<%BANNER%>

Is Aging Only Skin Deep?

Permanent Link: http://ufdc.ufl.edu/UFE0022072/00001

Material Information

Title: Is Aging Only Skin Deep? Assessing Change in Facial Bone Curvature with Age
Physical Description: 1 online resource (123 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: aging, bone, craniofacial, curvature, face, facial, maxilla, morphometrics, orbit, semilandmarks, shape, zygomatics
Anthropology -- Dissertations, Academic -- UF
Genre: Anthropology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The passage of time is etched on the human face as we transition from the round, indistinct visage of infancy into the individuality of adulthood and eventually the mature face of old age. As life expectancies have increased over the last century, this transition has become increasingly expansive. Incongruously, adult facial bone aging is a relatively under-investigated aspect of craniofacial research. This issue is apparent in the repetition of studies that have concentrated on the application of linear and angular measurement techniques on relatively homogenous populations (i.e., European/European-American) over the last 150 years. Given this focus, scholars have been unable to capture subtle changes in facial morphology with age in diverse human populations or visualize these changes within the three-dimensional space from which they were derived. As a counterpoint, my study of facial aging embraces three-dimensional geometric morphometric techniques. This allowed me to quantify and visualize differences in facial bone curvature among assorted human age groups. I hypothesized that significant differences in facial bone shape exist between age groups globally (entire facial skeleton), regionally (eye orbits, zygomatic arches, nasal aperture, maxillary ridge) and locally (e.g., curves representing superior orbital rims). Furthermore, I hypothesized that these age-related shape changes manifest differently in different human populations. In the course of data collection, I partitioned the face into structural curves. This allowed curves constituting such areas as the eye orbits to be analyzed both in isolation from, and in conjunction with, other areas such as the zygomatic arches. I collected curvature data in the form of three-dimensional semilandmarks from nearly 700 African-American and European-American adult crania. These socially-determined racial classifications were then sub-divided into three age groups (i.e., young adult, middle-aged, and elderly). I transformed the semilandmark data into a common coordinate system and used principal component analysis to reduce the dimensionality. I then performed various multivariate statistical procedures to evaluate the presence and nature of interactions between age, sex, and race with respect to facial shape. Finally, I visualized these age-related shape differences in facial bone curvature using vector plots. In support of my hypotheses, the results reveal a complex pattern of curvature change wherein the facial skeleton displays age-related shape differences as a whole, regionally, and locally, as well as age*race or age*race*sex interactions at these various levels. Visually, these differences may appear as superoinferior compression or expansion, lateral expansion or recession, mediolateral compression, or anterior recession depending on the region analyzed and the age groups compared. Thus, my findings indicate that facial bone curvature alters with age at different levels and these changes vary across human populations. My results, then, not only help create an integrated model of facial aging, but emphasize the importance of considering social and biological factors such as race and sex in such analyses.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Falsetti, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022072:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022072/00001

