the possible prevalence of malaria at other sites. In 

particular, additional sites from the Meroitic, Post-Meroitic, and medieval periods could provide 

173 

 
 
a more detailed view on malarial dynamics during the time in which saqia irrigation is thought to 

have been introduced and subsequently intensified. Future research on malaria at Mis Island 

could include testing the association between the skeletal lesions associated with malaria and the 

skeletal lesions and indicators studied by Soler (2012) and Hurst (2013) for a closer examination 

of potential co-morbidities.  

The association between the porous lesions studied in this dissertation and age of the 

individual is consistent with a hypothesized etiology of anemia. Furthermore, significant 

associations among these lesions suggest shared processes involved in their formation. The three 

cribrous lesions demonstrated associations in the Mis Island sample, although the relationship 

between cribra orbitalia and the postcranial cribrous lesions may depend on the population. 

However, as observed in other studies, humeral cribra seems to form from shared processes with 

femoral cribra. Understandings of this relationship and the process of their formation would 

benefit from assessing the role of cortical thickness at these lesions’ locations. While spinal 

porosity has not received as much attention as the cribrous lesions, the findings from this 

dissertation suggest its association with the other porous lesions but may potentially be related to 

older subadult age groups. Finally, the association between periosteal reaction and the porous 

lesions may vary between populations and reflect a different response to malarial infection, as 

well as more varied etiological responses. One limitation in examining the skeletal lesions of 

interest is the potential for co-existing conditions influencing the expression of these lesions. 

Future research could focus on closely examining the relationship between lesions suggestive of 

malaria and those associated with other conditions to better understand the interaction of 

processes leading to their formation. Finally, while the cribrous lesions have been a focus of 

recent research, understandings of the five skeletal lesions implicated in malarial infection as a 

174 

 
 
whole would benefit from investigating their patterns in additional skeletal samples from varied 

contexts, including those involving other types of endemicity for malaria. 

The investigation on femoral cribra regarding its potential activity and severity provided 

preliminary support for interpreting the lesion’s status using features from the external surface. 

Applying Mensforth and colleagues’ (1978) approach to lesion activity, the sharp-edged femoral 

cribra lesions estimated to be active and the smooth-edged femoral cribra lesions estimated to be 

healing followed the expected age-related pattern that would support these designations. 

Furthermore, micro-CT analyses of a selected subsample of femora demonstrated trabecular 

differences that are consistent with the lesions that appear “healing” from the surface to have 

undergone more formative processes as compared to lesions that look “active”. The cortical 

window observed in the femora from the Mis Island sample may be associated with and 

accompany “healing” of the porous lesion; however, may not be a sign on which to base an 

interpretation.  

Regarding possible severity of femoral cribra, examination of Radi and colleagues’ 

(2013) scoring system revealed that these scores may represent “less severe” and “more severe” 

stages in the development of the lesion. In addition to the larger lesion size observed with the 

higher score, results of the micro-CT analysis tentatively suggests a progression of severity 

between the two scores that may involve resorptive processes. Furthermore, when lesions scored 

as a Radi et al. Score 2 with different extents of trabecular exposure were examined, the 

trabecular differences tentatively suggest varying stages of possible marrow hyperplasia in the 

score’s expression. However, a limitation of the micro-CT research component of this 

dissertation was the small subsample sizes, as well as a focus solely on subadults with age-at-

death estimates between 2–7 years for the investigation into severity.  

175 

 
Therefore, current understandings of femoral cribra would benefit from continued 

research into its trabecular microarchitecture in a greater number of femora across age groups, as 

well as from different populations. This line of work would also be aided by the inclusion of 

additional femora without the lesion, as processes that did not result in alterations on the bone’s 

surface may have still influenced the trabecular bone used in the comparisons. Finally, the micro-

CT investigation of femoral cribra was consistent with a possible etiology of anemia and thus 

potentially associated with the hypothesized malaria in this skeletal sample. Future micro-CT 

research from other populations could contribute further etiological considerations on femoral 

cribra, as well as humeral cribra and the association between the two lesions. 

176 

 
 
 
 
LITERATURE CITED 

Adams WY. 1964. Post-Pharaonic Nubia in the light of archaeology I. J Egypt Archaeol 50:102–
120. 

Adams WY. 1965. Post-Pharaonic Nubia in the light of archaeology II. J Egypt Archaeol 50:160–
178. 

Adams WY. 1966. Post-Pharaonic Nubia in the light of archaeology III. J Egypt Archaeol 
50:147–162. 

Adams WY. 1974. Sacred and Secular Polities in Ancient Nubia. World Archaeol 6:39–51. 

