|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
BIOCULTURAL VARIATION OF SKELETAL TRAUMA
IN CONTEMPORARY GREEKS
SUZANNE MARIE ABEL
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Suzanne Marie Abel
I thank my committee members Drs. Anthony Falsetti, Michael Warren, Sue
Boinski, and Thomas Hollinger for their help in making this dissertation possible. They
each provided necessary support in one way or another, and I greatly admire their
individual talents in anthropology and anatomy.
This research was funded in part by University of Florida Foundation grants, the
William R. Maples Memorial Scholarship, and by Drs. Anthony Falsetti and Michael
My parents Vernon Abel, Gerry Buchanan and Joyce Buchanan have given me 35
years of love, emotional support and financial help, for which I am eternally grateful.
Somehow they knew I would stick with it and finish, even when I was unsure.
I thank Dr. Sotiris Manolis for allowing me access to the skeletal collections at the
University of Athens and for his gracious hospitality. Constantine Eliopoulos provided
daily camaraderie in the lab while making the best coffee in all of Greece over a simple
lab burner. Anna Lagia also offered valuable information about the skeletal collections.
I would like to thank all the citizens of Greece I met during my time in their country.
They are truly peerless in their warmth and hospitality.
Stateside, I thank Dr. Ted Rathbun, whose infective enthusiasm about anthropology
and teaching changed my life. Anyone who has had the pleasure of meeting Ted knows
there is no one else like him. HK forever!
This dissertation could not have been done were it not for Ann Ross, Katie
Jemmott, and the rest of the CAPHIL gang. I thank them for their faithful support, the
Halloween parties and the times we sat around the lab and commiserated.
Last but certainly not least, I thank Dr. Wolf Bueschgen. He was my classmate and
fellow traveler years ago, and he continues to travel with me through life now as my
husband. What patience he possesses to put up with me.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iii
LIST OF TABLES ............. ........... ...... .................. vii
LIST OF FIGURES ......... ....... .................... .......... ....... ............ ix
A B S T R A C T ............................................ ... ......... ................................... x
1 IN T R O D U C T IO N ............................................................................... .............. ..
2 MECHANISMS OF SKELETAL TRAUMA............... ...................................5
C h ild h o o d T rau m a .............................................................................. .. .......... .. .. .
C h ild A b u se .......................................................... ................ 7
A du lt T rau m a ...................................... ................................................ .. 9
Interpersonal Violence.................. .. .. ............ ..... ....... 12
3 RELEVENT BIOARCHAEOLOGICAL LITERATURE .......................................14
Skeletal Traum a Studies Using Greek Collections ..................... ........................... 16
F ra n c h th i C av e ....................................................................................................... 1 6
C a ta l H u y u k ...................................................................................................1 7
Lerna ............... ..................................................... 18
K hirokitia ..............................19..............................................
K a rata s ................................................................2 0
Cephallenia ..................................................... .... 20
O th e r site s .............. .................................... ..................................................2 1
4 RELEVANT CONTEMPORARY LITERATURE .............................................. 24
5 SOCIOECOLOGICAL ASPECTS OF PRIMATE BEHAVIOR AND TRAUMA ..30
Population D ensity and Resource Availability ...........................................................31
In fa n tic id e ...................................................................................................... 3 2
Coalitionary Killing .................. 34......... ....... .........34
P rim ate T rau m a ................................................................3 5
6 M ATERIALS AND M ETHODS ........................................ ......................... 36
The University of Athens Sam ple............................................................................36
M e th o d s ..............................................................................3 7
7 R E S U L T S .............................................................................4 1
Descriptive Demographics of the University of Athens Skeletal Sample..................41
S ex an d A g e ............................................................................... 4 1
Cause of Death ................................ ........ 43
O occupation ...................... ............... .. .. .................................43
T raum a F requencies......... ........................................................................ ... ..45
Statistical Results ............... ......... .............. ................... 47
D iet an d h e alth ...................... .. ............. .. ...............................................5 1
Health Care Funding....... ...... ....... ............ ......... .............. .. ........ ..52
O osteoporosis and Fractures...................... .... ............ ..................... ............... 57
Geography ......................... ....................... 58
P o p u latio n ........... ...... ..... .. ................. .................................................5 8
Com prison to Archaeological Sam ples................................. ...................... 71
8 D IS C U S S IO N ...................... .. ............. .. ................................................7 4
Sum m ary of findings .................... .. ........... ............. .. ..... ........ .... ....... ..74
D iet an d D ise a se ................................................................................................... 7 5
Substance Abuse ................................................................................ .............................. 76
Child Abuse and Other Interpersonal Violence ..................................................77
Suggestions For Further Research................................................... .. ................ .. 79
A ZOGRAFOU CEMETERY AND MAUSOLEUM, ATHENS..............................80
B SAMPLE DOCUMENTATION FORMS ................................................83
L IST O F R E F E R E N C E S ........................................................................ .....................9 1
BIOGRAPHICAL SKETCH ............................................................. ............... 105
LIST OF TABLES
3-1 Summary of skeletal analyses of Greek and Turkish archaeological sites
published by A ngel ......................... ....... .... .. ...... ............ 23
7-1 Age category frequency by sex for the UA sample ............................................42
7-2 Percentages of deaths by age category for UNECE data (1997) and the UA
skeletal sam ple ............ ...... ....... ... ......... ...................... 42
7-3 Cause of death categories and frequencies for the UA skeletal sample ..............44
7-4 Select cause of death categories and percentages for all Greek deaths in 1997
and the U A skeletal sam ple......................................................... ............... 45
7-5 Occupation category, number and percentage for the UA skeletal sample...........45
7-6 Select employment sectors, frequencies and percentages for Greece, 1997..........45
7-7 Trauma frequency percentages by element for the UA skeletal sample................46
7-8 Select socioeconomic indicators for Greece, neighboring Mediterranean
countries and the U .S. ...................... .. .... ........................................... 49
7-9 Crime indicators, per 100,000 population, for Greece and the U.S., year 2000....50
7-10 Violent death rates per 100,000 population for males/females, Greece and the
U .S., year 1997................................................... ..................... ..... ....... 50
7-11 Total population increases for Greece and greater Athens for select decades.......59
7-12 The 1997 death rates per 100,000 population for males/females, Greece and the
U united States ..................................................................... ..........61
7-13 Life expectancy at birth for males/females: Greece, neighboring
Mediterranean countries and the United States .............................................. 62
7-14 Loci used in the loglinear model and their respective skeletal elements ..............67
7-15 Cross-classification of craniofacial, thoracic and appendicular trauma in the
U A sa m p le ....................................................... ................ 6 8
7-16 Three-dimensional contingency table using the UA data................................69
7-17 Research questions examined in this study and their results...............................73
LIST OF FIGURES
4-1 General relationship between socioeconomic status and trauma frequency among
hospital patients ............ ..... ...... .... ..................... 25
7-1 Age category frequencies by sex for the UA sample............... ....... ............42
7-2 Fractured distal left radius (left) and normal contralateral right radius (right),
c a rp a l v iew ...............................................................................................................5 4
7-3 Fractured distal left radius (left) and normal contralateral radius (right), lateral
v ie w ...................................... ................................... ................ 5 5
7-4 Fractured right proximal femur .................... .......................... .... ........... 55
7-5 Orthopaedic fixation in a fractured greater trochanter and neck of the right femur,
an terio r v iew ...................................... ............ ............... ................ 5 6
7-6 Fractured neck and proximal diaphysis of the right femur ................................56
7-7 Craniofacial trauma caused by a motor vehicle accident in a 43-year-old male......60
7-8 Tibial fracture in a 34-year-old m ale.................................... ........................ 61
7-9 Age category frequencies for all Greeks, 1990 and 2000 ..................................63
7-10 Left femur from the UA sample with Allen's fossa located inferior to the anterior
articular surface of the femoral head ............................ .................................... 72
A-1 Primary burials maintained in Zografou Cemetery, looking northwest ................. 80
A-2 Extreme eastern perimeter of Zografou Cemetery ................................................80
A-3 One of two communal subterranean bone vaults on the southern perimeter of
Z ografou C em etery ..................... .................. .............. .... ........... .. 1
A -4 Close-up of a subterranean vault.................................... ........................... ......... 81
A-5 Interior of the Zografou Cemetery Mausoleum ...................................................82
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
BIOCULTURAL VARIATION OF SKELETAL TRAUMA
IN CONTEMPORARY GREEKS
Suzanne Marie Abel
Chair: Anthony B. Falsetti
Major Department: Anthropology
Bioarchaeological research has highlighted the importance of eco-geographic
factors on skeletal trauma frequencies in ancient populations. Clinical research
addressing trauma patterns in hospital populations shows that socioeconomic variables
such as income, education level, minority/ethnic status, and substance abuse are
significantly correlated with overall trauma.
Research based on the behavioral mechanism of trauma in contemporary skeletal
samples is notably scarce, mainly because of the limited number of available collections.
As such, most of these studies are limited to fracture patterns of isolated elements or
skeletal regions, and assume that injuries found in certain regions are predictive of overall
accidental or aggressive behavior.
A collection (n=121) of fully represented and documented individuals of recent
death from Athens, Greece, presented the unique opportunity to test the influence of
various demographic, ecological, and environmental variables on individual trauma in a
contemporary population. The sample consists of the remains of individuals donated to
the Biological Laboratory at the University of Athens where the collection is currently
housed for research purposes.
Morphological analysis of this sample suggests that most of the skeletal trauma
present is the result of cumulative accidental episodes. Observed injuries show no
distinct constellations suggesting intentional, interpersonal violence. Significant trauma
clustering is seen in the thoracic vertebrae, ribs, radii, and femora of both males and
females. However, these injuries are somewhat typical in an aged, osteoporotic
population such as the UA sample. Crosstabulations, ANOVA, and loglinear analyses
show statistical associations between sex and trauma location, and between sex and cause
of death category, and some degree of association among concomitant trauma locations.
The UA collection seems to mirror that of contemporary Greek society at large and
can be regarded as a subset of a modern society with relatively limited physical
aggression. Despite geographic constraints, high unemployment rates, focalized resource
competition, and a surging urban population, violent mortality rates in Greece are among
the lowest globally. Thus, contrary to socioeconomic stress theories suggested in
bioarchaeological and clinical literature, contemporary Greeks do not show increased
trauma from assault, abuse, and other violent crimes due to economic and populational
The ability of the anthropologist to recognize and interpret skeletal trauma provides
valuable information to investigators regarding circumstances surrounding death. Victim
identification is assisted through the analysis of individualizing traits observed on the
skeleton as well as the matching of trauma seen at a postmortem examination to
antemortem medical records. The anthropologist's knowledge of trauma mechanism
provides valuable insight into whether bony lesions are the result of violent injury or
Additionally, trauma analyses bring information on intra- and inter-group
differences in skeletal trauma. Demographic trends may illuminate which biological
profile is more apt to sustain injuries within a group, thus illustrating relationships among
sex, age, and trauma. Patterns can then be compared to geographically and temporally
diverse groups, to determine similarities or disparities in the cultural patterning of trauma.
Within anthropology, bioarchaeological research has especially highlighted the
importance of eco-geographical factors on trauma frequencies in ancient and historical
populations. Variables include the effect of local economy (Walker 1989), resource
competition (Milner et al. 1991; Standen & Arriaza 2000), and uneven geophysical
terrain (Alvrus 1999; Kilgore et al. 1997) on skeletal trauma.
Clinical research based on global hospital admittance records has shown that
socioeconomic variables such as income, education level, minority/ethnic status, and
substance abuse are significantly correlated with overall trauma frequencies (Harries
1997; Cubbin et al. 2000a; 2000b; Faelker et al. 2000; Lyons et al. 2000; Wagner et al.
2000; Hasselberg et al. 2001; Blakely et al. 2002). These same variables also influence
the frequency of abuse to women, children, and elders (Grisso et al. 1999; Kyriacou et al.
1999; Caetano et al. 2000; Cunradi et al. 2000).
Research based on the behavioral mechanism of trauma in contemporary skeletal
samples is notably scarce, due mainly to the limited number of available collections. As
such, studies focusing specifically on trauma are limited to fracture patterns of isolated
elements such as the forearm (Mensforth et al. 1987; Mensforth & Latimer 1989) or
cranium (Walker 1997). By focusing on specific anatomical elements or regions of the
skeleton, researchers assume that injuries found in certain regions are predictive of
overall accidental or aggressive behavior. Novak (1999) addressed this analytical bias in
a study on skeletal trauma manifestations of domestic assault versus accident in a modem
British casualty unit. Novak found the coexistence of certain injuries to be most
predictive of behavior. Tandem craniofacial, thoracic, and distal upper extremity injuries
were more often associated with domestic assault, while singular, isolated fractures to the
upper and lower distal extremities were more predictive of accidental episodes. Thus, if
preservation allows, an examination of the entire skeleton is critical when one attempts to
decipher the mechanism of trauma.
A cemetery collection (n=121) of well-preserved individuals of recent death from
Athens, Greece, presents a unique opportunity to test the influence of various
demographic and environmental variables on individual trauma. The skeletal remains of
these individuals are on average 95% complete, and thus an analytical approach to trauma
may be done using the entire skeleton. Associated death certificates include data on sex,
age at death, place of birth and cause of death. Also included are frequent notations on
last known occupation.
Greek burial customs dictate that interred bodies are exhumed after a period of 3
years and stored in mausoleums located on the cemetery grounds, unless one has made
monetary provisions for permanent burial. Those individuals or their families who could
not afford permanent burial sites make up the collection.
Using this data set, several research objectives were addressed. First tested was the
relationship between the relatively low socioeconomic status of the average Athenian and
possible increased trauma frequencies. Compared to more westernized cosmopolitan
cities, the population of Athens is composed of citizens of mostly low income and low
education. This skeletal collection in particular is composed of those less wealthy
citizens who could not afford permanent gravesites in the mausoleums of the Zografou
Cemetery in Athens. Low income precludes access to adequate diet and health care, the
evidence of which will possibly be seen in the remains as poorly treated fractures and the
osseous reaction to dietary distress.
Next, the relationship between the specific ecogeographical characteristics of
Athens was tested for possible increased trauma frequencies. Contemporary Athens has
undergone very recent urban modernization and population surging, with population
levels increasing almost 50% in the last 3 decades (Greece in Figures, National Statistical
Service of Greece 2003). Growth in the capital is physically constrained within naturally
limiting geographical boundaries, compounded by mountainous and uneven terrain.
Combined with the limited tourist-based local economy, Athenians are faced with
stringent focal resource competition in a confined, often treacherous space.