Material Information

Title: Is Aging Only Skin Deep? Assessing Change in Facial Bone Curvature with Age
Physical Description: 1 online resource (123 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: aging, bone, craniofacial, curvature, face, facial, maxilla, morphometrics, orbit, semilandmarks, shape, zygomatics
Anthropology -- Dissertations, Academic -- UF
Genre: Anthropology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The passage of time is etched on the human face as we transition from the round, indistinct visage of infancy into the individuality of adulthood and eventually the mature face of old age. As life expectancies have increased over the last century, this transition has become increasingly expansive. Incongruously, adult facial bone aging is a relatively under-investigated aspect of craniofacial research. This issue is apparent in the repetition of studies that have concentrated on the application of linear and angular measurement techniques on relatively homogenous populations (i.e., European/European-American) over the last 150 years. Given this focus, scholars have been unable to capture subtle changes in facial morphology with age in diverse human populations or visualize these changes within the three-dimensional space from which they were derived. As a counterpoint, my study of facial aging embraces three-dimensional geometric morphometric techniques. This allowed me to quantify and visualize differences in facial bone curvature among assorted human age groups. I hypothesized that significant differences in facial bone shape exist between age groups globally (entire facial skeleton), regionally (eye orbits, zygomatic arches, nasal aperture, maxillary ridge) and locally (e.g., curves representing superior orbital rims). Furthermore, I hypothesized that these age-related shape changes manifest differently in different human populations. In the course of data collection, I partitioned the face into structural curves. This allowed curves constituting such areas as the eye orbits to be analyzed both in isolation from, and in conjunction with, other areas such as the zygomatic arches. I collected curvature data in the form of three-dimensional semilandmarks from nearly 700 African-American and European-American adult crania. These socially-determined racial classifications were then sub-divided into three age groups (i.e., young adult, middle-aged, and elderly). I transformed the semilandmark data into a common coordinate system and used principal component analysis to reduce the dimensionality. I then performed various multivariate statistical procedures to evaluate the presence and nature of interactions between age, sex, and race with respect to facial shape. Finally, I visualized these age-related shape differences in facial bone curvature using vector plots. In support of my hypotheses, the results reveal a complex pattern of curvature change wherein the facial skeleton displays age-related shape differences as a whole, regionally, and locally, as well as age*race or age*race*sex interactions at these various levels. Visually, these differences may appear as superoinferior compression or expansion, lateral expansion or recession, mediolateral compression, or anterior recession depending on the region analyzed and the age groups compared. Thus, my findings indicate that facial bone curvature alters with age at different levels and these changes vary across human populations. My results, then, not only help create an integrated model of facial aging, but emphasize the importance of considering social and biological factors such as race and sex in such analyses.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Falsetti, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022072:00001


This item has the following downloads:


Full Text

PAGE 1

IS AGING ONLY SKIN DEEP?: ASSESSING CHANGE IN FACIAL BONE CURVATURE WITH AGE By SHANNA E. WILLIAMS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Shanna E. Williams 2

PAGE 3

To all who encouraged me to strive for more, to push the boundaries, and most importantly to never settle 3

PAGE 4

ACKNOWLEDGMENTS I thank my committee members, Anthony Falsetti (chair), Thomas Hollinger, Dennis Slice, David Daegling, and Michael Warren, for making time in their busy schedules to offer me support and guidance. I am al so grateful to David Hunt, Lyman Jellema, Lee Jantz, and Heather Edgar, whose permission, patience, and assistance in making skeletal material accessible to me made this endeavor possible. Thank you also to the Ford Foundation for its generous support. A special thank you goes to friend and co-conspirator, Alana Lynch, for her critical eye for editing, generous ear of listeni ng, and creative mind of ideas. I sincerely thank my family. Their constant support and guidan ce kept me moving forward, even when I wanted desperately to stay planted in one spot. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.......................................................................................................................10 LIST OF ABBREVIATIONS........................................................................................................11 ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 2 BONE......................................................................................................................... ............19 Bone Structure........................................................................................................................19 Osteogenesis...........................................................................................................................21 Age-Related Morphological Changes to Bone.......................................................................21 Microcosm to Macrocosm Bone Change...............................................................................23 3 METRIC CRANIOFACIAL AGING LI TERATURE: REVIEW AND CRITQUE.............24 Cross-Sectional Studies..........................................................................................................25 Longitudinal Studies........................................................................................................... ....32 Literature Critique..................................................................................................................37 Population Dynamics.......................................................................................................37 Methodological Constraints.............................................................................................38 4 GEOMETRIC MORPHOMETRICS......................................................................................40 Traditional Mo rphometrics.....................................................................................................4 1 Geometric Mo rphometrics......................................................................................................42 Ridge Curve Analysis.............................................................................................................43 Equidistant Semilandmarks.............................................................................................44 Sliding Semilandmarks....................................................................................................46 5 MATERIALS AND METHODS...........................................................................................48 Skeletal Collections................................................................................................................48 Robert J. Terry Anatomical Collection...........................................................................48 Hamann-Todd Collection................................................................................................49 Maxwell Museums Documented Skeletal Collection....................................................50 William M. Bass Donated Skeletal Collection................................................................51 5