Adams WY. 1991. The United Kingdom of Makuria and Nobadia: A Medieval Nubian Anomaly. 
In: Davies WV, editor. Egypt and Africa: Nubia from Prehistory to Islam. London: British 
Museum Press in association with the Egypt Exploration Society. p 257–263.  

Adams WY. 1998. Toward a comparative study of Christian Nubian burial practice. Archeologie 
du Nil Moyen 8:13–41.  

Adams WY. 2001. Meinarti II: The Early and Classic Christian Phases. Oxford, England: 
Archaeopress. 

Ageep TB, Cox J, Hassan MM, Knols BGJ, Benedict MQ, Malcolm CA, Babiker A, El Sayed 
BB. 2009. Spatial and temporal distribution of the malaria mosquito Anopheles arabiensis in 
northern Sudan: influence of environmental factors and implications for vector control. Malaria 
Journal 8:123. 

Ahmed SM. 2014a. An introduction to the Merowe Dam Archaeological Salvage Project 
(MDASP). In: Anderson JR, Welsby DA, editors. The Fourth Cataract and Beyond: Proceedings 
of the 12th International Conference for Nubian Studies. Leuven: Peeters. p 5–8. 

Ahmed SM. 2014b. Kushites at the Fourth Cataract. In: Anderson JR, Welsby DA, editors. The 
Fourth Cataract and Beyond: Proceedings of the 12th International Conference for Nubian 
Studies. Leuven: Peeters. p 111–118. 

Al-Khafif GD, El-Banna R, Khattab N, Rashed TG, Dahesh S. 2018. The immunodetection of 
non-falciparum malaria in ancient Egyptian bones (Giza Necropolis). BioMed Research 
International 1–6. 

Allen H. 1882. A System of Human Anatomy, Sec. II, Bones and Joints. Philadelphia. 

Analyze 14.0 (AnalyzeDirect, Overland Park, KS) 

Angel JL. 1964. The reaction area of the femoral neck. Clin Orthop Relat Res 32:130–142. 

177 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Angel JL. 1966. Porotic hyperostosis, anemias, malarias and marshes in the prehistoric Eastern 
Mediterranean. Science (153):760–763. 

Aufderheide AC, Rodríguez-Martin C. 1998. The Cambridge Encyclopedia of Human 
Paleopathology. Cambridge: Cambridge University Press. 

Bain BJ, Clark DM, Wilkins BS. Infection and Reactive Changes. 2019. In: Bone Marrow 
Pathology, Fifth Edition. Hoboken, New Jersey: John Wiley & Sons, Ltd. p 109–184. 

Barbour KM. 1961. The Republic of the Sudan: A Regional Geography. London: University of 
London Press. 

Bianucci R, Mattutino G, Lallo R, Charlier P, Jouin-Spriet H, Peluso A, Hingham T, Torre C, 
Ramino Massa E. 2008. Immunological evidence of Plasmodium falciparum infection in an 
Egyptian child mummy from the Early Dynastic Period. Journal of Archaeological Science 
35:1880–1885. 

Blondiaux J, de Broucker A, Colard T, Haque A, Naji S. 2015. Tuberculosis and survival in past 
populations: A paleo-epidemiological appraisal. Tuberculosis 95:S93–S100. 

Brass M. 2015. Interactions and pastoralism along the southern and southeastern frontiers of the 
Meroitic State, Sudan. J World Prehist 28:255–288. 

Brickley M, Ives R. 2006. Skeletal manifestations of infantile scurvy. American Journal of 
Physical Anthropology 129:163–172. 

Brickley MB. 2018. Cribra orbitalia and porotic hyperostosis: A biological approach to 
diagnosis. American Journal of Physical Anthropology 167:896–902. 

Brickley MB, Ives R, Mays S. 2020. The Bioarchaeology of Metabolic Bone Disease. Elsevier. 

Buikstra JE, Ubelaker DH. 1994. Standards for Data Collection from Human Skeletal Remains. 
Fayetteville: Arkansas Archaeological Survey. 

Butzer KW. 1976. Early Hydraulic Civilization in Egypt: A Study in Cultural Ecology. Chicago: 
University of Chicago Press. 

Buzon MR, Sanders KS. 2016. Skeletal evidence of malaria at Tombos: disease patterns from the 
New Kingdom through Napatan Periods in Upper Nubia. Poster session presented at the 
American Association of Physical Anthropologists, Atlanta, GA. 

Carter R, Mendis KN. 2002. Evolutionary and historical aspects of the burden of malaria. 
Clinical Microbiology Reviews 15(4):564–594. 

178 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Coates D, Redding-Coates TA. 1981. Ecological problems associated with irrigation canals in the 
Sudan with particular reference to the spread of bilharziasis, malaria and aquatic weeds and the 
ameliorative role of fishes. International Journal of Environmental Studies 16:207–212. 