The age bias in the collection was then tested for possible increases in skeletal
trauma. The advanced age of many of the individuals in the collection should result in
increased cumulative antemortem injuries. Given the low socioeconomic status of the
individuals, these injuries should also be more frequent and noticeable due to a lifetime
of suboptimal health care.
The overall nature of the injury pattern was explored to decide whether it provides
population-specific evidence of accidental versus intentional injury. Those demographic
and environmental variables most associated with skeletal injury were examined, and, as
much as can be tested, the extent to which reported occupation results in trauma was also
reviewed. Lastly, the contemporary Greek trauma pattern was compared to those
described in the archaeological literature from Greece and neighboring Mediterranean
MECHANISMS OF SKELETAL TRAUMA
Before reviewing the bioarchaeological and contemporary research in skeletal
trauma, it is useful to briefly examine what specific injuries suggest a certain mechanism
for skeletal injury. Because of individual resistance to injury, health status, cultural
allowances, or barriers to trauma, and so on, there can be no cross-cultural list of injury
constellations suggesting specific trauma mechanism. Keeping these limitations in mind,
the following discussion explores the macroscopic qualities of regional skeletal trauma
and their possible mechanisms in both the subadult and adult skeleton.
The growing skeleton has specific biological and mechanical qualities that result in
fractures and patterns of healing unlike those seen in the adult skeleton. The more
biologically active subadult metaphyses and periostea, as well as the patent nature of the
growth plates, are some of the features precluding children to unique fractures and
Simple linear fractures to the cranial vault may result from direct blows or birth
trauma. While not usually fatal on their own, Papaefthymiou et al. (1996) report on the
increasing phenomenon of "growing" skull fractures in infants who were delivered with
the assistance of vacuum extraction. The cranial bones of the infant are thin and fragile,
and aggressive pulling during delivery with a vacuum apparatus may result in a fracture
that is soon infiltrated with blood and/or cerebrospinal fluid. Soon after delivery, the
infant's head begins to assume the more normal rounded shape, and the originally small
fracture spreads (or grows) as the vault changes shape.
Fractures to the clavicle are very common in children, with fractures to the shaft
making up 85% of all clavicular injuries (Webb & Mooney 2003). The most common
mechanism is a fall onto the shoulder. Clavicular fracture may also occur as a result of
birth trauma, when an infant's upper chest is either squeezed through the narrow birth
canal or the infant is assisted out of the birth canal by pulling on an arm.
Trauma to any of the large growth plates located at the longitudinal margins of the
long bones tends to be caused by accidental falls (Webb & Mooney 2003) or direct blows
(Neer 1985). Humeral shaft fractures in children less than 3 years of age are highly
correlated with child abuse, especially if the fractures are spiral (Webb & Mooney 2003).
Spiral fractures are caused by torque applied in a twisting motion to the limb as one
would see with children being grasped violently by the arm.
Elbow trauma is extremely common in children and accounts for up to 65% of all
fractures and dislocations in children (Green 2003). Specifically, the supracondylar area
of the distal humerus is commonly injured from falls on an outstretched hand with the
elbow hyperextended, or by direct falls on a flexed elbow.
Fractures to the forearm are common accidental injuries in children. The most
common cause is a fall in or around the home or in sports-related activities.
Approximately 80% of all forearm fractures occur in the distal third of the radius and/or
ulna, with the site of the fracture becoming more proximal with advancing skeletal age.
Just over 50% of these fractures are greenstick, or incomplete transverse fractures
(Armstrong et al. 2003).
Fractures to the femoral shaft usually result from high-energy impact sources such
as automobile accidents or falls from great heights, although child abuse may be a culprit.
Child abuse accounts for 67% of femoral shaft fractures in children less than 1 year of
age (Nork et al. 1998). Fractures to the distal femoral metaphysis are most commonly
caused by a direct blow to the anterior or lateral aspect of the thigh or by a fall from a
height (Zionts 2003). In children less than 1 year of age, child abuse should be
considered as children this young are not fully ambulatory. These same forces cause
injury to the adjacent physis.
Fractures to the tibia and fibula are mostly transverse or oblique injuries resulting
from falls, motor vehicle accidents, or sports such as skiing or soccer (Thompson &
Behrens 2003). Again, as in the femur or any major long bone, child abuse may be
suspected if the child is less than 1 year of age or nonambulatory. Children aged between
1 and 6 years commonly experience a toddler's fracture at the distal tibial shaft (Dunbar
et al. 1964). The injury presents as an oblique fracture line crossing the distal tibial
diaphysis that terminates medially. It usually results from innocuous activity such as
tripping while walking, or falling from a modest height.
Injuries to the foot and ankle are commonly caused by indirect forces usually
hypereversion or hyperinversion of the foot (Crawford & Al-Sayyad 2003). They may
also be caused via direct violence by falls from a height or by motor vehicle accidents.
There is much debate as to which bones or skeletal regions are the most frequently
injured in known cases of child abuse. Some find the middle and proximal diaphyses of
the long bones to be injured most often (Beals & Tufts 1983), while others cite the vault
(Kowel-Vern et al. 1992; Lodor & Bookout 1991; Skellem et al. 2000) or ribs (Akbamia
et al. 1974). Worlock et al. (1986) found different patterns for infants and toddlers:
abused infants exhibited more thoracic and vault trauma and toddlers presented with
more long-bone injuries.
Spiral fractures of the long bones may result from violent twisting of a limb or from
shaking an infant or child while suspending it by a limb (O'Neill et al. 1973).
Anteroposterior compression of the thorax (holding a child by the chest and squeezing)
often results in rib fractures (Kleinman et al. 1992). Acute axial loading of the head
against the neck (slamming a child on a floor) may produce fractures to the first ribs. The
relatively heavier head of the child acts as a compressive force, transmitting energy to the
first ribs via the neck muscles (Strouse & Owings 1995). Thomas et al. (1991) list spiral
fractures to the long bones, especially in children younger than 1 year, and all humeral
fractures other than supracondylar as the more common injuries resulting from abuse.
Walker et al. (1997) found that the presence of active or healed subperiosteal bone
formation, rib fractures without major chest trauma, metaphyseal fractures, and simple
linear cranial fractures were possible indicators of child abuse in forensic cases.
One commonly cited pathognomonic feature of child abuse is the presence of
multiple fractures in different stages of healing (Kocher & Kasser 2000; Walker et al.
1997). However, Shaw et al. (1997) found that spiral fractures of the humerus in children
under 3 years of age were more likely the result of accidental episodes than abuse, and
that neither age nor fracture pattern were diagnostic of abuse. Whether the fractures were
isolated or in association with other injuries also did not lead to suspicion of abuse.
Indeed, Lodor & Bookout (1991) found isolated acute fractures without signs of other
trauma to be the most frequent pattern in battered children.
While there is no set age for a child to be most vulnerable to abuse, research has
shown that it mostly occurs in younger children, usually less than 1 year of age (Beals &
Tufts 1983; Kowel-Vern et al. 1992).
An awareness of conditions producing changes in bone that mimic child abuse is
useful for differential diagnoses. As stated above, Walker and colleagues (1997) noted
subperiosteal lesions as possibly indicative of abuse, although they also concede that new
bone formation in the metaphysis is normal in infants of 6 to 8 months of age and could
be confused with trauma. Osteogenesis imperfecta, hemophilia, leukemia, and congenital
indifference to pain should always be considered in clinical differential diagnoses
(Walker et al. 1997).
Beginning with the craniofacial skeleton, focal depressed fractures limited to the
outer table only are termed pond fractures because of their shallow, rounded appearance
(Knight 1991). Pond fractures may be accidental in origin, although they are often
attributed to interpersonal aggression. Walker (1989) found numerous pond fractures and
other nonlethal craniofacial trauma in a sample of prehistoric Native Americans from
southern California. These cranial wounds were attributed to culturally mediated acts of
intentional violence (nonlethal face-to-face combat) resulting from resource competition
stress in the geographically circumscribed area.
In addition to being a common result of intentional blunt force aggression, facial
fractures are frequent injuries from vehicular accidents when the occupant is thrown face-
first into the dashboard or windshield (Gurdjian 1975). Maxillary and nasal fractures
may result from punches, vehicular accidents, falls, and so on. Essentially any direction
of force will fracture these delicate bones, although lateral blows tend to be most
common (Watson-Jones 1941). Lateral impact to the cheek will usually result in trauma
to the zygomatic bone. More posterior lateral blows will usually fracture the arch of the
zygomatic bone. Frontal impact may result in fractures to the projecting malar tubercle.
Fractures involving the lower orbital floors may be associated with complex LeFort
midfacial trauma or they may be an isolated injury called a blowout fracture. They are
the result of direct force over the orbit, where hydraulic changes in the globe of the eye
cause compression of the bony floor. Frontal impact to the lower orbit may also lead to
buckling of the floor. Blowout fractures result from impact with a fist or a fist-sized
object (Rogers 1992).
Vertebral injuries have a number of mechanisms. Injuries to the lower cervical
spine tend to occur indirectly as a result of a blow to cranium, rapid deceleration,
hyperflexion, hyperextension, axial loading, or extreme rotation of the cranium (Mirza &
Anderson 2003). Excessive axial loads to the thoracolumbar region may result in
compression, wedge, or burst fractures of vertebrae. Extreme lateral flexion, extension,
and rotation of the spine may cause fractures to the lateral masses, spinous processes,
and/or dislocation of adjacent vertebrae.
The clavicle is a frequently fractured bone in adults. Mechanisms for injury
include fall from a height, motor vehicle accidents, sports injury, and direct blows (Ring
& Jupiter 2003).
Humeral fractures are relatively rare in adults, accounting for only 4-5% of all
fractures (Green & Norris 2003). The most common mechanism for humeral and related
shoulder injury is a direct blow to the anterior, lateral, or posterolateral aspects of the
humerus. An axial load applied to the humerus through a flexed elbow may also result in
fracture. Much more common is rotator cuff tearing, although the evidence of such a soft
tissue injury in bioarchaeological settings is limited to possible myositis ossificans and/or
periosteal reaction of the humerus.
Fractures to the tubular bones of the hand are very common injuries in adults,
especially fractures at the necks of the metacarpals (Jupiter et al. 2003). They are
frequently the result of direct impact on the metacarpal heads with the hand in a clenched
fist (boxer's fracture). Axial loads to the radial half of the palm while the wrist is in
extension (falling on the palm of the hand) may fracture the scaphoid, distal radius and
radial head. Fractures to the distal radius are extremely common, accounting for almost
20% of all fractures seen in the emergency department (Cohen et al. 2003). Colles
(1814) first described this injury resulting from a fall on an outstretched hand, where
energy absorbed from the impact travels through the carpus into the distal radius. Parry
fractures of the medial-distal ulna result from direct blows to the forearm as it is raised to
ward off a blow to the head.
Femoral neck fractures occur mostly in individuals older than 50 years
(Swiontkowski 2003). While such fractures may be caused by falls from significant
heights or vehicular trauma (especially in younger individuals), low-energy falls from a
standing position account for 90% of fractured femoral necks in the older cohort. Age
related osteoporosis, balance problems from decreased strength and agility, and
neuromuscular disease make older individuals prone to falls and subsequent fractures of
the femoral neck and intertrochanteric region. Fractures to the femoral diaphysis and
distal femur result mainly from falls from heights, motor vehicle accidents, and sporting
injuries (Court-Brown 2003).
Injuries to the tibial plateau occur mostly as a result of direct force to the proximal
tibia as in a car bumper fracture (Watson & Schatzker 2003). Transverse tibial
diaphyseal fractures may result from a direct blow to the shins, especially if the adjacent
fibula remains intact. Spiral and transverse fractures to the tibia also commonly result
from sporting accidents, especially skiing. Transverse boot top fractures occur when the
top of the ski boot acts as a fulcrum over which the tibia is broken. Spiral fractures occur
when the foot is stable and the body is twisted over the foot (Trafton 2003). Stress
fractures of the tibia result from repeated loading, with ultimate failure from fatigue.
Tibial stress fractures are mostly seen in the proximal physis of individuals (such as
soldiers, dancers, and runners) who place significant demands on their lower extremities.
Fractures to the distal end of the tibia (pilon) are most commonly caused by a fall from a
height, or by a motor vehicle accident (Bartlett et al. 1997). Forced abduction of the
pronated foot is responsible for many injuries to the medial and lateral malleoli. Usually
referred to as "twisted ankles", these injuries result from tension created by the pull of the
lateral or medial ligaments when the ankle is sharply inverted or everted, as when one
falls off a curb. Falls from heights or motor vehicle accidents account for most foot
injuries (DiGiovanni et al. 2003).
In general, craniofacial trauma tends to be more indicative of interpersonal
aggression. Victim identity is focused in the face, and thus aggression to the individual is
often directed there (Galloway 1999). In addition to the face, Fonseka (1974) found that
the thorax and ventral surfaces were areas most injured during episodes of spousal abuse.
Appendicular trauma tends to be more suggestive of accidental situations. Fractures to
the wrist and ankle commonly result from falls, especially in uneven terrain (Wells
The distinction between accidental and intentional injuries is blurred; attempts to
define the behavioral mechanism behind injuries are difficult, and should be pursued with
caution. The distinction between accidental- and violent-based skeletal trauma is highly
dependent on cultural influences. Attempts to determine the mechanism of trauma in
skeletal remains must be made in a contextual manner, where the examination of the
entire skeleton, when possible, can lead to more accurate determination of the manner in
which injuries were sustained.
Although many studies have addressed the physical manifestation of aggressive
versus accidental trauma, just as many have failed to adequately propose a set pattern for
skeletal expression distinguishing between the two. While due partly to differing
documentation protocols, what seems to be missing from the mass of literature is the
notion that aggression is dictated mainly by cultural norms. Every society will have its
own mode of dealing with stressors, whether they are of environmental or behavioral
origin. The physical display of intentional violence will thus show much intercultural
RELEVENT BIOARCHAEOLOGICAL LITERATURE
Archaeologically, evidence of traumatic injuries are seen as bony changes
associated with fractures, callus development, remodeling after joint dislocations, and
ossifications that occur within injured muscle, tendon, and periosteum. From a
behavioral perspective, it is important to distinguish among injuries suffered before death
(antemortem) and those around the time of death (perimortem). For an injury to be
considered antemortem, evidence of healing must be present. In antemortem skeletal
injuries, fracture edges are either rounded, have woven callus formation, or are
remodeled, depending on how long the individual lived with the injury. Perimortem
injuries are distinguished from postmortem damage by the lack of healing activity, by
differential coloration of fractured ends, and by properties of the fracture. The diagnostic
features of fractures produced by blunt, sharp, or projectile forces and the principles
guiding the interpretations of proximate cause are well understood by forensic
anthropologists. The ultimate cause is much more difficult and requires consideration of
both intrinsic biological variables such as age and sex, and extrinsic factors relating to the
physical and sociocultural context (Walker 2001).