PAGE 6

Inclusion Criteria....................................................................................................................51 Cranial Regions and Age Groups...........................................................................................53 Hypotheses..............................................................................................................................54 Data Collection Procedure......................................................................................................55 Special Curve Considerations.................................................................................................55 Alveolar Curve................................................................................................................56 Nasal Curve.....................................................................................................................56 Superior Orbital Curve....................................................................................................56 Inferior Orbital Curve......................................................................................................57 Outer Orbital, Superior Temporal, Infe rior Temporal, and Inferior Zygomatic Curves..........................................................................................................................57 Intra-Observer Error...............................................................................................................57 Data Manipulation and Analysis............................................................................................57 Semilandmark Extraction................................................................................................57 Multivariate Statis tical Analysis......................................................................................58 6 RESULTS...................................................................................................................... .........64 Preliminary Data Analysis...................................................................................................... 64 Pooling the Data..............................................................................................................6 4 Dental State in the Elderly Age Group............................................................................65 Main Data AnalysisMANOVA............................................................................................66 Global Effects..................................................................................................................66 Regional Effects..............................................................................................................67 Local Effects....................................................................................................................67 Main Data AnalysisPost-Hoc Tests......................................................................................68 Eye Orbits..................................................................................................................... ...68 Zygomatic Arches...........................................................................................................71 Maxillary Ridge...............................................................................................................7 2 Shape Patterns in the Facial Curvature...................................................................................73 7 DISCUSSION................................................................................................................... ......87 Level of Analysis.............................................................................................................. ......88 Racial Differences in Faci al Morphology with Age...............................................................89 Visualizing Regional Shape Differences with Age................................................................91 Functional Matrix Hypothesis................................................................................................92 Skeletal Units...................................................................................................................93 Functional Matrices.........................................................................................................93 Loading as an Epigenetic Factor.....................................................................................94 Bone Responsiveness to Loading....................................................................................96 Aging Muscle and the Functional Matrix Hypothesis.....................................................98 Other Contributing Functional Matrices.........................................................................99 Absence of Nasal Aperture Change......................................................................................100 Limitations.................................................................................................................... ........101 8 CONCLUSION................................................................................................................... ..102 6

PAGE 7

7 APPENDIX A AGE GROUP POPULATION DEMOGRAPHICS.............................................................108 B LANDMARK DEFINTIONS...............................................................................................109 LIST OF REFERENCES.............................................................................................................110 BIOGRAPHICAL SKETCH.......................................................................................................123

PAGE 8

LIST OF TABLES Table page 5-1 Curve definitions.......................................................................................................... ......62 5-2 Number of semilandmarks extracted for each curve.........................................................63 6-1 MANOVA output for the evaluation of collection effect..................................................75 6-2 MANOVA output for the evalua tion of dental state in th e eye orbits and zygomatics arches......................................................................................................................... ........75 6-3 MANOVA output for the evaluation of dental state in the eye orbits...............................75 6-4 MANOVA output for the evaluation of de ntal state in the zygomatic arches...................75 6-5 MANOVA output for the entire face.................................................................................76 6-6 MANOVA output for the eye orbits..................................................................................76 6-7 MANOVA output for the zygomatic arches......................................................................76 6-8 MANOVA output for the maxillary arch...........................................................................76 6-9 MANOVA output for the nasal aperture...........................................................................77 6-10 MANOVA output for the superior orbital rim...................................................................77 6-11 MANOVA output for the inferior orbital rim....................................................................77 6-12 MANOVA output for the outer orbital curve of the zygomatic arches.............................77 6-13 MANOVA output for the superior temporal curve of the zygomatic arches....................78 6-14 MANOVA output for the in ferior zygomatic curve of the zygomatic arches...................78 6-15 MANOVA output for the in ferior temporal curve of the zygomatic arches......................78 6-16 Age group contrast tests for Af rican-American female eye orbits....................................78 6-17 Age group contrast tests for Af rican-American male eye orbits.......................................79 6-18 Age group contrast tests for Eu ropean-American male eye orbits....................................80 6-19 Age group contrast tests for Eur opean-American female eye orbits.................................81 6-20 Age group contrast tests for Eu ropean-American zygomatic arches.................................83 8