Dalton M, Spencer N, Macklin MG, Woodward, JC, Ryan P. 2023. Three thousand years of river 
channel engineering in the Nile Valley. Geoarchaeology 38:565–587.  

DeWitte SN, and Stojanowski CM. 2015. The Osteological Paradox 20 Years Later: Past 
Perspectives, Future Directions. Journal of Archaeological Research 23:397–450. 

Djuric M, Milovanovic P, Janovic A, Draskovic M, Djukic K, Milenkovic P. 2008. Porotic 
lesions in immature skeletons from Stara Torina Late Medieval Serbia. International Journal of 
Osteoarcheology 18:458–75. 

Djuric M, Janovic A, Milovanovic P, Djukic K, Milenkovic P, Draskovic M, Roksandic M. 2010. 
Adolescent health in medieval Serbia: signs of infectious diseases and risk of trauma. Journal of 
Comparative Human Biology 61:130–149. 

Djuric M, Zagorac S, Milovanovic P, Djonic D, Nikolic S, Hahn M, Zivkovic V, Bumbasirevic 
M, Amling M, Marshall RP. 2013. Enhanced trabecular micro-architecture of the femoral neck in 
hip osteoarthritis vs. health controls: a micro-computed tomography study in postmenopausal 
women. International Orthopaedics 37:21–26. 

Dukeen MYH, Omer SM. 1986. Ecology of the malaria vector Anopheles arabiensis Patton 
(Diptera: Culiciadae) by the Nile in northern Sudan. Bulletin of Entomological Research 76:451–
467. 

Edwards DN. 1994. Post Meroitic (‘X-Group’) and Christian burials at Sesebi, Sudanese Nubia. 
J Egypt Archaeol 80:159–178. 

Edwards DN. 1996. The Archaeology of the Meroitic State: New Perspectives on its Social and 
Political Organization. Oxford: Tempus Reparatum: Distributed by Hadrian Books. 

Edwards DN. 1999. Meroitic Settlement. In: Welsby DA, editor. Recent Research in Kushite 
History and Archaeology: Proceedings of the 8th International Conference for Meroitic Studies. 
British Museum Occasional Paper. London: British Museum Press. p 19–67. 

Edwards DN. 2001. The Christianisation of Nubia: some archaeological pointers. Sudan & 
Nubia 5:89–96.  

Edwards DN. 2004. The Nubian Past: an archaeology of the Sudan. London, United Kingdom: 
Routledge. 

Edwards DN. 2014. Medieval Nobadia. In: Anderson JR, Welsby DA, editors. The Fourth 
Cataract and Beyond: Proceedings of the 12th International Conference for Nubian Studies. 
Leuven; Paris; Wapole, MA: Peeters. p 171–182. 

179 

 
 
 
 
 
 
 
 
 
 
 
 
 
El-Najjar MY, Ryan DJ, Turner CG, and Lozoff B. 1976. The etiology of porotic hyperostosis 
among the prehistoric and historic Anasazi Indians of Southwestern United States. American 
Journal of Physical Anthropology 44:477–487. 

el-Tayeb M. 2010. The Post-Meroitic from Kirwan to the present. Sudan & Nubia 14:2–14. 

Elmesery. 2020. Evolution of irrigation system, tools, and technologies. In: Omran EE, Negm 
AM, editors. Technological and Modern Irrigation Environment in Egypt: Best Management 
Practices and Evaluation. p 59–73. 

Fajardo RJ, Müller R. 2001. Three-dimensional analysis of nonhuman primate trabecular 
architecture using micro-computed tomography. American Journal of Physical Anthropology 
115:327–336. 

Farfour E, Charlotte F, Settegrana C, Miyara M, Buffet P. 2012. The extravascular compartment 
of the bone marrow: a niche for Plasmodium falciparum gametocyte maturation? Malaria Journal 
11(285):1–4. 

Finnegan M. 1978. Non-metric variation of the infracranial skeleton. Journal of Anatomy 
125:23–37. 

Finnegan M, Faust MA. 1974. Variants of the femur. Research Report 14: Bibliography of 
Human and Non-human, Non-Metric Variation. 3. 

Fuller DO, Parenti MS, Hassan AN, Beier JC. 2012. Linking land cover and species distribution 
models to project potential ranges of malaria vectors: an example using Anopheles arabiensis in 
Sudan and Upper Egypt. Malaria Journal 11:264. 

Fuller DQ, Lucas L. 2020. Savanna on the Nile: Long-term agricultural diversification and 
intensification in Nubia. In: Emberling G, Williams BB, editors. The Oxford Handbook of 
Ancient Nubia. New York: Oxford University Press. p 927–953. 