Archaeological research has highlighted the importance of sociocultural and
ecogeographic factors in trauma occurrence. Using a prehistoric Indian sample from the
Channel Island area of southern California, Walker (1989) demonstrated the influence of
local economy on skeletal injury. Numerous cranial wounds found were attributed to
intentional violence resulting from resource competition stress in the geographically
circumscribed area. Similar resource competition and resultant warfare are also cited as
explanations for like trauma found in several other locations, including the late
prehistoric sites in Illinois (Milner et al. 1991) and Tennessee (Smith 2003), late
Woodland Michigan (Wilkinson & Van Wagenen 1993), precontact central California
(Jurmain & Bellifemine 1997), and the preceramic Chinchorro population from northern
Chile (Standen & Arriaza 2000). In the late Woodland Michigan site, in particular,
numerous females were found with cranial fractures. These injuries are suggested to
originate from both spousal abuse and violence associated with female capture during
Other significant bioarchaeological literature has shown the absence of trauma
stemming from aggressive behavior. Hershkovitz and colleagues (1995) found a very
low percentage of violent trauma in a collection of exclusively male skeletons associated
with a Byzantine monastery in Judean Desert. The authors attribute the lack of cranial
trauma to isolation from the secular population, regular food supply and absence of
warfare. In a late Woodland foraging group from Ohio, Lovejoy & Heiple (1981) found
evidence of mostly accidental traumatic episodes, with the highest frequencies found in
young males and females. The lack of sex differential in injuries suggests that warfare
was not the cause. Additionally, Smith (1996) found no sex bias in tandem parry
fractures to the ulna and craniofacial injuries in a large sample from late archaic
Tennessee. If a significant number of females had been found with such injuries, it might
suggest sex-specific abuse where blows to the head and face were deflected with raised
Geographic factors have been associated with elevated skeletal injuries and
reported in bioarchaeological studies. The uneven physical terrain of Nubia is suggested
as the particular cause of elevated accidental appendicular fracture frequencies found at
the neighboring sites of Semna South (Alvrus 1999) and Kulubnarti (Kilgore et al. 1997).
Both sites are in the Batn el Hajar region just south of the Egyptian border, an area
characterized by a dry and boulder-strewn landscape.
Skeletal Trauma Studies Using Greek Collections
Mediterranean archaeological sites, and Greek sites in particular, are infamous for
poor skeletal preservation. Thin-to-absent topsoil and wildly fluctuating temperatures
both result in speedy degradation of delicate bones. However, a few collections from
Greece and other neighboring Mediterranean sites have survived these environmental
challenges. J. Lawrence Angel, the late Smithsonian Institution physical anthropologist,
published numerous site reports on Mediterranean skeletal material from the early 1940s
through the mid 1970s. His reports are of varying length and detail, and customary of the
time, are found in the appendices of archaeological site reports. Although Angel was
primarily interested in tracing Greek social biology through cranial morphometrics
(Angel 1944, 1946), he also attempted to document traumatic and pathological lesions
found on archaeological remains.
In a report on the skeletal remains from Franchthi Cave, Angel (1969) described
trauma he observed on 10 individuals from this early Mesolithic to late Neolithic site in
southern Greece. One adult male in the collection had a "vertical scar down his right jaw
ramus" and a young adult female had "injured knuckles of [the] left hand" (Angel
1969:380). Another young adult male exhibited healed fractures of the left 1st
carpometacarpal joint and 2 healed depressed fractures of the frontal bone above the
browridge. No illustrations were provided of these injuries, nor does Angel speculate on
the behavioral mechanism associated with the trauma.
Angel's (1971a) thorough work at Catal Huyuk in Turkey provides a glimpse into
the early farming/proto-city lifestyle of people during the early Mediterranean Neolithic.
Social differentiation of trauma is evident in the frequency of skull wounds found in the
Catal Huyuk sample. Six of 22 adult male crania exhibit vault injuries, while only 2 of
32 females express similar trauma. Unfortunately, no description of the head injuries is
presented, thus limiting any speculation on the possible cause of the trauma. The fact
that 27% of males exhibit injuries compared to only 6% of females provides some
possible evidence of (intramale) violence. Fractures of the clavicle, humerus, ulna,
radius, and femur are also found on various individuals and are suggestive of accidental
trauma. Angel provides interesting scenarios for these injuries, ranging from bull goring
incidents to falls from house ladders in the dark (Angel 1971a). There is no sex bias in
any of these postcranial injuries.
The most noted skeletal pathology of this sample is porotic hyperostosis, which
results from anemia severe enough to cause hypertrophy of the vault diploe. Of the 143
adults, 41% display this characteristic thickening. The location of Catal Huyuk on the
inland drainage of the Konya Plain meant certain and close contact with anopholine
mosquitoes, which are transmitters of thalassemia, a Mediterranean version of malaria.
In addition to anemia, stresses from hunting, warfare, and trading trips would perhaps
have introduced sufficient life stresses to result in interpersonal violence.
Postural indicators are found on numerous individuals in this sample and provide
some indication as to how these people interacted with their environment. Reactive areas
on the femoral neck (Allen's and Poirier's fossae) are frequent. These lesions develop
when the ilio-femoral ligament is pressed against the zona orbicularis during running or
descending a slope. Such lesions are found in 70% of adult males and 84% of adult
females. Their frequent presence, in addition to the number of distal appendicular
fractures, suggests interaction with rough terrain. Squatting facets (depressions on the
distal tibia and/or superior talus) are found on approximately 50% of adult tibiae and tali
and suggest marked ankle flexion associated with a squatting posture and/or frequent
Additionally, numerous individuals in the collection (46% of 43 femora) have a
backward direction of the lesser trochanter. The iliopsoas tendon inserts at this area and
serves as a lever to tighten flexion and rotation of the hip. This posterior twisting of the
lesser trochanter may be another reaction to flexion and stability of the thigh necessary
for frequent climbing; although Angel suggests that the twisting developed to supply
added leverage for quick turning and poising needed in dancing or animal games, as
shown in frescoes of this time period (Angel 1971a).
Angel (1971b) found similar postural indicators in individuals from Lerna, an early
Neolithic to Roman era site on the Argos Plain of the Peloponnese peninsula. Evidence
of strong hip muscle development and frequent squatting facets indicate that climbing or
descending steep terrain was a common occurrence. Evidence of possible intentional
trauma is seen in fractures to the ulna, 5th metacarpal (boxer's fracture), nasals, scapula,
and vault. The overall frequency of these injuries is low (10% of the sample) and the
distribution of injuries is equally spread between males and females. The collection is
quite fragmentary, however, and true frequencies may be higher.
The Middle Bronze Age people of Lerna represent the start of a proto-urban
tradition. The era is a cultural spinoff from the transition from hunting, gathering, and
settled farming to urban trade and unification, and is presented by Angel as a time of
relative prosperity and cultural infusion. The general health of the individuals from this
time period is somewhat better than it is for those from earlier occupations. Porotic
hyperostosis from thalassemic malaria is seen in 26% of subadults and 16% of adults
(Angel 1971b). Additionally, signs of seasonal malnutrition or severe childhood disease
occurring as growth arrest lines on tooth enamel are minimal.
Earlier, Angel (1953) examined the skeletal remains of 45 subadults and 78 adults
from Khirokitia, a Neolithic farming village site on the Mediterranean island of Cyprus.
Craniofacial trauma is found on 3 of 39 adult male crania. Injuries are seen especially on
the zygomatic, frontal, parietal, and occipital bones. Angel does not provide detailed
descriptions of the injuries, nor are photographs provided. He suggests that the
craniofacial trauma is of sharp force origin, although the injuries could represent healed
linear blunt force fractures. All cranial trauma is found exclusively on males, suggesting
that the injuries are of an intentionally violent origin. Certain metric and genetic
observations of this collection, such as short-headedness, paedomorphism, and metopism
are presented as strong evidence that the Khirokitians were an inbred and parochial
population that had few contacts with other breeding groups. Indeed, the high infant
death rate, short life span, and short adult stature in the group may be evidence that the
Khirokitian villagers had little leisure time and were more concerned with trying to stay
alive (Angel 1953).
Angel's long-term work at the Turkish necropoli at Karatas (1966, 1968, 1970,
1976) has provided valuable information regarding life during the early to middle Bronze
Age on the Anatolian plateau. It is a large sample (n=584) important for its information
on the health and population changes made during the development of a proto-urban
economy out of an early farming subsistence. Cranial vault injuries are found on 5 males
and 1 female (over 10% of all intact crania) and suggest warfare or some other form of
intentional aggression (Angel 1970). Injuries are of blunt force and sharp force
mechanisms (depressed fractures and axe wounds, respectively). Fractures to the
midshaft of the ulnae are seen on 4 of 50 male ulnae and 1 of 30 female ulnae. Femoral
shaft fractures occurred in 1 of 122 male femora, and 2 of 160 female femora. Schmorl's
herniations and fatigue fractures of the 5th lumbar vertebrae are cited as common,
although no percentage frequencies are provided (Angel 1970). Minor postcranial trauma
found on the femora, tibiae, clavicles, humeri and radii occur equally between the sexes.
In a series of 40 crania from 5 different cemeteries dating to the 12th century B.C.,
from the Mediterranean island of Cephallenia, Angel (1943) found that 32-47% of each
subsample exhibits cranial injuries. All injuries were found in males, with the exception
of 1 female bearing a depressed fracture to the right parietal. Trauma resulting from
sharp force injury was found on 5 males, with 4 of them occurring on the left side. One
other male had a fractured left zygomatic bone of blunt force mechanism. The
preponderance of sharp force trauma in males, in addition to the frequency of injuries on
the left side of the skull, demands an ethnological explanation. Angel notes that the
decline of the Mycenaean civilization that occurred during the 12th century B.C. may
have resulted in physical encounters with invaders from the west. Additionally,
participation in the Trojan War would explain the frequency of violent trauma found at
this site (Angel 1943).
Other reports exist describing Greek skeletal trauma found on remains from Lema
(Wesolowsky 1973), Nichoria (Wade 1983; Bisel 1992), Attica (Angel 1945), Corinth
(Angel & Bums 1973), Diros (Papathanasiou et al. 2000) and the Athenian Agora (Little
& Papadopoulos 1998). However, these reports focus on either single, unique individuals
or on remains far too fragmentary or commingled to provide adequate comparative data.
Poor preservation aside, a few observations may be made after reviewing Angel's
anthropological reports. The relatively high frequency of fractures to the distal
extremities in the majority of samples is suggestive of consistent interaction with the
rough, sloping and mountainous terrain characteristic of Greece. Indeed, the prevalence
of postural indicators (i.e., squatting facets, Allen's fossae) and the degree of muscular
development among the individuals supports this assumption. There is also a relatively
high frequency of cranial trauma suggesting elevated cultural stressors such as
interpersonal (intramale) aggression, internecine warfare, and defense of the community
against warring invaders.
Table 3-1 compiles Angel's more notable skeletal analyses of Greek and Turkish
sites discussed above. When Angel gave information on whether injuries were found in
males or females, it is indicated. It should be remembered that skeletal trauma was not
the focus of Angel's work. Thus, blank cells do not necessarily mean that injuries were
absent. In many of his reports, postcranial trauma was not discussed at all. Actual
cranial and postcranial trauma occurrences are likely higher than reported, as the
archaeological material Angel analyzed tended to be quite fragmentary.
Table 3-1. Summary of skeletal analyses of Greek and Turkish archaeological sites published by Angel
Site Francthi Cave Catal Huyuk Lerna Khirokitia Karatas Cephallenia
Total sample size
Usable sample size
Early Mesolithic- Late
6 m, 2 f
5 m, 1 f
12th cent. B.C.
8 m, 1 f
1 m, 1 f
1 m, 1 f
RELEVANT CONTEMPORARY LITERATURE
Contemporary clinical research has shown that socioeconomic variables such as
income, education level, minority/ethnic status and substance abuse are all significantly
correlated with overall trauma frequencies in hospital patients (Harries 1997; Cubbin et
al. 2000a; 2000b; Faelker et al. 2000; Lyons et al. 2000; Wagner et al. 2000; Hasselberg
et al. 2001; Blakely et al. 2002). These same variables also influence the frequency of
abuse to women, children and elders (Breiting et al. 1989; Grisso et al. 1999; Kyriacou et
al. 1999; Caetano et al. 2000; Cunradi et al. 2000).
Figure 4-1 graphically displays the general relationship between socioeconomic
status and trauma frequency among hospital patients. Deprived individuals are
collectively those with low income, low education, minority status, history of substance
and alcohol abuse, incarceration and high unemployment rate. Conversely, affluent
individuals are those with relatively higher income, higher education, majority status, less
substance and alcohol abuse, and higher employment rates. As one's socioeconomic
status improves, or becomes more affluent, one's overall frequency of trauma tends to
Social researchers have also found socioeconomic status to be an important factor
in the type of skeletal injury encountered. Lyons et al. (2000) addressed the influence of
annual family income on childhood trauma in their study of injuries in affluent and
deprived areas of Wales. When matched for sex and age, children from wealthier areas
had higher rates of sports-related fractures while those in poorer areas had more assault-
related injuries. In addition, poorer children had considerably higher rates of death from
injury than their more affluent counterparts. Moustaki et al. (2001) noted a lower overall
injury rate in Greek children as compared to this same Welsh sample. They recorded a
rate one-third the frequency found by Lyons et al. (2000), citing the better nutrition of
Greek children as the biological mechanism for this decrease; presumably the better diet
of Greeks led to stronger bones and less fractures. Greek children did, however, have
twice as many cranial fractures as the Welsh sample. Variation in preventive measures
(lack of safety helmet use) and geographic qualities (rocky terrain) are given as reasons
for this difference.
Figure 4-1. General relationship between socioeconomic status and trauma frequency
among hospital patients
An enormous amount of clinical literature on skeletal trauma exists in relevant
medical journals. Most reports are based on hospital admittance records. The majority of
reports, however, are case studies focusing on specific regions or elements of the body,
such as the craniofacial skeleton (Schultz 1967; Luce et al. 1979; Voss 1982; Brook &
Wood 1983; Scherer et al. 1989; Hussain et al. 1994), upper limb (Alffram & Bauer
1962), wrist (Fleege et al. 1991), and foot (Wenig 1990; Koch & Rahimi 1991).