Garcia E, Berrocal MI, Baxarias J, Campillo D, Subirá ME. 2002. Cribra and trace 
elements in the Prat de la Riba Necropolis (Tarragona, Spain, 3rd–5th centuries AD). 
Antropologia Portuguesa 19: 71–83. 

Gibbon VE, Buzon MR. 2018. A diachronic examination of biomechanical changes in skeletal 
remains from Tombos in ancient Nubia. Journal of Comparative Human Biology 69:158–166. 

Ginns A. 2006. Preliminary report on the excavations conducted on Mis Island (AKSC), 2005–
2006. Sudan & Nubia, 10:13–19. 

Ginns A. 2007. Preliminary report on the second season of excavations conducted on Mis Island 
(AKSC). Sudan & Nubia 11:20–26. 

Ginns A. 2010a. Medieval cemetery 3-J-10. Sudan Archaeol. Res. Soc. Draft publications. 

180 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
Ginns A. 2010b. Medieval cemetery 3-J-11. Sudan Archaeol. Res. Soc. Draft publications. 

Godlewski W. 2014. The Kingdom of Makuria. In: Anderson JR, Welsby DA, editors. The 
Fourth Cataract and Beyond: Proceedings of the 12th International Conference for Nubian 
Studies. Leuven; Paris; Wapole, MA: Peeters. p 155–170. 

Göhring A. 2021. Allen’s fossa—An attempt to dissolve the confusion of different nonmetric 
variants on the anterior femoral neck. International Journal of Osteoarchaeology. 
https://doi.org/10.1002/oa.2968 

Gomes RAMP, Petit J, Dutour O, Santos AL. 2022. Frequency and co-occurrence of porous 
skeletal lesions in identified non-adults from Portugal (19th to 20th centuries) and its association 
with respiratory infections as cause of death. International Journal of Osteoarchaeology 32:1061–
1072.   

Gowland RL, Western AG. 2012. Morbidity in the marshes: using spatial epidemiology to 
investigate skeletal evidence for malaria in Anglo-Saxon England (AD 410–1050). American 
Journal of Physical Anthropology 147:301–311. 

Hawass Z, Gad YZ, Ismail S, Khairat R, Fathalla D, Hasan N, Ahmed A, Elleithy H, Ball M, 
Gaballah F, et al. 2010. Ancestry and pathology in King Tutankhamun’s family. JAMA: The 
Journal of the American Medical Association 303:638–647. 

Hengen OP. 1971. Cribra orbitalia: pathogenesis and probable etiology. Journal of Human 
Evolution 22:57–75. 

Hildebrand T, Laib A, Müller R, Dequeker J, Rüegsegger P. 1999. Direct three-dimensional 
morphometric analysis of human cancellous bone: microstructural data from spine, femur, iliac 
crest, and calcaneus. Journal of Bone and Mineral Research 14(7):1167–1174. 

Holland TD, and O’Brien MJ. 1997. Parasites, porotic hyperostosis, and the implications of 
changing perspectives. American Antiquity 62(2):183–193. 

Hurst CV. 2013. Growing up in Medieval Nubia: health, disease, and death of a medieval 
juvenile sample from Mis Island, PhD dissertation. East Lansing: Michigan State University. 

Joice R, Nilsson SK, Montgomery J, Dankwa S, Egan E, Morahan B, Seydel KB, Bertuccini L, 
Alana P, Williamson KC, et al. 2014. Plasmodium falciparum transmission stages accumulate in 
the human bone marrow. Science Translational Medicine 6(244):244re5. 

Judd M. 2004. Gabati: Health in Transition. Sudan & Nubia 8:84–89. 

Kate BR. 1963. The incidence and cause of cervical fossa in Indian femora. Journal of the 
Anatomical Society of India 12(2):69–76. 

181 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
Kemp B, Stevens A, Dabbs GR, Zabecki M, Rose JC. 2013. Life, death and beyond in 
Akhenaten’s Egypt: excavating the South Tombs Cemetery at Amarna. Antiquity 87(335):64. 

Kirwan L. 1984. The birth of Christian Nubia: some archaeological problems. Revista degli 
Studi Orientali 63:119–134.  

Kirwan L. 2002. Post-Meroitic Nubia—A Reappraisal. In: Kirwan L, Hägg T, Törö, L, Welsby 
DA, editors. Studies on the History of Late Antique and Christian Nubia. Aldershot: Ashgate 
Publishing Limited. p 1–9. 

Kivell TL, Skinner MM, Lazenby R, Hublin J. 2011. Methodological considerations for 
analyzing trabecular architecture: an example from the primate hand. Journal of Anatomy 
218:209–225. 