Few clinicians have attempted a systematic approach to skeletal trauma, and those
that publish such data lack a sociocultural approach in their analysis. Buhr & Cook
(1959) describe fracture patterns in a sample composed of British hospital admission
records for the years 1938 through 1955. Fractures (no other skeletal trauma were
investigated) were considered only by age and sex of the injured. The authors found that,
below age 50, fractures occurred mostly in males, but above that age there was a sharp
increase of trauma to women. By the decade 70-79 years, there were four times as many
women with fractures as men. In a similar study of British hospital records, Donaldson
and colleagues (1990) reported on age and sex specific fracture rates in a 3-year patient
cohort in the early 1980s. The authors found an identical elderly female dominance in
fracture frequencies with age-related disease cited as the mechanism behind most
injuries. Sahlin (1990) analyzed the incidence of fractures according to age, curiously
omitting sex as a variable. Using admittance records of a Norwegian hospital from 1985-
1986, he found advanced age to be the most important variable in fracture occurrence.
Harries (1997) provides an interesting report focusing on the influence of
sociocultural stress and physical trauma. Using an epidemiological approach, the author
investigated spatially distributed phenomena (violence) and the conditions (social
stresses) associated with them. A stressorr" is defined as a condition producing some
degree of social dysfunction. The stress of poverty was most influential and was
accompanied by a related set of deprivations in housing and other basic necessities.
Using a number of variable stressors (including the number of homicide/aggravated
assault incidents, large households with no male figure, unemployment, poverty levels,
vacant housing units nearby, adult per capital income and education level), Harries found
clusters of specific areas with shared traits in terms of the underlying set of stressors.
Harries did not focus on specific, physical examples of trauma. Rather, he used a public
health approach to investigate why certain areas in an urban setting were exhibiting high
frequencies of trauma.
Research based on contemporary skeletal sample populations is scarce, due mainly
to the limited number of available collections. At present, there are two large and
relatively well-documented accessible collections, both of which are composed mostly of
American individuals who either willed their bodies to science or who were unclaimed
hospital deaths. The Hamann-Todd Osteological Collection at the Cleveland Museum of
Natural History consists of the skeletal remains of anatomical specimens who were
originally unclaimed bodies retrieved from the nearby county morgue and city hospitals
from 1893-1938 (Jones-Kern & Latimer 1996). The Robert J. Terry Anatomical Skeletal
Collection curated at the National Museum of Natural History at the Smithsonian
Institution consists mainly of cadaver skeletons retrieved from the Washington University
Medical School. These bodies were primarily obtained from hospital and institutional
morgues in the St. Louis, Missouri area from 1920-1967 (Hunt DR, August 17, 2004,
www.nmnh.si.edu/anthro/cm/terry.htm). The collection also contains some individuals
who donated their bodies to science.
Although much skeletal research has been done using these two samples, studies
focusing specifically on trauma are curiously limited. Mensforth et al. (1987) and
Mensforth & Latimer (1989) observed fracture patterns of specific elements of
individuals in the Hamann-Todd collection and compared them to the same injuries in
individuals of slightly more contemporary origins. Angel (1974) included a sub-sample
from the Terry Collection who had willed their bodies to science in his report on the
comparative fracture patterns of individuals from the Neolithic to modern times. Angel
specifically ignored the dissecting-room component of the collection, believing that
"disadvantaged people exposed to tougher genetic, nutritional, and socioeconomic
forces... [were] not comparable with ancient populations" (Angel 1974:9). Angel
assumed the health of the contemporary industrial city-dweller was worse than that of the
archaeological sample, although a large suite of environmental stressors no doubt
affected the ancient groups as well. Lastly, Walker (1997) included a sample of crania
from both the Terry and Hamann-Todd collections as part of a larger study examining the
skeletal evidence for the cultural patterning of violence in diverse Western populations.
Most of these studies focused on specific anatomical elements or regions of the
skeleton. The researchers thus assume that trauma found at various sites is predictive of
overall accidental or aggressive behavior. Walker (1997) assumes that certain cranial
injuries are more predictive of intentionally violent situations than other regions and he
omits the postcranial skeleton in his study. Although Mensforth et al. (1987) indicate
they were only looking for fracture patterns in certain bones, the inclusion of the entire
skeleton would have provided a more accurate picture of the behavior associated with
trauma. Indeed, in a study on the skeletal trauma manifestations of patients involved in
domestic assault versus accidents in a modern British casualty unit, Novak (1999) found
the coexistence of certain injuries to be most predictive of behavior. Tandem
craniofacial, thoracic and distal upper extremity injuries were more often associated with
domestic assault while isolated fractures to the upper and lower distal extremities were
more predictive of accidental episodes. Thus, if preservation allows, an examination of
the entire skeleton is necessary when one attempts to decipher the ultimate mechanism of
trauma. Selecting specific elements may ease analysis, but it presents an artificial
isolation of elements, and therefore introduces bias.
SOCIOECOLOGICAL ASPECTS OF PRIMATE BEHAVIOR AND TRAUMA
Research on primate behavior is of interest to anthropologists because, to the extent
that non-human animals engage in planning, cooperation, and manipulation of
individuals, relationships, and alliances, their behavior must involve features which are to
some degree similar to those which characterize human behavior (Quaitt & Reynolds
1993). Anthropologists look to our primate cousins to discover characteristics that are
shared among the humans and non-human primate evolutionary continuum. Primate
studies are also pursued to discover characteristics that are distinctly human, as opposed
to those that might be part of the primate heritage.
Of special interest to primate socioecological research is the focus on within- and
between-group competition. Competition can be defined simply as when species
simultaneously seek essential resources of an environment that are in limited supply, be
they food, social/sexual partners or safe places to live or hide (Sussman 1999).
Competition depends on the availability and distribution of resource items, the
monopolization of which plays a part in structuring the nature of social relationships
between members of a primate social group (Van Schaik 1989). In order to acquire
resources, it pays for an individual to invest in behavioral dispositions that increase his or
her superiority. Consequently, there will be a selective pressure toward individual
dispositions that are dominance-oriented.
Population Density and Resource Availability
Ecological theories on aggressive behavior hold resource scarcity to be the crucial
determinant in human societies (Harris 1974, 1979; Ross 1985, 1986). Likewise, in cases
where preferred or necessary food items are naturally restricted to a few places, conflict
and high levels of intragroup aggression are also observed in non-human primate groups
(Nagel & Kummer 1974). Borries and colleagues (1991) found that in saturated habitats
of Hanuman langurs, rates of within-group aggression are more elevated than they are in
groups living at lower population densities with more abundant food supplies. Here, one
would expect environmental circumstances to affect the frequency of aggressive acts.
Diminished resources increase the importance of winning; it seems only logical,
therefore, to predict an increase in antagonistic behaviors under such circumstances.
Food shortages should result in increased competition and status hierarchies representing
the order of access to food.
Primate research on population density and aggressive behavior among various
species also reports conflicting results, however. In a study of chimpanzees living at the
Yerkes Regional Primate Research Center, Aureli & deWaal (1997) found that the rate of
agonistic behaviors actually occurred less frequently under high-density conditions. This
was interpreted as an inhibition strategy to reduce opportunities for conflict when
interindividual distances were reduced. This strategy was effective only in the short run,
however, as anxiety levels were simultaneously elevated, suggesting increased social
tension under high population density conditions. When free-ranging rhesus monkeys in
Cayo Santiago were faced with a temporary food shortage, the frequency of fights
decreased significantly while the frequency of grooming, play and mating also decreased.
The animals became socially lethargic and spent most of the time slowly searching for
food (Loy 1970). Southwick (1976) noted that food shortages also resulted in decreased
agonistic encounters among rhesus monkeys in Calcutta. Fighting, grooming, playing
and mating behavior simultaneously declined during food shortages, suggesting tension
was present even if outwardly agonistic behavior was not.
Additional research suggests that primates use various affiliative mechanisms (i.e.,
hand-holding, altruistic behavior) to adjust their behavior when population density
increases such that potentially adverse consequences of crowding and food shortages are
present (Bercovitch & Lebron 1991). Social stability then, tends to be a more important
determinant of primate aggression than population density.
Studies on human aggression suggest that cultural features may likewise override
ecological factors involved in agonistic behavior. Robarchek & Robarchek (1992)
performed a comparative study on two equatorial societies in an effort to determine what
factors were involved in aggressive behavior in two eco-geographically similar societies.
The warlike Waorani of the Ecuadorian Amazon live in a low population density region
with plentiful resources. There are no specific cultural or individual values to maintain
group cohesion and thus no internalized controls on conflict or violence. The Semai
Senoi of the Malaysian rainforest experience higher population density with scarcer
resources. Violent behavior is heavily constrained, however, by individual and cultural
values that stress nonviolence, and by the internalized need to avoid any disruption within
the kindred and band. Thus, differing cultural constructions between the two groups
structured their behavior.
Darwin's sexual selection hypothesis is believed to play an important role in the
systematic or opportunistic killing of infants (Darwin 1871). Sexual selection refers to
the struggle between one sex for access to the other with the result for the unsuccessful
competitor being not death, but fewer or no offspring. A male increases his reproductive
success by killing unrelated infants if the infant's death makes the mother resume estrus
sooner than she would otherwise. Unweaned infants are at the highest risk (Hrdy et al.
In addition to sexual selection, infanticide may occur as a response to competition
for territory and food resulting from environmental pressures (Hrdy & Hausfater 1984;
Hrdy et al. 1995). Roda & Pontes (1998) documented targeted killing of unweaned
infants in a group of common marmosets as a consequence of environmental disturbance.
The population had surpassed the limited carrying capacity of the environment, which
increased competition for food, mates, and territory. Troop members had no opportunity
to set up territories elsewhere and so were forced to compete for the scarce resources in
their natal group.
Infanticide may also act to reduce the number of future competitors in a group.
Newton-Fisher (1999) observed two cases of male infanticide by extra-troop males in
wild chimpanzees. By reducing the number of males reaching adulthood in a
neighboring community, the aggressors reduce its territorial strength, which in turn
makes range expansion, recruitment of females, and extinction of the neighboring
communities more feasible for the infanticidal males. Arcadi & Wrangham (1999) found
male and female cooperation in infanticide. Females appear to gain long-term benefits by
reducing resource competition or the risk of their own infants being attacked. Males
benefit by spurring females into estrus sooner and eliminating future rivals. However,
the fact that some of the males were related to their victims and that some did not always
mate with the mother highlights how difficult it is to attribute infanticide to any one
Field studies also show that increased general aggression toward females with
infants occurs during boundary patrols (Watts & Mitani 2000). Chimpanzees are often
seen cannibalizing the bodies of killed infants (Goodall 1986) suggesting that infanticide
may sometimes occur as a food resource exploitation.
Intercommunity relations among certain primate species are predictably hostile.
Muller (2002) found that coalitionary intergroup attacks are a regular feature of
Kanyawara chimpanzee society. Male activities during border patrols for territorial
defense are especially violent. Female chimpanzees are found to be aggressive primarily
in the context of feeding competition, although levels of interpersonal aggression are
rarely severe. One possible explanation is that competition for space is not as
pronounced as it is at other regions, thus the benefits of female competition for high rank
are less. Coalitionary attacks act to reduce the coalitionary strength of neighbors as well
as expand territories.
Other references cite tolerant and affiliative aspects of primate relationships,
especially in all-male groups. Examples include spider monkeys (Chapman et al. 1989),
squirrel monkeys (Boinski 1994; Mitchell 1994) and chimpanzees (De Waal 1982).
Primatologists have found that cooperation, bonding reciprocity and negotiation among
primate groups seems to depend partly on the degree of genetic relatedness (Van Hooff
Targeted aggression results, then, from an amalgamation of adaptive, ecological,
social, and physiological mechanisms. Environmental fluctuations influence the
regularity of food patches and in turn affect the social and physical well being of primate
troops (Roda & Pontes 1998; Saito et al. 1998). The introduction of stepparents or
sudden group movement may spur stress among troop members resulting in "social
pathology" (Snowdon & Pickhard 1999). Aggressive behavior may result from attempts
at population control (Digby 1999) or predation (Goodall 1986). Researchers have found
that aggressive behavior differs among species (Boinski 1999) and within species (Saito
et al. 1998), the latter suggesting that personality differences of troop members may
reflect an individual's own experiences more so than their genetic background (Clarke &
Severe wounding or death from infanticide and generalized aggression are frequent
occurrences in most primate species. Various intraspecific social, physical, and
ecological processes influence injury caused by aggressive and accidental behavior.
Traumatic, healed fractures in primates occur in significant frequencies and may
reflect patterns of primate aggressive behavior as well as physical adaptations to
environmental settings and locomotor differences. There may be different intraspecific
"cultural" norms existing that regulate aggressive behavior. Individual personality
differences and the influences of life histories may also direct intraspecific injury rates.
Species may also be prone to sustain injuries because of the dangers of different
environmental settings (terrain, climate, predators). Such wide variances in fracture
frequencies may also be the result of differing documentation protocols with varying
ideas of what constitutes trauma. Most likely, all these factors come into play.
MATERIALS AND METHODS
The following discussion describes in more detail the Greek sample used to address
the research objectives introduced earlier. Documentation protocols and statistics used in
this study are also provided.
The University of Athens Sample
The University of Athens (UA) skeletal sample is composed entirely of individuals
from an ossuary located in the eastern Athenian suburb of Zografou. They are the result
of recent exhumations within the last 5 years from a cemetery on the same grounds. The
remains have been donated to the University of Athens, Department of Human and
Animal Physiology for research purposes.
Greek culture follows pseudo-secondary burial practices. Unembalmed decedents
are initially buried for a period of three years to allow for complete skeletonization,
exhumed, and then finally stored (or re-buried) in individual containers in several large
repositories located on the cemetery grounds. Those individuals who had not made
provisions for the storage of their remains after exhumation are, after a period of time,
eventually pulled out of the repositories and disposed of in large communal underground
vaults (Appendix A). All the individuals in this collection have either been donated to
the University of Athens by family members of the deceased or are individuals who were
on their way to the communal bone vault (C. Eliopoulos, pers. comm.). Full and friendly
cooperation exists between the ossuary caretakers and the University of Athens.
Most skeletons were relatively free of soft tissue and odor on exhumation. Those
individuals who required further cleaning were immersed in a weak solution of warm
sodium perborate trihydrate for a period of 2-3 days in order to extract additional fat from
the bone. After cleaning, the bones were air-dried under a protective hood, labeled, and
then stored in archival boxes in the biological laboratory at the University of Athens.
Currently, 171 individuals are in the Greek collection, with additional remains
received periodically from the Zografou ossuary. The age at death, place of birth and
death, occupation, and cause of death are mostly known. A few caveats must be
presented explaining the quality of the data. While sex, age and places of birth and death
are established forthright, other variables are somewhat questionable. Cause of death as
listed on the death certificate is described interchangeably as either the primary or
secondary cause. Thus, an individual may exhibit obvious evidence of a long-term
systemic disease such as cancer, but the death certificate will cite the cause of death as
'acute respiratory failure'. Additionally, the last known occupation provided by the death
certificate is just that. In what aspect the individual was or was not employed preceding
the last known occupation is not known. Many of the individuals in this collection are of
advanced aged and are listed as pensioners, rendering the occupation variable unusable
for those individuals.
Individuals from the sample were examined for gross antemortem and perimortem
skeletal injuries. For an injury to be considered antemortem, evidence of healing must be
present. In antemortem skeletal injuries, fracture edges are either rounded, have woven
callus formation or are remodeled, depending on how long the individual lived with the
injury. Perimortem injuries are distinguished from postmortem damage by the lack of
healing activity, differential coloration of fractured ends and properties of the fracture.
Both sexes and all ages are included in this study. In addition to fractures, skeletal
evidence of soft tissue trauma (i.e., muscle and ligament injuries) was recorded. No
destructive procedures were undertaken.
The first data collection on a subsample (n=31) of the cemetery collection at the
University of Athens during the month of October 2001 served as a pilot study. For this
trip as well as the subsequent trip in September-October 2002, a thorough and efficient
recording protocol was developed. Protocols suggested by Lovell (1997) and Buikstra &
Ubelaker (1994) provided the bases for the documentation procedure. Data recording
included separate forms for inventory, morphological and metric assessments of sex, age
and pathology. The inventory forms recorded data on individual element preservation by
region and side. The pathology forms included detailed information on fracture type,
shape and sequelae. The length of the afflicted element, as well as the apposition,
rotation, and angulation at the fracture site were recorded. Observations were also made
for evidence of dislocations, amputations, arthritis (as possible evidence for joint injury)
and skeletal evidence of antemortem soft tissue trauma. Sample data collection forms are
located in Appendix B.
Although data on the sex, age, and ancestry for the Greek individuals were
already known from associated death certificates, confirmation of the biological profile
data was accomplished via morphological and metric assessments of diagnostic skeletal
features, the most accurate of which are found on the pelvis and cranium. For the pelvis,
Phenice's (1969) technique for adult sex determination on the pubic region was used.
This method includes examination of the ventral arc, subpubic concavity and ischiopubic
ramus. Additionally, the morphology of the greater sciatic notch and preauricular sulcus
were considered. On the cranium, development of the nuchal crest, mastoid processes
and supraorbital margins were further diagnostic indicators for sex.
Subadult sex determination is notoriously difficult due to the lack of sexually
dimorphic skeletal features, which are under hormonal control and thus do not appear
until puberty. Although features of the auricular surface (Weaver 1980; Mittler &
Sheridan 1992) and greater sciatic notch (Boucher 1957; Weaver 1980; Schutkowski
1993) are suggested as dimorphic in subadults, none of these methods is highly accurate.
Thus, morphological confirmation of sex determination on subadults was not attempted.
Adult age indicators were confirmed mainly from pelvic features. The
morphology of the pubic symphysis according to the Suchey-Brooks scoring system
(Suchey & Katz 1986; Brooks & Suchey 1990) provided the most accurate age ranges.
The appearance of the auricular surface of the ilium (Lovejoy et al. 1985) supplemented
age determination. Interestingly, age ranges produced using gross morphology of the
sternal rib ends (Iscan et al. 1984a, 1984b, 1985) did not appear to match the pelvic
indicators in this collection, indicating the need for a population-specific criteria. This
problem is currently being addressed in a forthcoming thesis by Mr. Constantine
Eliopoulos, MSc, in his research using this same collection. Lastly, the degree of cranial
suture closure (Meindl & Lovejoy 1985) and degree of vertebral arthritis and joint
degeneration (McKern & Stewart 1957) were used to confirm age in the absence of more
accurate criteria as well as to support age estimation.
Subadult age is determined mainly by long bone lengths, union of primary
ossification centers (Stewart 1979), tooth formation (Moorees et al. 1963a, 1963b), tooth
eruption (Schour & Massler 1941; Ubelaker 1999) and epiphyseal union (Buikstra &
Determination of ancestry is best accomplished by examination of cranial
features. Morphological and metric aspects of the midfacial paranasall) area, facial
profile and vault shape have shown to be especially accurate for ancestral determination
(Krogman 1962; Byers 2002). Overall, contemporary Greeks conform to the classic
European standards of craniofacial morphology. Interestingly, personal observation of
the Greek sample showed frequent metopism in both males and females as well as
pronounced supraorbital ridge development in many females.
To explore statistical relationships among the data, the prevalence of trauma by
bone/region and the distribution of trauma patterns were tested for independence and
significance for each biological and demographic category (sex, age, cause of death, last
known occupation). Summaries of percentages, means, cross-tabulations, ANOVAs, chi
square tests for independence between variables and loglinear modeling of multivariate
categorical data explored patterns in the data.
Descriptive Demographics of the University of Athens Skeletal Sample
Sex and Age
The total sample size from the University of Athens (UA) collection is 121
individuals. There are 62 (51.2%) males and 59 (48.8%) females. Ages span from 3 to
99 years. The overall mean age is 58 years. When the three subadults in the sample are
omitted (3, 6 and 14 years-old), the mean age increases to 59 years. Table 7-1 displays
the frequency of individuals in each age category by sex. Figure 7-1 shows how age at
death in this collection is skewed toward individuals in the middle adult to older adult
years for both males and females.
To compare the demography of the UA collection with the population of Greece
at large, the United Nations Economic Commission for Europe (UNECE) provides data
for Greece for the year 1997 (Table 7-2). This year was chosen as the target year for
comparison as this was when the majority of individuals in the Greek skeletal sample
died. Disparities in the percentages of individuals making up each age category between
the two groups do not actually indicate true population differences. Rather, the UA
collection was amassed primarily to establish new identification protocols (namely, aging
and sexing techniques) for contemporary Greeks. Individuals in the collection were
chosen specifically for skeletal preservation and specific age than for any true
demographic representation. Table 7-2 illustrates this bias in collection choices.
Table 7-1. Age category frequency by sex for the UA sample
Table 7-2. Percentages of deaths by age category for UNECE data (1997) and the
UA skeletal sample
Age category UNECE* UA sample
65-74 21.8 16.5
75-84 32.0 14.9
85+ 27.4 5.0
*UNECE Statistics for Europe and North America, Demographic Database.
Oto 15to 25to 35to 45to 55to 65to 75to 85+
14 24 34 44 54 64 74 84
Figure 7-1. Age category frequencies by sex for the UA sample
Cause of Death
Causes of death as listed on associated death certificates ranged widely from
common cancer- and heart-related illnesses to infrequent drug overdoses and
asphyxiations. To facilitate statistical analysis, causes of death were compiled into
somewhat discrete categories (Table 7-3). Chronic diseases of the heart were overall the
most frequent illnesses leading to death in the UA sample, followed by numerous forms
For comparison, the World Health Organization (WHO) provides data on mortality
for all of Greece. Again, the year 1997 was chosen as the target year for comparison.
Table 7-4 lists select cause of death categories and frequencies for the UA sample and for
all Greeks who died in 1997. By and large, chronic cardiac- and cancer-related illnesses
are the leading causes of death between the two groups. Respiratory illness was not as
common as stroke in the Greek population at large in 1997; the opposite is true for the
UA sample. Renal and cerebral illnesses, in addition to suicide, remain the least frequent
cause of death among Greeks.
Associated death certificates provide information on last known occupation for
nearly half of the individuals. Many in this collection are elderly pensioners and are
included in the unknown category. Table 7-5 graphically displays the distribution of last
known occupation among the UA skeletal sample. For comparison, Table 7-6 displays
employment categories and frequencies by select sectors for all Greek citizens during the
Table 7-3. Cause of death categories and frequencies for the UA skeletal sample
Cause of death Cause of death as listed on death n %
Cardiac heart attack 43 35.5
vegetative myocardis [sic]
Cancer generalized 37 30.6
Respiratory pneumonia 11 9.1
Stroke stroke 7 5.8
Other sipshaemic [sic] shock 6 5.0
Trauma heavy craniocerebral injury 5 4.1
skull and brain trauma
Renal chronic renal deficiency 3 2.5
TB-C of kidney
Cerebral degeneration of brain 3 2.5
advanced necrosis of brain matter
Unknown unknown 2 1.7
Drug Overdose drug overdose 1 0.8
Table 7-4. Select cause of death categories and percentages for all Greek deaths in 1997
and the UA skeletal sample
Category All Greeks* UA sample
Cardiac 14.3 35.5
Cancer 23.4 30.6
Respiratory 3.5 9.1
Stroke 18.4 5.8
Trauma 2.3 4.1
Renal 1.5 2.5
Cerebral 1.6 2.5
Suicide 0.4 1.7
*Based on 99,738 deaths. Compiled from WHO Annual Statistics: Table 1: Numbers of
deaths and death rates. Greece, 1997.
Table 7-5. Occupation category, number and percentage for the UA skeletal sample
Occupation n %
Unknown 56 46.28
Domestic 36 29.75
Private sector 23 19.01
Military 5 4.13
Civil servant 1 0.83
Table 7-6. Select employment sectors, frequencies and percentages for Greece, 1997
Category n %
Private sector 2,886,600 27.5
Civil servant 967,500 9.2
Total employed* 3,854,100 36.7
Total unemployed* 440,400 4.2
*Based on individuals 15+ years of age. Compiled from The Yearbook of Labour
Statistics. International Labour Organization, Bureau of Statistics, LABORSTA,
The skeletal remains of the individuals comprising the UA sample are very well
preserved. Brief burial time and careful handling by cemetery caretakers result in the
majority being fully represented. Injury frequencies were first calculated by the per bone
method (number of injured elements divided by the number of elements present). Injury
frequencies were also calculated by the per individual method (number of injured
elements divided by the number of individuals present). Finally, injury frequencies were
calculated by sex. See Table 7-7.
Table 7-7. Trauma frequency percentages by element for the UA skeletal sample.
Element % per bone % per % per male % per
Facial 12.7 12.4 21.9 1.8
Cranium 11.7 11.6 14.1 8.8
Cervical vertebrae 2.5 5.8 6.2 5.3
Thoracic vertebrae 7.3 26.4 29.7 22.8
Lumbar vertebrae 3.6 11.6 15.6 7.0
Sacrum 11.2 9.1 7.8 10.5
Left rib 5.1 26.4 26.6 26.3
Right rib 5.6 27.3 25.0 29.8
Sternum 3.1 2.5 0.0 5.3
Left clavicle 1.8 1.7 3.1 0.0
Right clavicle 1.7 1.7 3.1 0.0
Left scapula 1.7 1.7 1.6 1.8
Right scapula 2.5 2.5 3.1 1.8
Left humerus 4.2 4.1 4.7 3.5
Right humerus 5.1 5.0 4.7 5.3
Left ulna 7.8 7.4 10.9 3.5
Right ulna 5.2 5.0 4.7 5.3
Left radius 12.2 11.6 10.9 12.3
Right radius 8.5 8.3 9.4 7.0
Left carpal 2.0 2.5 4.7 0.0
Right carpal 3.3 3.1 3.5
Left metacarpal 1.1 3.3 4.7 1.8
Right metacarpal 3.3 4.7 1.8
Left carpal phalanx 0.2 0.8 1.6 0.0
Right carpal phalanx 0.0 0.0 0.0
Left pelvis 2.5 2.5 1.6 3.5
Right pelvis 4.3 4.1 4.7 3.5
Left femur 5.9 5.8 4.7 7.0
Right femur 12.0 11.6 14.1 8.8
Left patella 0.0 0.0 0.0 0.0
Right patella 3.2 1.7 0.0 3.5
Left tibia 8.4 8.3 10.9 5.3
Right tibia 3.4 3.3 3.1 3.5
Left fibula 5.1 5.0 4.7 5.3
Right fibula 1.7 1.7 3.1 0.0
Left tarsal 1.3 5.0 6.3 3.5
Right tarsal 1.1 4.1 7.8 0.0
Left metacarpal 0.8 6.6 6.3 7.0
Right metacarpal 2.5 3.1 1.8
Left pedal phalanx 1.7 3.1 0.0
Right pedal phalanx 2.5 3.1 1.8
In order to address the research objectives presented in Chapter 1, numerous
statistical analyses were performed. Power analysis tested for sample size adequacy.
Next, each research question is addressed. The prevalence of trauma by bone/region and
the distribution of trauma patterns were tested for independence and significance for each
biological and demographic category (sex, age, cause of death, last known occupation).
Summaries of percentages, means, cross-tabulations, ANOVAs, chi square tests for
independence between variables and loglinear modeling of multivariate categorical data
explored patterns in the data.
Sample size adequacy is tested via power analysis. In order to maximize the
chances of estimating a sample statistic that closely approximates the actual parameter,
power analysis ensures that the sample size is adequate. The sample size for this study is
based upon a statistic that is statistically significant at the 0.05 level, with a 5%
confidence interval using the following formula proposed by Krejcie & Morgan 1970:
sample size =
where: 2 = chi-square formula for 1 degree of freedom (X2 = 3.841)
N = population size
P = population parameter of a variable (P = 0.5)
C = confidence einterval (5%)
The current population size of the UA sample is 171. Using this equation, a
statistically adequate sample size for this research is 188 individuals.
(.05)2(170) + (3.841)(.5)(.5)
Research Question 1: Does the Relatively Low Socioeconomic Status of the Average
Athenian Result in Increased Trauma in the UA Sample?
Compared to neighboring Balkan and Mediterranean countries, Greece appears to
prosper in certain socioeconomic aspects such as average annual income, level of
education and annual expenditure on health (Table 7-8). However, when viewed against
other westernized countries with cosmopolitan and economically competitive urban
centers such as the United States and Italy, the population of Greece has comparatively
low socioeconomic levels. This disparity reflects Greece's political, economic and
geographic location as a crossroads between Balkan repression and the modern European
Since entering the European Union in 1981, Greece's economy has rapidly changed
from a focus on the provincial agricultural lifestyle to an increased urban service industry
catering to the needs of modernization. Greece's claim on the commercial shipping
industry has grown during the last two decades to assume 9% of the global merchant fleet
(U.S. Dept. of State Post Report 2003). Despite recent urban modernization, this is still a
country that must import most of its food, machinery and raw materials.
Greece depends heavily on tourism for national income. Of the 71.6% gross
domestic product provided by the service sector in 2003, over 7% originated from
tourism receipts alone (UNECE Trends in Europe and North America, 2003). This
limited, tourist based economy is especially felt in Athens with its focal resource
competition and confined, physically treacherous environment. Additionally,
unemployment in Greece has remained consistently high for the last decade and at 10.2%
in 2001, it was the second highest rate in the European Union (UNECE Trends in Europe
and North America, 2003).