Kotsick E.L. 1963. Facets and imprints on the upper and lower extremities of femora from a 
Western Nigerian population. J. Anat. 97(3): 393–402. 

Kricun ME. 1985. Red-yellow marrow conversion: Its effect on the location of some solitary 
bone lesions. Skeletal Radiology 14:10–19. 

Lalremruata A, Ball M, Bianucci R, Welte B, Nerlich AG, Kun J.F.J, Pusch CM. 2013. Molecular 
identification of falciparum malaria and human tuberculosis co-infections in mummies from the 
Fayum Depression (Lower Egypt). PLoS ONE 8(4):e60307. 

Lamikanra AA, Brown D, Potocnik A, Casals-Pascual C, Langhorne J, Roberts DJ. 2007. 
Malaria anemia: of mice and men. Blood 110(1):18–28. 

Lazenby RA, Skinner MM, Kivell TL, Hublin J. 2011. Scaling VOI size in 3D µCt studies of 
trabecular bone: a test of the over-sampling hypothesis. American Journal of Physical 
Anthropology 144:196–203. 

Lee MSJ, Maruyama K, Fujita Y, Konishi A, Lelliott PM, Itagaki S, Horii T, Lin J, Khan SM, 
Kuroda E, et al. 2017. Plasmodium products persist in the bone marrow and promote chronic 
bone loss. Science Immunology 2:eaam8093. 

Lewis M. 2018. Hemopoietic and metabolic disorders. In: Paleopathology of Children: 
Identification of Pathological Conditions in the Human Skeletal Remains of Non-Adults. 
London, United Kingdom: Academic Press. p 193–223. 

Lohwasser A. 2014. Kush and Her Neighbors Beyond the Nile Valley. In: Anderson JR, Welsby 
DA, editors. The Fourth Cataract and Beyond: Proceedings of the 12th International Conference 
for Nubian Studies. Leuven; Paris; Wapole, MA: Peeters. p 125–134. 

Loufouma-Mbouaka A, Binder M, Noedl H, Gamble M. 2020. Evaluation of rapid diagnostic 
tests and Enzyme Linked Immunoassay in the detection of malaria in ancient human remains. 
Journal of Archaeological Science 116:105118 

182 

 
 
 
 
 
 
 
 
 
 
 
 
 
Malagelada F, Tayar R, Barke S, Stafford G, Field RE. 2015. Anatomy of the zona orbicularis of 
the hip: a magnetic resonance study. Surgical and Radiologic Anatomy 37:11–18. 

Malcolm CA, Welsby DA, El Sayed BB. 2007. SIT for the malaria vector Anopheles arabiensis 
in northern state, Sudan: an historical review of the field site. In: Vreysen MJB, Robinson AS, 
Hendrichs J, editors. Area-Wide Control of Insect Pests. From Research to Rield 
Implementation. Dordrecht: Springer. p 361–372. 

Mangas-Carrasco E, López-Costas O. 2021. Porotic hyperostosis, cribra orbitalia, femoralis and 
humeralis in Medieval NW Spain. Archaeological and Anthropological Sciences 13:169. 

Mann RW, Hunt DR, Lozanoff S. 2016. Photographic regional atlas of nonmetric traits and 
anatomical variants in the human skeleton. Springfield, Illinois: Charles C. Thomas Publisher, 
Ltd. 

Marchand C, Chen H, Buschmann MD, Hoemann CD. 2011. Standardized three-dimensional 
volumes of interest with adapted surfaces for more precise subchondral bone analyses by micro-
computed tomography. Tissue Engineering 17(4):475–484. 

Marciniak S, Herring DA, Sperduti A, Poinar HN, Prowse TL. 2018. A multi-faceted 
anthropological and genomic approach to framing Plasmodium falciparum malaria in Imperial 
period central-southern Italy (1st-4th c. CE). Journal of Anthropological Archaeology 49:210–
224. 

Mays S. 2018. How should we diagnose disease in paleopathology? Some epistemological 
considerations. International Journal of Paleopathology 20:12–19. 

McIlvaine BK. 2015. Implications of reappraising the iron-deficiency anemia hypothesis. 
International Journal of Osteoarchaeology 25:997–1000. 

Mendiela S, Rissech C, Haber M, Pujol-Bayona A, Lomba J, Turbón D. 2014. Childhood growth 
and health in Camino del Molino (Caravaca de la Cruz, Murcia, Spain) a collective burial of the 
III Millennium cal. BC. A preliminary approach. In: Adés, Ao, editors. Estudis D’evolució 
Etolgia 101–106. 

Mensforth RP, Lovejoy OC, Lallo JW, Armelagos GJ. 1978. The role of constitutional factors, 
diet, and infectious disease in the etiology of porotic hyperostosis and periosteal reactions in 
prehistoric infants and children. Medical Anthropology 2(1):1–59. 

Meyer AW. 1924. The “cervical fossa” of Allen. American Journal of Physical Anthropology 
7:257–269. 

Meyer AW. 1934. The genesis of the fossa of Allen and associated structures. American Journal 
of Anatomy 55:469–510. 

183 

 
 
 
 
 
 
 
 
 
 
 
 
 
Miller RL, Ikram S, Armelagos GJ, Walker R, Harer WB, Shiff CJ, Bagget D, Carrigan M, Maret 
SM. 1994. Diagnosis of plasmodium falciparum infections in mummies using the rapid manual 
ParaSight™-F test. Transactions of the Royal Society of Tropical Medicine and Hygiene 88:31–
32. 

Miquel-Feucht MJ, Polo-Cerdá M, Villalaín-Blanco JD. 1999. El sindrome criboso: cribra 
femoral vs. cribra orbitaria. In: Sistematización metodológica en Paleopatología, Actas V 
Congreso Nacional AEP, Sánchez JA (ed.). Asociación Española de Paleopatología, Alcalá la 
Real: Jaén, Spain. p 221–37. 

Molleson T, Andrews P, Boz B, Sofaer Derevenski J, Pearson J. 1998. Human remains up to 
1998. Catalhoyuk 1998 Archive Report. 

Moreau R, Malu DT, Dumais M, Dalko E, Gaudreault V, Roméro H, Martineau C, Kevorkova O, 
Dardon JS, Dodd EL, Bohle DS, Scorza T. 2012. Alterations in bone and erythropoiesis in 
hemolytic anemia: Comparative study in bled, phenylhydrazine-treated and Plasmodum-infected 
mice. PLoS ONE 7(9): e46101. doi:10.1371/journal.pone.0046101 

Morgan JA. 2014. The methodological and diagnostic applications of micro-CT to 
paleopathology: A quantitative study of porotic hyperostosis, PhD thesis. London, Ontario: 
University of Western Ontario. 

Naidoo S, London L, Burdorf A, Naidoo RN, Kromhout H. 2011. Occupational activities 
associated with a reported history of malaria among women working in small-scale agriculture in 
South Africa. American Journal of Tropical Medicine and Hygiene 85(5):805–810. 

Näser C. 2007. The Humboldt University Nubian expedition 2005: Works on Sherari and Us. In: 
Näser C, Lange M, editors. Proceedings of the Second International Conference on the 
Archaeology of the Fourth Nile Cataract. Meroitica 23:118-133. 

Nathan H, Haas N. 1966. “Cribra orbitalia” A bone condition of the orbit of unknown nature: 
Anatomical study with etiological considerations. Israel Journal of Medical Sciences 2(2):171–
191. 

Nerlich AG, Schraut B, Dittrich S, Jelinek T, Zink AR. 2008. Plasmodium falciparum in ancient 
Egypt. Emerging Infectious Diseases 14:1317–1319. 

Obaldia N III, Meibalan E, Sa JM, Ma S, Clark MA, Mejia P, Moraes Barros RR, Otero W, 
Ferreira M, Mitchell JR, et al. 2018. Bone marrow is a major parasite reservoir in Plasmodium 
vivax infection. mBio 9(3):e00625-18. 

Obłuski A. 2020. The Archaeology of Medieval Nubian Kingdoms. In: Emberling G, Williams 
BB, editors. The Oxford Handbook of Ancient Nubia. New York: Oxford University Press. p 
829–846. 

184 

 
 
 
 
 
 
 
 
 
 
 
 
Odgers P.N.B. 1931. Two details about the neck of the femur: (1) The eminentia. (2) The 
empreinte. Journal of Anatomy 65:352–362. 

Oftadeh R, Perez-Viloria M, Villa-Camacho JC, Vaziri A, Nazarian A. 2015. Biomechanics and 
mechanobiology of trabecular bone: a review. Journal of Biomechanical Engineering 
137(1):010802-1–010802-15. 

Ortner DJ. 2003. Identification of Pathological Conditions in Human Skeletal Remains. San 
Diego, CA: Academic Press. 

Oxenham MF, Cavill I. 2010. Porotic hyperostosis and cribra orbitalia: the erythropoietic 
response to iron-deficiency anemia. Anthropological Science 118:199–200. 

Paredes J, Ferreira MT, Wasterlain SN. 2015. Growth problems in a skeletal sample of children 
abandoned at Santa Casa da Misericórdia, Faro, Portugal (16th–19th centuries). Anthropological 
Science 123:49–59. 

Posit team. 2022. RStudio: Integrated Development Environment for R. Posit Software, PBC, 
Boston, MA. URL http://www.posit.co/. 