Table 7-8. Select socioeconomic indicators for Greece, neighboring Mediterranean
countries and the U.S.
income per capital
total students per
1,000 pop., 2001
on health (%
Life expectancy at
birth (yrs), m/f,
per km2, 2003
% urban, 2003
Greece Cyprus Turkey
3,739 15,863 24,436
75.4/80.7 65.3/67.2 66.8/72.5 70.5/74.8 76.7/82.9 74.1/79.5
*Former Yugoslav Republic of Macedonia. Compiled from UNECE Trends in Europe
and North America, The Statistical Yearbook of the Economic Commission for Europe,
Given Greece's marginal socioeconomic levels, it is surprising that crime and
violent mortality rates in the country are among the lowest in the EU and also well below
the United States (Tables 7-9 and 7-10). Overall homicide rates are below the EU
average and suicide rates continue to be the absolute lowest for all reference countries
(lerodiakonou et al. 1998).
Table 7-9. Crime indicators, per 100,000 population, for Greece and the U.S., year 2000
Serious assault 69.8 323.6
Reported rapes 2.3 32.1
Robbery, violent theft 16.6 144.9
Homicides 2.8 5.5
Compiled from INTERPOL International Crime Statistics 1989-1990 and 2000,
www.interpol.int/Public/Statistics/ICS, and the Fifth United Nations Survey of Crime
Trends and Operations of Criminal Justice Systems, 1990-1994, UNCJIN.
Table 7-10. Violent death rates per 100,000 population for males/females, Greece and
the U.S., year 1997
Suicide 6.2/1.0 18.7/4.4
Homicide 2.6/0.6 13.8/3.9
Compiled from WHO Annual Statistics, Table 1: Numbers of death and death rates.
Greece, 1997, and United States of America, 1997.
Socioeconomic variables such as low income, low education, high unemployment
rates and high population density are significantly associated with increased frequencies
of assault, abuse and other violent crimes. Accordingly, one should expect to see
increased frequencies of skeletal trauma in a sample based on a population with low
socioeconomic status. Instances of poorly set or non-treated fractures where individuals
either could not afford adequate medical care or it was simply not available might be
As stated in Chapter 2, craniofacial trauma tends to be more indicative of
interpersonal aggression. Victim identity is focused in the face and thus aggression to the
individual is often directed there (Galloway 1999). In addition to the face, Fonseka
(1974) found that the thorax and ventral surfaces were areas most injured during episodes
of spousal abuse.
Only twelve individuals (9.9%) in the UA sample display antemortem, non-motor
vehicle accident trauma to the cranium, mostly found in the paranasal region. All trauma
is found in males, except for one female, which suggests some form of sex-specific
(intramale?) aggression is taking place. Of the 11 males, 5 injuries are found on the left
side and 9 are on the right. The female exhibits a perimortem linear fracture to the left
temporal extending into the left parietal. Her cause of death is noted generically as a
'brain hemorrhage' on the death certificate, so it is possible that this injury was due to a
motor vehicle accident or it could be a fracture (contra coup?) from blunt force trauma of
unknown mechanism. She does not display any additional skeletal trauma except for a
perimortem compaction-type fracture to the distal epiphysis of the left ulna.
As seen in Table 7-7, rib fractures are frequently encountered in this sample. No
significant preference for side or exact region on the rib can be discerned. Several
individuals have multiple rib fractures in different stages of healing, which might suggest
aggression. However, the lack of concomitant craniofacial trauma, as well as the
osteoporotic state of the majority of skeletons, suggests either accidental (fall, motor
vehicle accident) or underlying pathological origins (cancer-related osteoporosis) of the
No skeletal indications of overtly intentional violence (i.e., parry fractures or lethal
sharp force/blunt force/projectile trauma) are found in this collection. Nor is there any
indication of chronic physical abuse such as multiple fractures in various stages of
healing outside of the above mentioned rib fractures.
Diet and health
Low socioeconomic status also generally precludes access to adequate diet and
health care. Evidence of poor nutrition such as chronic anemia caused by iron deficiency
in the diet can be seen osteologically as perpendicular orientation and expansion of the
trabeculae in the cranial diploe, thinning of the compact cranial bone (porotic
hyperostosis) and thickening of the orbital roof (cribra orbitalia). These lesions result
from hypertrophy of the blood-forming tissues in the marrow in order to increase the
production of red blood cells in response to the anemia. The increase in marrow
production results in the replacement of the outer table of compact bone with exposed
diploic bone, which gives the appearance of raised and porous zones of skeletal tissue
usually on the cranial vault and/or orbits (Stuart-Macadam 1987). A variety of
nondietary or genetic factors may also cause iron deficiency including thalassemia, sickle
cell anemia, traumatic blood loss, parasitic infection and chronic diarrhea (Stuart-
Only three individuals (1 male, 2 females) in the UA sample display cranial vault
porotic hyperostosis, while 5 individuals (3 males, 2 females) have cribra orbitalia in
various stages of activity and healing at death. As age and cause of death range widely
among those affected, no particular pattern is discerned regarding underlying diet,
behavior or disease.
Health Care Funding
Funding is a perpetual problem in the development and improvement of Greek
medical care. National funding cuts resulted in a decrease in the total number of hospital
units in all of Greece from 595 in 1984 to 341 in 1998 (Greece in Figures 2003, National
Statistical Service of Greece). Financial roadblocks also affect research. Indeed,
attempts to organize a basic trauma registry at a teaching hospital in Crete were met with
both financial setbacks as well as general lack of enthusiastic participation from
overworked hospital staff (Sanidas et al. 2000). Additionally, a recent study out of a
hospital in Piraeus found that an alarming 42% of all DOAs could possibly have been
prevented if pre-hospital emergency medical care had received funding for basic
technological upgrading (Papadopoulos et al. 1996).
There are 201 health centers situated in Greek rural and semi-urban areas
(Theodorou 2003). These facilities are administratively linked to large urban health
centers and are mostly staffed by recent medical graduates, who are required to spend one
year of service in a rural area upon graduation. Limited human and financial resources,
organizational problems and a lack of clinical experience by doctors raises concerns
about the quality of service delivered at satellite health centers. Although it is not known
for certain, some of the injuries found in the UA collection may have occurred outside of
greater Athens in rural areas lacking optimum health care. This collection in particular is
composed of those less wealthy citizens who for the most part could not afford permanent
placement in the mausoleums of the Zografou Cemetery in Athens. The following
highlights trauma found on some of the individuals.
Figure 7-2 illustrates a well-healed fracture to the distal left radius in a 59-year-old
female. During a fall onto an outstretched pronated arm, the dorsal surface of the arm is
placed under compression while the ventral surface is under tension. The tensile forces
cause a transverse fracture and there is a subsequent crumbling of the posterior (and
sometimes carpal) surface. These fractures tend to result most often from low energy
trauma such as a fall from a standing height (Galloway 1999). The carpal articular
surface of this individual is comminuted and the afflicted radius is 13 mm shorter than
the contralateral element. There is a 5-degree dorsal angulation. No evidence of internal
or external fixation (i.e., implanted hardware, resorbed drill tunnels) is seen.
Residual deformity from a poorly treated radial fracture may have subsequent
effects on wrist movement due to the precise geometry of the wrist joint. Loading
patterns across the wrist are affected by very minor changes in distal radial geometry as
little as 2.5 mm of radial shortening significantly shifts force loading to the distal ulna.
This disturbs the relationships and the forces at the distal radioulnar joint, which
manifests as pain and limitation in forearm rotation (Cohen et al. 2003). Although the
injury to this specific individual healed very well without surgical intervention, it is
surprising that open reduction was not pursued.
Alffram & Bauer (1962) found that in older women, distal forearm fractures occur
more often with simultaneous fractures of the proximal end of the femur, as might be
seen in falls from a standing height. Of the 12 radii with distal fractures and the 11
proximal femora with fractures in the UA sample, only 3 individuals have fractures to
both sites. Two of the individuals are elderly males and the other is an elderly female.
Figure 7-2. Fractured distal left radius (left) and normal contralateral right radius (right),
Figure 7-3 shows another fractured left distal radius, this one in a 73-year-old
female. There is marked dorsal angulation and 9 mm shortening of the afflicted. Again,
there is no evidence of open reduction.
Figure 7-3. Fractured distal left radius (left) and normal contralateral radius (right),
Figure 7-4 shows a fractured right proximal femur in an 81-year-old male. Post-
operatively, the femoral neck retains noticeable posterior angulation with 16 mm
shortening of the afflicted element. There is marked resorption around the neck
component of the appliance. Poor appliance placement on the diaphysis resulted in
minimal investment of the cortical screws. Such placement results in a biomechanically
disadvantage femur where axial loads are not fully supported during healing.
(a) (b) (c)
Figure 7-4. Fractured right proximal femur: (a) posterior, (b) posterior-lateral close-up
of appliance and (c) posterior close-up of cortical screw in diaphysis
Figure 7-5 shows a complication of flexible orthopaedic appliances (Ender nails)
used to internally fix a fractured greater trochanter and neck of the right femur in a 78-
year-old male. Either excessive force during insertion or inappropriate post-operative
axial loading by the patient resulted in the nails bursting through the distal anterior femur
superior to the condyles.
Figure 7-5. Orthopaedic fixation in a fractured greater trochanter and neck of the right
femur, anterior view
Figure 7-6 shows a fractured neck and proximal diaphysis of the right femur in an
85-year-old male. The neck component of the orthopaedic appliance is poorly aligned;
threads are seen exposing from the cortex of the neck. The femoral head is angled
anterior and marked reactive bone embedding the appliance suggests inappropriate post-
operative movement of the fractured ends.
(a) (b) (c)
Figure 7-6. Fractured neck and proximal diaphysis of the right femur: (a) anterior, (b)
posterior and (c) posterior close-up
Osteoporosis and Fractures
Compared with North American, Finnish and German populations, both Greek men
and women have consistently lower bone mineral density (BMD) in the decades over 40
years. In a study on BMD of the vertebrae, proximal femora and calcaneii in normal
Greeks, Hadjidakis et al. (1997) found that the total bone loss between ages 20 and 70 is
29.5% for the vertebrae and 32% for the femoral neck in women, whereas the values for
men are 19.5% and 29% respectively. Hip fractures, the most dramatic complication of
osteoporosis, has shown an average annual increase of 7.6% in Greeks from 1977-1992,
due partly to an aging demographic (Paspati et al. 1998). While not clinically measured,
the majority of individuals in the UA sample are to some degree osteoporotic. Fragile,
lightweight bones are seen especially in the females, but also in many of the males. The
advanced age of the average individual comprising the sample, as well as the high
prevalence of cancer and other chronic illnesses among the group, likely contribute to
frequently osteoporotic skeletons.
Research Question 2: Do the Specific Geographical and Population Characteristics
of Athens Result in Increased Trauma in the UA Sample?
Bioarchaeological research has highlighted the importance of various geographical
and population factors on trauma occurrence. Using a prehistoric Indian sample from the
Channel Island area of southern California, Walker (1989) demonstrated the influence of
resource competition stress in the geographically circumscribed area on skeletal trauma
frequencies. High population density and social stress brought on by resource
competition and land scarcities are likewise blamed for elevated craniofacial trauma
found in preceramic northern Chilean coastal communities (Standen & Arriaza 2000) and
Sudanese Nubians (Alvrus 1999).
Bioarchaeological studies have associated certain geographical features with
elevated skeletal injuries. The uneven geophysical terrain of Nubia is suggested as the
particular cause of elevated accidental appendicular fracture frequencies found at the
neighboring sites of Semna South (Alvrus 1999) and Kulubnarti (Kilgore et al. 1997).
Both sites are in the Batn el Hajar region just south of the Egyptian border an area
characterized by a dry and boulder-strewn landscape.
Greece is surrounded by water on three sides: the Ionian Sea to the west, the
Aegean Sea to the east and the Mediterranean Sea to the south. These adjoining seas are
studded with thousands of rocky islands, of which only 200 or so are habitable. Three-
fourths of Greece's terrain is rocky with little or no significant topsoil (U.S. Dept. of
State Post Report 2003).
Athens is located in the southeastern aspect of the Attican Peninsula of the Greek
mainland. It is the largest city of the country and with an estimated 3,700,000
inhabitants, is home to 40% of the Greek population (U.S. Dept. of State Post Report
2003). The city itself is geographically enclosed by the Parnis, Pendeli and Hymettos
mountain ranges on three sides. The Gulf of Saronikos the inlet of the Aegean Sea -
provides the fourth border.
The current population density of Greece is 80 individuals per km2. The city of
Athens, however, has a population density of 923 individuals per km2, with 60% of the
entire Greek population living in an urban setting (UNECE Trends in Europe and North
As in many countries in Europe, the total fertility rate of Greece is nearly below
replacement value. As a result, the rate of natural increase in Greece's population as a
whole has shown a slow growth of only about 0.06% per year, due mostly to immigrants
from Albania, the Former Yugoslav Republic of Macedonia and the Russian Federation
(WHO Highlights on Health in Greece 1998). Greater Athens, however, has undergone
significant recent population surging, with numbers increasing almost two-fold in the last
three decades due to an influx of foreign immigrants and rural Greeks into the capital
pursuing economic opportunity (Table 7-11). Growth within the country as well as in the
capital are physically constrained within naturally limiting geographical boundaries
compounded by mountainous and uneven terrain.
Table 7-11. Total population increases for Greece and greater Athens for select
1971 1981 1991 2001
Greece 8,894,981 9,667,336 10,134,534 10,206,539
Athens 1,985,221 2,276,750 2,519,661 3,700,000*
*2004 estimate. Compiled from Greece in Figures 2003, National Statistical Service
Despite a high unemployment rate, Greeks continue to make significant purchases
such as automobiles, which are unfortunately a necessity in the city due to limited mass
transportation. Over one million registered vehicles clogged the streets of Athens in 2003
(U.S. Dept. of State Post Report 2003). Given the geographic constriction of the city, this
increase in road traffic has led to a subsequent increase in motor vehicle accident (MVA)
fatalities. Indeed, Greece is currently the third highest for MVAs among all European
Union countries (Kardara & Kondakis 1997).
Four individuals in the UA collection display obvious perimortem high impact
trauma caused by vehicular accidents (it is not noted on the death certificates whether the
individual was a vehicle occupant or pedestrian). Skeletal trauma on these individuals
includes multiple craniofacial and postcranial fractures. Figure 7-7 shows craniofacial
injuries sustained by a 43-year-old male in a motor vehicle accident.