Rabino Massa E, Cerutti N, Marin A, Savoia D. 2000. Malaria in ancient Egypt: 
paleoimmunological investigation on predynastic mummified remains. Chungara 32(1):7–9. 

Radi N, Mariotti V, Riga A, Zampetti S, Villa C, Belcastro MG. 2013. Variation of the anterior 
aspect of the femoral head-neck junction in a modern human identified skeletal collection. 
American Journal of Physical Anthropology 152:261–72. 

Ratti V, Wallace DI. 2020. A malaria transmission model predicts holoendemic, hyperendemic, 
and hypoendemic transmission patterns under varied seasonal vector dynamics. Journal of 
Medical Entomology 57(2):568–584. 

Redfern R. 2010. Human Bioarchaeological Database. London: The British Museum 

Reisner GA. 1910. The Archaeological Survey of Nubia, Report for 1907–1908. Egypt Survey 
Department, Cairo. 

Rivera F, Mirazón Lahr M. 2017. New evidence suggesting a dissociated etiology for cribra 
orbitalia and porotic hyperostosis. American Journal of Physical Anthropology 164:76–96. 

Ruffini GR. 2012. Medieval Nubia: A Social and Economic History. New York; New York: 
Oxford University Press. 

Sabbatani S, Manfredl R, Fiorino S. 2010. Malaria infection and human evolution. Le Infezioni 
in Medicina 1:56–74. 

185 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Sandberg OH, Aspenberg P. 2016. Inter-trabecular bone formation: a specific mechanism for 
healing in cancellous bone. Acta Orthopaedica 87(5):459–465. 

Schats R. 2021. Cribriotic lesions in archaeological human skeletal remains. Prevalence, co-
occurrence, and association in medieval and early modern Netherlands. International Journal of 
Paleopathology 35:81–89. 

Schats R. 2023. Developing an archaeology of malaria. A critical review of current approaches 
and a discussion on ways forward. International Journal of Paleopathology. 41:32–42. 

Scheidel W. 2001. Death on the Nile: Disease and the Demography of Roman Egypt. Leiden, 
Netherlands: Brill Academic Publishers. 

Schofield G. 1959. Metric and morphological features of the femur of the New Zealand Maori. 
The Journal of the Royal Anthropological Institute of Great Britain and Ireland 89(1):89–105. 

Setzer TJ. 2014. Malaria detection in the field of paleopathology: a meta-analysis of the state of 
the art. Acta Tropica 140:97–104. 

Shinnie PL, Shinnie M. 1965. New light on medieval Nubia. J Afr Hist 6:263–273.  

Smith NE. 2015. The paleoepidemiology of malaria in the Ancient Near East, PhD dissertation. 
Fayetteville: University of Arkansas. 

Smith-Guzmán NE. 2015a. The skeletal manifestation of malaria: an epidemiological approach 
using documented skeletal collections. American Journal of Physical Anthropology 158:624–
635. 

Smith-Guzmán NE. 2015b. Cribra orbitalia in the ancient Nile Valley and its connection to 
malaria. International Journal of Paleopathology 10:1–12. 

Smith-Guzmán NE, Rose JC, Kuckens K. 2016. Beyond the differential diagnosis: new 
approaches to the bioarchaeology of the Hittite plague. In: Zuckerman MK, Martin DL, editors. 
New Directions in Biocultural Anthropology. Hoboken, New Jersey: John Wiley & Sons, Inc. p 
295–316. 

Smith-Guzmán NE, Dabbs GR, Davis HS, Shidner AE. 2018. Distinguishing skeletal lesions of 
malaria from comorbidities and coexisting metabolic conditions at Amarna, Egypt. Poster 
presented at the 87th Annual Meeting of the American Association of Biological Anthropologists. 

Snow RW, Guerra CA, Mutheu JJ, Hay SI. 2008. International funding for malaria control in 
relation to populations at risk of stable Plasmodium falciparum transmission. PLoS Med 5(7): 
e142.doi:10.1371/journal.pmed.0050142 

Soler A. 2012. Life and death in a Medieval Nubian farming community: the experience at Mis 
Island, PhD dissertation. East Lansing: Michigan State University. 

186 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
Spaulding J. 1995. Medieval Christian Nubia and the Islamic world: a reconsideration of the 
Baqt Treaty. Int J Afr Hist Stud 28:577–594.  

Stuart-Macadam P. 1987. Porotic hyperostosis: new evidence to support the anemia theory. 
American Journal of Physical Anthropology 74:512–526. 

Stuart-Macadam P. 1992. Anemia in past human populations. In: Stuart-Macadam P, and Kent S, 
editors. Diet Demography and Disease, Changing Perspectives on Anemia. New York: Aldine. p 
151–170. 