Figure 7-8 shows a 34-year-old male who sustained multiple fractures from some
type of (vehicular?) impact mainly to the left side of the body. The left tibia has an
oblique fracture to the proximal diaphysis. The afflicted element has 26 mm shortening,
90% apposition and lateral displacement of the proximal segment. The ipsilateral fibula
has a transverse fracture to the proximal 14 of the diaphysis, 5 mm shortening, 90%
apposition and anterior displacement of the proximal segment.
Figure 7-7. Craniofacial trauma caused by a motor vehicle accident in a 43-year-old
male: (a) trauma to the right frontal bone including medial orbital elements
and (b) close-up of the right orbital area. The sclerotic callus and healed
fracture in the right superior orbit lateral to the perimortem injury are
evidence of a previous traumatic episode.
Figure 7-8. Tibial fracture in a 34-year-old male: (a) left tibia, anterior view and (b) left
tibia, posterior view
Despite ecogeographical constraints, high unemployment rates, focalized resources
competition and a high urban population density, non-motor vehicle accidents and other
violent mortality rates in Greece are among the lowest in EU countries and are also well
below the United States (Table 7-12). Certainly, skeletal injury in the UA sample seems
to be overwhelmingly non-violent in origin.
Table 7-12. The 1997 death rates per 100,000 population for males/females, Greece and
the United States
Males Females Males Females
MVAs 34.3 10.4 21.3 10.6
Accidental falls 4.5 2.7 5.9 5.7
Suicide 6.2 1.0 18.7 4.4
Homicide 2.6 0.6 13.8 3.9
WHO Annual Statistics. Table 1: Numbers of death and death rates. Greece, 1997, and
United States of America, 1997.
Nineteen individuals (15.7%) in the UA collection sustained trauma to the forearm
and/or hand. Of these, 15 involved the distal radius, mainly taking the form of Colles',
Smith's and radial styloid fractures. Two scaphoid fractures are also found. The
locations of these injuries are suggestive of falls on outstretched hands. Distribution of
the injuries is spread evenly between males and females with an average age of 70 years.
The rocky and uneven terrain of Greece, as well as the osteoporotic state of the majority
of these aged individuals, possibly led to this relatively high frequency of wrist injuries.
Seven individuals (5.8%) have trauma to the distal tibia, fibula or calcaneus that are
suggestive of ankle-twisting injuries or falls. Fractures to the distal diaphysis of the tibia,
medial malleolus, fibular styloid process, and talar surface of the calcaneus are the most
common ankle injuries in the collection. As with the wrist injuries, the distribution of
ankle injuries is spread evenly among males and females. The average age of those with
ankle injuries is 69 years. Injuries in these ankle locations reflects accidental trauma,
probably while traversing uneven ground.
Does the Age Bias of the Individuals in the UA Sample Result in Increased Skeletal
The average life expectancy for Greek males and females in 2002 was 75.8 and
81.1 years, respectively (World Health Organization 2003). Greek life expectancy
continues to be one of the highest in the EU community and also surpasses the average
for the United States, even though Americans spend approximately 13% of their annual
income of health compared to 8.4% by Greeks (see Table 7-13).
Table 7-13. Life expectancy at birth for males/females: Greece, neighboring
Mediterranean countries and the United States
Country Life expectancy in years
Greece 75.8 81.1
Italy 76.8 82.5
Cyprus 75.5 79.1
Turkey 67.9 72.2
U.S. 74.6 79.8
WHO Annex Table 1: Basic indicators for all member states, 2002.
There has been a considerable change, however, in the population structure of
Greece; the number of people aged 65 years and over has increased from 11% to 15%,
while the percentage of the population aged less than 15 years has decreased from 25% to
17% (WHO Highlights on Health in Greece 1998). Low fertility, increased longevity and
emigration of young adults from Greece for education and employment opportunities
elsewhere have led to an aging of the population. Figure 7-9 displays this trend in the
Greek population for the decade 1990-2000.
Assuming that as one ages, one is chronologically exposed to more opportunities
for traumatic episodes, the advanced age of many of the individuals in the UA collection
should result in increased cumulative antemortem injuries. The Pearson r test for
association between age and the total number of injuries per individual resulted in a slight
positive correlation (r = .169). Thus, as age increases, so does total number of injuries.
0-14 15-24 25-34 35-44 45-54 55-64 65-74 75-84 85+
Figure 7-9. Age category frequencies for all Greeks, 1990 and 2000. Compiled from:
UNECE Trends in Europe and North America. The Statistical Yearbook of
the Economic Commission for Europe 2003.
Biological Variables Associated with Skeletal Trauma
Known biological variables for the UA collection include sex and age at death.
Crosstabulations were computed for sex and general trauma location in order to
determine if any significant relationship between the two variables. To facilitate analysis,
all skeletal trauma was compiled into categories based on location. The categories
include: face, head, rib/sternum, vertebrae, shoulder girdle, humerus, arm/hand, pelvic
girdle, femur and leg/foot. Significance relationships exist between sex face (X2
12.536, p=.00) and sex vertebrae (2 12.969, p=.00), with males having significantly
more trauma in these two locations than females. No relationship was detected between
sex and any other location.
Chi-square tests for independence were conducted on age and the same general
trauma locations described above. Numerical age for each individual was first
categorized into one of eight categories: 1 (21-30), 2 (31-40), 3 (41-50), 4 (51-60), 5 (61-
70), 6 (71-80), 7 (81-90) and 8 (91-100). No significance was found. Age categories
were then compressed to include just three categories: 1 (21-40), 2 (41-60) and 3 (61-
99). Still no significant associations were found.
Signifance testing was also performed to determine if males experienced more
cumulative trauma than females. While there is a difference in the frequency of multiple
trauma by sex (average 1.5 injuries per male, 1.9 per female), a one-way ANOVA test
proved it not significant (p=. 115).
Population-Specific Evidence of Accidental Versus Intentional Injury
The study of interpersonal violence as inferred from the skeleton is highly
intriguing to both scholars and the lay public. However fascinating it is to read such sexy
reports, osteologists often have difficulty determining if trauma observed on an individual
skeleton is the result of accidental or intentional mechanisms in the absence of associated
weaponry. Certain types of trauma, such as parry fractures to the ulna and cranial pond
fractures, are often cited as proof positive that intentionally violent encounters occurred
(Wells 1964), although the true causal behavioral link suggested is unknown.
Keeping these factors in mind, one can still carefully attempt to examine patterns of
skeletal trauma, with the hope that causal mechanism will be illuminated. Such an
analysis must take into consideration all known demographic and cultural variables of the
collection. For example, age and sex are important dimensions of the modern violence
pattern in that various cultural factors may make one sex or age group more vulnerable to
aggression than another. Modern assault victims show a distinctive distribution of
skeletal injuries with high facial trauma rates, especially in cases of abuse to females.
The upper limb is typically the next most common injury site. Novak (1999) found that
tandem craniofacial and thoracic injuries suggest aggression, while solo appendicular
injuries suggest accidental mechanisms.
To test for possible relationships between the coexistence of craniofacial, thoracic
and appendicular injuries in the UA sample, loglinear modeling was applied using SPSS
Advanced Models 12.0. The loglinear technique models the means of cell counts in
multi-dimensional contingency tables by describing the association patterns among a set
of categorical variables without specifying any variable as a response (dependent)
variable. It is structured to fit hierarchical linear models to crosstabulations using
iterative proportional-fitting algorithms (SPSS Advanced Models 12.0). These
techniques allow the analysis of chi-square-type data using regression-like models. They
are essentially multiple linear regression models in which the classification variable and
their interaction terms are the independent (predictor) variables, and the dependent
variable is the natural logarithm of the frequency of cases in a cell of the frequency table.
Using the natural log of the frequencies produces a linear model.
Loglinear analysis focuses on studying associations between pairs of variables
rather than modeling the response on one of them in terms of the others. The loglinear
model formulae express the logs of cell expected frequencies in terms of dummy
variables for the categorical variables and interactions between those variables (Agresti &
Finlay 1997). Thus, all variables in the multi-dimensional contingency table used in
loglinear models are response variables, rather than one a response and the others
explanatory. The resulting model describes associations in partial tables that relate, for
example, two of the variables while controlling for the third one.
The term loglinear comes from the form of the model; the natural logarithm of cell
counts is modeled as a linear function of the effects of categorical variables and their
relationships. For example, to investigate relationships between three categorical
variables X, Y and Z, the full (saturated) loglinear model is:
log(m) = t + kX + X v + Y z + kXY + kXZ + kYZ + kXYZ
where: kx, kv and kz represent the 1st order main effects of the independent variables
iXY, Xxz and vYz represent the 2nd order interactive effects
XXYZ represents the 3rd order interactive effect of all variables X, Y and Z
In the saturated model, all terms correspond to all possible main effects and interactions,
and thus the model fits the data perfectly. This does not provide any information
regarding the effects of any variable or possible interactions of variables, so one then
estimates a non-saturated model containing a subset of the parameters from the saturated
model and tests the difference between that and the saturated version. Such tests are
called "goodness of fit" tests, because they tell whether the model in question is
significantly worse than the saturated, or perfect, model. The likelihood ratio best tests
the goodness of fit. If the reduced model is true, the likelihood ratio statistic has a
distribution that approaches a chi-square distribution as the sample size increases.
To preserve statistical power, there need to be at least 5 times the number of cases
as cells in the data. For example, for a 2 x 2 x 3 contingency table, one needs to have at
least 60 cases. If the required amount of cases is not met, then the sample size needs to
be increased, the number of variable categories needs to be compressed or variables need
to be eliminated.
To perform loglinear modeling on the UA data, trauma locations were first
organized into six skeletal loci: craniofacial, thoracic, proximal upper appendicular,
distal upper appendicular, proximal lower appendicular and distal lower appendicular.
Table 7-14 defines the elements of each locus. The resultant contingency table contained
an excessive number of cells with less than the minimum of five counts, so the loci were
compressed to include craniofacial, thoracic and a broader appendicular category
including all elements in the previous last four loci.
Table 7-14. Loci used in the loglinear model and their respective skeletal elements
Craniofacial all facial, cranial, and mandibular elements
Thoracic ribs, vertebrae, sternum
Proximal upper appendicular clavicle, scapula, humerus
Distal upper appendicular radius, ulna, carpals, metacarpals, phalanges
Proximal lower appendicular innominates, femur
Distal lower appendicular patella, tibia, fibula, tarsals, metatarsals, phalanges
Likelihood ratio output from loglinear testing of a hierarchical model compared to
the saturated model indicated no significance at the 1st' 2nd or 3rd order levels (p>.05). To
test the individual 1st order effects within themselves, crosstabulations and chi-square
tests for independence between the pairs craniofacial thoracic, craniofacial *
appendicular and thoracic appendicular were conducted. All tests resulted in non-
significant results at the .05 level (p=.217, .711, .104 respectively).
To further test for association between the categorical variables, odds and odds
ratios were computed for 2 x 2 tables cross classifying craniofacial thoracic and
craniofacial appendicular loci (Table 7-15). Since the odds ratio treats variables
symmetrically, it does not require identifying a response variable. This makes the odds
ratio a natural measure when there is no obvious distinction between the variables, such
as when they are both response variables (Agresti & Finlay 1997).
For individuals with craniofacial trauma, there are 2.14 (15/7) individuals with
concomitant thoracic trauma for every one individual without thoracic trauma. For
individuals without craniofacial trauma, the odds of having thoracic trauma equal 1.16
(51/44). This means that there are 1.16 individuals with thoracic trauma for every one
individual without thoracic trauma. For individuals with craniofacial trauma, the odds of
having concomitant thoracic trauma are 1.84 (2.14/1.16) times the odds of having
thoracic trauma without concomitant craniofacial trauma.
Table 7-15. Cross-classification of craniofacial, thoracic and appendicular trauma in the
Craniofacial Yes No Total Craniofacial Yes No Total
Yes 15 7 22 Yes 13 9 22
No 51 44 95 No 52 43 95
For individuals with craniofacial trauma, there are 1.44 (13/9) individuals with
concomitant appendicular trauma for every one individual without. For individuals
without craniofacial trauma, the odds of having appendicular trauma equals 1.21. This
means that there are 1.21 individuals with appendicular trauma for every one individual
without. For individuals with craniofacial trauma, the odds of having concomitant
appendicular trauma are 1.19 (1.44/1.21) times the odds of having appendicular trauma
without concomitant craniofacial trauma.
Put in simpler terms, these cross-classifications indicate that individuals with
craniofacial trauma also tend to sustain trauma in both the thoracic and appendicular
trauma. Also, individuals are somewhat more likely to display craniofacial and thoracic
trauma together than they would craniofacial and appendicular.
Another method to determine possible patterns in the coexistence of trauma at
different skeletal loci is to construct a 3-dimensional contingency table. Table 7-16 lists
the general location of skeletal trauma on all individuals in the UA collection. Each
location (craniofacial, thoracic and appendicular) has one of two possible responses (yes
Table 7-16. Three-dimensional contingency table using the UA data
Craniofacial Thoracic Appendicular % yes
Yes Yes 9 6 60.0
No 4 3 57.1
No Yes 31 19 62.0
No 20 24 45.4
From this table, the following statements can be made: (1) for individuals with
craniofacial trauma, appendicular trauma was seen 60% of the time when thoracic trauma
was also present and 57% of the time when thoracic trauma was not also present, (2) for
individuals without craniofacial trauma, appendicular trauma was seen 62% of the time
when thoracic trauma was present and 45% of the time when thoracic trauma was not
present, (3) for individuals without craniofacial trauma, appendicular trauma was seen
16.6% (62.0-45.4) more often when thoracic trauma was present that when it was absent
and (4) controlling for craniofacial trauma (by keeping it fixed), the percentage of
appendicular trauma is higher when thoracic trauma is also present.
While no obvious population-specific pattern of intentional versus aggressive
injury is found, these findings further underscore the importance of considering the entire
skeleton in trauma analysis when it is available. Lastly, although the UA sample does not
exhibit evidence of large-scale interpersonal aggression, the prevalence of craniofacial
injury in the group is modestly high (10.7%). The fact that 12 of the 13 individuals
(92.3%) are male suggests some type of cultural factor is involved with the frequency of
trauma to this area.
Occupation, Cause of Death and Sex
Crosstabulations were performed to test for possible relationships between
occupation and cause of death. After all cells with <5 were omitted (thus including only
the domestic and private sector employee categories for occupation and unknown, cardiac
and cancer categories for cause of death), no association was found between any
occupation or cause of death category (p=.611).
Similar crosstabulations were performed to test for possible relationships between
sex and cause of death category. After all cells with <5 omitted in the cause of death
categories (thus only including cardiac and cancer), significance is found between the
two variables (p=.04). A possible association exists, then, between sex and cause of
death such that females succumb more often to cardiac illness and males succumb more
often to cancer.
Comparison to Archaeological Samples
Trauma frequencies of the UA sample were compared to geographically similar
archaeological collections. Table 3-1 compiles skeletal analyses of various Greek and
Mediterranean sites published by J. Lawrence Angel. Several important caveats must be
taken into consideration when examining this table. As previously discussed, skeletal
trauma was not the focus of Angel's work. Rather, he was primarily interested in tracing
Greek social biology through cranial morphometrics (Angel 1944, 1946). He did,
however, attempt to document traumatic and pathological conditions found on certain
individuals. When Angel gave information on whether injuries were found in males or
females, it is indicated in the table. Blank cells do not necessarily mean that injuries were
absent in many of his reports, postcranial trauma was not discussed at all. Actual
cranial and postcranial trauma occurrences are likely higher than reported, as the
archaeological materials Angel analyzed were usually highly fragmented.
Archaeologically, fractures to the distal extremities are suggestive of consistent
interaction with the rough, sloping and mountainous terrain characteristic of Greece.
Indeed, the prevalence of postural indicators (i.e., squatting facets, Allen's fossae) and
the degree of muscular development noted among the individuals noted by Angel (and
discussed in Chapter 3) supports this assumption. There is also a relatively high
frequency of male cranial trauma suggesting elevated cultural stressors such as
interpersonal (intramale) aggression, internecine warfare, and probable defense of the
community against warring invaders.
As with the ancient samples, numerous individuals in the contemporary UA
collection display reactive areas on the femora called anterior cervical imprints, or
Allen's fossae (Capasso et al. 1999). This facet is found on the anterior aspect of the
medical femoral neck and is due to hyperextension of the femur and rotation of the head
in the acetabulum, usually from downhill walking or traversing mountainous landscapes
(Figure 7-10). The zona orbicularis normally resists hyperextension of the femur by
tightening around the femoral neck. The fossa forms where the zona bifurcates around
the iliofemoral ligament.
Figure 7-10. Left femur from the UA sample with Allen's fossa located inferior to the
anterior articular surface of the femoral head
The prevalence of male cranial trauma over female cranial trauma is also found in
the contemporary UA sample. Some type of cultural factor is causing elevated
frequencies of craniofacial injury in Greek males. Angel's warfare proposal makes sense
given the temporal and geographic nature of his ancient samples. Contemporary Greek
males, however, are not in the same political position as their ancient forefathers. This
sample in particular, while rather aged, is not of the correct demographic to have
participated in any large-scale political upheavals or warfare. Given the lack of other
credible reasons for frequent male cranial injuries, one is left with two possible scenarios:
(1) males are simply more physically active than females and therefore expose
themselves to more opportunities for accidents or (2) males are engaging in some type of
interpersonal (probably intramale) aggression.
The UA collection also contains more individuals with wrist and ankle injuries than
in Angel's ancient samples. As individuals from both the ancient and contemporary
groups were exposed to identical geography, it makes sense that both would exhibit
similar accidental trauma patterns to some degree. A likely reason why Angel's samples
show less appendicular trauma is due to the fact that he simply was not as concerned with
postcranial trauma as he was obvious craniofacial injuries and skeletal morphology.
Table 7-17. Research questions examined in this study and their results.
Does deprived socioeconomic No evidence of increased trauma, assault
status result in increased trauma, or abuse, but poorly treated fractures are
assault and abuse, medical care? common
Do limiting geography,
and increased population density
result in increased trauma?
Does advanced age result in
Which biological variables are
most associated with skeletal
Is there population-specific
evidence of accidental vs.
What is the relationship between
occupation and cause of death?
What is the relationship between
sex and cause of death?
Comparison to archaeological
Increased distal appendicular injuries, but
increased population density does not
necessarily lead to increased overall
Increased facial and vertebral trauma in
No. Concomitant trauma (craniofacial,
thoracic and appendicular) is more likely
than singular craniofacial trauma
Females die more frequently from cardiac
disease while males die more frequently
Similar male craniofacial trauma
dominance and increased distal
Summary of findings
Morphological analysis of a sample (n=121) of individuals from the University of
Athens skeletal material suggests that the majority of skeletal trauma is the result of
cumulative accidental episodes. Injuries observed show no distinct constellations
suggesting intentional, interpersonal violence. Significant trauma clustering is seen in the
thoracic vertebrae, ribs, radii and femora. These fracture loci are typical in an aged,
osteoporotic population. Crosstabulations, ANOVA and loglinear analysis show
associations between sex and trauma location, sex and cause of death category and some
degree of association between concomitant trauma locations. No statistical relationship is
found between age and trauma location.
Despite what is suggested in the bioarchaeological and clinical literature, the
presumed relatively low socioeconomic status of the individuals in the UA collection did
not result in increased skeletal trauma. Crime rates, violent mortality, and homicide and
suicide rates in Greece continue to be among the very lowest in the European Union and
well below rates in the United States. Only 9.9% of the individuals sustained non-motor
vehicle accident craniofacial trauma, a surprisingly low frequency given the significant
social and economic stresses this population is under. Only 2.5% exhibited active or
healed porotic hyperostosis and/or cribra orbitalia at death, the majority of which are
likely due to chronic diseases associated with death rather than any dietary deficiencies.
There is no evidence of overtly intentional trauma.
Injuries from traversing the treacherous terrain of Greece are found in the ankle and
foot area of 16% of the sample. Some 6% of the sample shows skeletal injury of the
forearm and/or hands, probably the result of falls. Population surging and environmental
circumscription have led to increasing trauma from motor vehicle accidents.
The UA collection seems to mirror that of the overall contemporary Greek
population by representing a subset of a modern, relatively nonviolent society. Despite
ecogeographical constraints, high unemployment rates, focalized resource competition
and a surging urban population, violent mortality rates in Greece are among the lowest
globally. Thus, contrary to socioeconomic stress theories suggested in the
bioarchaeological and clinical literature (Harries 1997; Cubbin et al. 2000a; 2000b;
Faelker 2000; Lyons et al. 2000; Wagner et al. 2000; Hasselberg et al. 2001; Blakely et
al. 2002), contemporary Greeks do not show increases in trauma from assault, abuse and
other violent crimes due to economic and populational stressors.
Diet and Disease
A Mediterranean diet high in natural oils, fish and vegetables has been shown to
improve longevity, especially coronary disease related mortality (Trichopoulou et al.
2003). Indeed, factors that most influence a significant reduction in coronary events
globally include physical activity, high levels of education and adherence to a traditional
Mediterranean diet (Panagiotakos et al. 2002). Several emerging lifestyle risk factors
among Greeks, however, such as increasing mental depression, smoking and adopting a
less healthful (fast food) diet have contributed to the steady increase in death rates from
coronary and chronic respiratory disease seen beginning in the 1960s (Chimonas 2001;
Karakatsani et al. 2003). The eradication of malaria, rheumatic fever, decrease in
infectious diseases, improvement of medical care system and the rise of the populations'
socioeconomic level during the last 5 decades in Greece have reduced all-cause mortality
(Chimonas 2001). Statistically, Greece remains the country with the lowest death rates
from coronary heart disease among major European countries (Menotti et al. 1999) and
with reference to total cancer mortality, Greeks have the lowest rates in the European
Union (Black et al. 1997; Levi et al. 1999).
As the UA sample is biased toward the elderly, perhaps the advanced age cohort
refrained more from the newer fast foods and kept to a healthier traditional
Mediterranean diet. Even so, chronic cardiac- and cancer-related illnesses are the leading
causes of death among all individuals in the UA collection but are especially common in
those individuals of advanced years.
Regarding macroscopic skeletal evidence of dietary distress or nutrition among
individuals in the UA collection, only three individuals (1 male, 2 females) display
cranial vault porotic hyperostosis, while 5 individuals (3 males, 2 females) have cribra
orbitalia in various stages of activity and healing at death. As age and cause of death
range widely among those affected, no particular pattern is discerned regarding
underlying diet, behavior or disease.
Greece has recently experienced an increase in drug and alcohol abuse as well as
drug related illegal behavior. Kokkevi and colleagues (1993) found male gender,
polydrug use, unemployment and low education level to be among the main factors
significantly predicting criminality of drug abusers in Greece.
It is important to consider the sociocultural context of substance use in the Greek
population. Especially among students, substance abuse has traditionally been
understood differently from that of many other European and North American countries.
Greece is a tobacco and alcohol producing country and there are no restrictions
whatsoever on their sales. This, in conjunction with the fact that they are both socially
accepted substances, explains their widespread use among the Greek population. While
the percentage of frequent alcohol consumers in Greece is among the highest in Europe,
the majority of the youth tend to drink only occasionally.
Among students, cannabis, regular tobacco and illicit drug use have shown sharp
increases during the 1990s while alcohol and unprescribed psychoactive medicines have
decreased (Kokkevi et al. 2000). So, while Greece is no longer among the low
prevalence European countries in illicit drug use, it still remains at much lower levels
than the United States.
Only one individual in the UA collection reportedly died of a known drug
overdose. However, the advanced age of the majority of individuals likely produces a
lower than expected prevalence of drug abuse as this age cohort probably did not engage
in such risk taking behaviors as much as younger individuals.
Child Abuse and Other Interpersonal Violence
Greece's recent transition from a traditional to an industrialized society has resulted
in cascading effects in Greek culture. The primary unit of change has involved the family
structure, where a gradual shift is observed in organization from collectivism to
individualism. The large, traditional extended family is beginning to change into a more
contemporary, nuclear unit with young family members often leaving the home and
country for economic pursuits elsewhere. These changes bring with it a number of high
risk factors at the individual, family, community and social levels concerning violence
due to social stressors. While still relatively low compared to other European countries,
Greece is just beginning to experience violent crime rates slightly above prevalence rates
in the previous decades. Petridou and colleagues (2001) found that an average of 0.23%
of all Greek children presenting at hospital emergency departments have injuries
attributed to acts of violence. These injuries are more common among migrant children
who are faced with their own suite of social, cultural and economic hurdles.
As in other areas of violence, income and education seem to be the most influential
factors associated with child abuse, where the mother's level of education mostly affects
the level of prenatal/perinatal care, infant mortality rate and the infant's future growth
and health care (Kafatos et al. 1978). Agathonos-Georgopoulou & Browne (1997) found
other high risk predictors to include a child's poor state of hygiene upon presentation to
the emergency ward, parents with mental health problems, poor relationship between
parents, parents with adverse life experiences and a mother strictly disciplined by her
No subadults in the UA collection bear skeletal evidence of abuse. The fact that
there are only three subadults in the entire sample should be taken into consideration,
however. There is also no skeletal evidence of abuse to the elderly.
No population-specific data on skeletal manifestations of interpersonal violence
among contemporary Greeks is presented here due mainly to the limited sample size
upon which such standards would be drawn. Assuming the data were adequate, a general
overview of the collection shows no patterns of trauma suggesting either accidental or
intentional mechanism. The nature of this essentially non-forensic collection an aged
population with mostly naturally occurring deaths precludes assessment of
interpersonal trauma patterns. Males in the collection do have significantly more facial
trauma than females (22% and 2%, respectively). Whether males are engaging more in
interpersonal aggression or are sustaining such injuries due to more physically active
interaction with the environment is unknown.
Suggestions For Further Research
Skeletal analyses on future remains received from the Zografou ossuary in Athens
will be helpful in better answering the hypotheses and questions proposed. It would be
very helpful to amass more detailed demographic information including the length of
time individuals actually resided in Athens (were they long term residents or recently
transported there for medical care?). It would also be helpful to attain supporting medical
documents in order to discover both the actual cause and manner of death. Any skeletal
fracture details included in the hospital documentation could be used to perform
postmortem analyses of fracture repair technique and outcome. If postmortem
radiographic studies could be performed, fracture-healing rates could be measured.
Metric evaluation of the long bone diaphysis might provide information on any
correlation between diaphyseal length and fracture occurrence. Lastly, three-dimensional
digitization of the crania and postcrania could assist with the development of population-
specific identification protocols.
ZOGRAFOU CEMETERY AND MAUSOLEUM, ATHENS
Primary burials maintained in Zografou Cemetery, looking northwest.
Photo by the author.
Figure A-2. Extreme eastern perimeter of Zografou Cemetery. A maintenance facility is
seen in the background. Photo by the author.
Figure A-3. One of two communal subterranean bone vaults on the southern perimeter of
Zografou Cemetery. Religious icons, incense containers and photographs of
deceased individuals are in the foreground. A skeletonized body still
wrapped in its burial shroud can be seen within the vault. Photo by the
Figure A-4. Close-up of a subterranean vault. Photo by the author.
Figure A-5. Interior of the Zografou Cemetery Mausoleum. After exhumation, most
individuals are secured in metal boxes and placed on shelves in this facility,
where family members can come to visit. Often, the family places pictures
of the deceased on the boxes. Religious icons, incense and burning candles
are seen in the right foreground. Photo by the author.
SAMPLE DOCUMENTATION FORMS
1 = 75-100% present (complete)
2 = <75% present (fragmentary)
3 = absent
Ribs Left Right
Carpal Phalanges Left
Ischiopubic ramus ridge
Greater sciatic notch
Estimated sex, pelvis
Left Mid Right
Estimated sex, skull
4th sternal rib
Estimated age: Very young adult (-18)
Young adult (20-35)
Middle adult (35-50)
Old adult (50+)
Lesser wing of sphenoid
Greater wing of sphenoid
Body of sphenoid
Immature Age Epiphyseal Union
Blank = unobservable
0 = open
1 = partial union
2 = complete union
Lateral to squama
Basiliar to squama
Skeletal Pathology Code Sheet
I. Bone: element
II. Exact: region of element
III. Side: R, L, R&L
Proximal epiphysis, intra-articular Middle 1/3 diaphysis
Proximal epiphysis, extra-articular Distal 1/3 diaphysis
Proximal 1/3 diaphysis Distal 2/3 diaphysis
Proximal 2/3 diaphysis Distal epiphysis, intra-articular
Middle 1/3 diaphysis Midshaft
Type: Complete Partial (Greenstick)
Bowed (note direction) Impaction (Compression)
Segmental (#) Avulsion
Burst Comminuted (#)
Depressed, outer table only
Depressed, inner and outer tables
Shape: Round Ellipsoidal
Edged (sft) Projectile (entry, exit, embedded)
Number of defects
Length, depth of each
Ante/Peri/Postmortem: Clear, Ambiguous
Woven callus Sclerotic
Healed Nonunion /pseudoarthrosis
Traumatic arthritis Ankylosis
Length (mm): Normal, Shortened
Rotation: Internal, External
Dislocation: Traumatic, Congenital, Ambiguous