Tatem AJ, Guerra CA, Kabaria CW, Noor AM, Hay SI. 2008. Human population, urban 
settlement patterns and their impact on Plasmodium falciparum malaria endemicity. Malaria 
Journal 7:218. 

Thomas R. 2008. The origin and use of ceramics on the islands of Mis and Umm Muri, in the 
Late Meroitic to Christian periods. Sudan & Nubia 12:64–73.  

Trigger BG. 1969. The Royal tombs at Qustul and Ballana and their Meroitic antecedents. The 
Journal of Egyptian Archaeology 55:117–128. 

Tsutsumi M, Nimura A, Utsunomia H, Akita K. 2021. Dynamic changes of the joint capsule in 
relation to the zona orbicularis: An anatomical study with possible implications for hip stability 
mechanism. Clinical Anatomy 34:1157–1164. 

Tumwiine J, Mugisha JYT, Luboobi LS. 2007. On oscillatory pattern of malaria dynamics in a 
population with temporary immunity. Computational and Mathematical Methods in Medicine 
8(3):191–203. 

Wagner FV, Negráo JR, Campos J, Ward SR, Parviz H, Trudell DJ, Resnick D. 2012. Capsular 
ligaments of the hip: Anatomic, histologic, and positional study in cadaveric specimens with MR 
arthrography. Radiology 263(1):189–198. 

Wagner WE, Gillespie BJ III. 2019. Using and interpreting statistics in the social, behavioral, and 
health sciences. Thousand Oaks, CA: SAGE Publications, Inc. 

Walker PL, Bathurst RR, Richman R, Gjerdrum T, and Andrushko VA. 2009. The causes of 
porotic hyperostosis and cribra orbitalia: a reappraisal of the iron-deficiency anemia hypothesis. 
American Journal of Physical Anthropology 139:109–25. 

Wapler U, Crubézy E, Schultz M. 2004. Is cribra orbitalia synonymous with anemia? Analysis 
and interpretation of cranial pathology in Sudan. American Journal of Physical Anthropology 
123:333–339. 

Welsby DA. 1996. The Kingdom of Kush: The Napatan and Meroitic Empires. London: 
Published for the Trustees of the British Museum by British Museum Press. 

187 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
Welsby DA. 2002. The Medieval Kingdoms of Nubia: Pagans, Christians and Muslims Along the 
Middle Nile. London: British Museum Press.  

Welsby DA. 2003. The Amri to Kirbekan Survey: the 2002–2003 season. Sudan Nubia 7:26–32. 

Welsby DA. 2004. The Archaeology and History of Sudan. In: Welsby DA, Anderson JR, editors. 
Sudan: Ancient Treasures. British Museum Press. p 12–19. 

Welsby DA. 2006. The Merowe Dam Archaeological Survey. Excavations in the vicinity of ed-
Doma (AKSE), 2005–2006. Sudan Nubia 10:8–13. 

Welsby DA. 2014. The Kingdom of Alwa. In: Anderson JR, Welsby DA, editors. The Fourth 
Cataract and Beyond: Proceedings of the 12th International Conference for Nubian Studies. 
Leuven; Paris; Wapole, MA: Peeters. p 183–200. 

Welsh F. 2013. Architecture and life in villages of the Fourth Nile Cataract in the region of et-
Tereif. Draft report. 

White NJ. 2018. Anaemia and malaria. Malaria Journal 17:371 https://doi.org/10.1186/s12936-
018-2509-9 

Whitehouse WJ. 1975. Scanning electron micrographs of cancellous bone from the human 
sternum. Journal of Pathology 116:213–224. 

Whitehouse WJ, Dyson ED. 1974. Scanning electron microscope studies of trabecular bone in 
the proximal end of the human femur. Journal of Anatomy 118:417–444. 

Wickramasinghe SN, Abdalla SH. 2000. Blood and bone marrow changes in malaria. Bailliére’s 
Clinical Haematology 13(2):277–299. 

Wood JW, Milner GR, Harpending HC, and Weiss KM. 1992. The Osteological Paradox. Current 
Anthropology 33(4):343–370. 

Żurawski B. 2006. Nubian mortuary complex of the Christian Period. In: Caneva I, Roccati A, 
editors. Acta Nubica: Proceedings of the X International Conference of Nubian Studies, Rome 9-
14 September 2002. Rome: Universitá di Roma “La Sapienza.” 

Żurawski B. 2014. The Fourth Cataract in the Medieval Period. In: Anderson JR, Welsby DA, 
editors. The Fourth Cataract and Beyond: Proceedings of the 12th International Conference for 
Nubian Studies. Leuven; Paris; Wapole, MA: Peeters. p 135–154. 

188