PAGE 9

6-21 Age group contrast tests for Af rican-American zygomatic arches....................................83 6-22 Age group contrast tests for Af rican-American maxillary ridge.......................................85 6-23 Age group contrast tests for Eu ropean-American maxillary ridge....................................85 A-1 African-American age groups..........................................................................................108 A-2 European-American age groups.......................................................................................108 B-1 Landmark definitions....................................................................................................... 109 9

PAGE 10

LIST OF FIGURES Figure page 5-1 Curve tracings of the face................................................................................................. .62 5-2 Extracted semilandmarks for individual with all curves present.......................................63 6-1 Canonical variates pl ot of the African-American male eye orbits.....................................79 6-2 Vector plot of African-Ame rican male eye orbit: young to elderly vectors of change.....79 6-3 Canonical variates pl ot of the European-American male eye orbits..................................80 6-4 Vector plot of European-American male eye orbit: young to middle-aged vectors of change................................................................................................................................80 6-5 Vector plot of European-American ma le eye orbit: young to elderly vectors of change................................................................................................................................81 6-6 Canonical variates plot of the European-American female eye orbits..............................81 6-7 Vector plot of European-American fe male eye orbit: young to middle-aged vectors of change............................................................................................................................82 6-8 Vector plot of European-American female eye orbit: middle-aged to elderly vectors of change............................................................................................................................82 6-8 Canonical variates pl ot of the African-American zygomatic arches.................................83 6-9 Vector plot of African-American z ygomatic arch: young to elderly vectors of change................................................................................................................................84 6-10 Vector plot of African-A merican zygomatic arch: middle-aged to elderly vectors of change................................................................................................................................84 6-11 Canonical variates plot of the European-American maxillary ridge.................................85 6-12 Vector plot of European-American maxillary ridge: young to middle-aged vectors of change................................................................................................................................86 6-13 Vector plot of European-American maxilla ry ridge: middle-aged to elderly vectors of change................................................................................................................................86 10

PAGE 11

LIST OF ABBREVIATIONS CVA canonical variates analysis DF: degrees of freedom BMU: basic multicellular units FMH: functional matrix hypothesis GM: geometric morphometrics GPA: Generalized Procrustes analysis MANOVA: multiple analysis of variation PC: principal component PCA: principal component analysis 11

PAGE 12

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IS AGING ONLY SKIN DEEP?: ASSESSING CHANGE IN FACIAL BONE CURVATURE WITH AGE By Shanna E. Williams May 2008 Chair: Anthony B. Falsetti Major: Anthropology The passage of time is etched on the human f ace as we transition from the round, indistinct visage of infancy into the individuality of adulth ood and eventually the mature face of old age. As life expectancies have increas ed over the last century, this tr ansition has become increasingly expansive. Incongruously, adult facial bone agin g is a relatively under-in vestigated aspect of craniofacial research. This issue is apparent in the re petition of studies that have concentrated on the application of linear and angular meas urement techniques on relatively homogenous populations (i.e., European/European-American) ove r the last 150 years. Given this focus, scholars have been unable to capture subtle changes in facial morphology with age in diverse human populations or visualize these changes wi thin the three-dimensional space from which they were derived. As a counterpoint, my study of facial ag ing embraces three-dimensional geometric morphometric techniques. This allowed me to quantify and visual ize differences in facial bone curvature among assorted human age groups. I hypothe sized that significant differences in facial bone shape exist between age groups globally (entire faci al skeleton), regionally (eye orbits, zygomatic arches, nasal aperture, maxillary ridge) and locally (e.g., curves representing superior 12

PAGE 13

orbital rims). Furthermore, I hypothesized th at these age-related shape changes manifest differently in different human populations. In the course of data collection, I partitioned the face into structural curves. This allowed curves constituting such areas as the eye orbits to be analyzed both in isolation from, and in conjunction with, other areas such as the zygomatic arches. I co llected curvature data in the form of three-dimensional semilandmarks fr om nearly 700 African-American and EuropeanAmerican adult crania. These socially-determine d racial classifications were then sub-divided into three age groups (i.e., young adult, middl e-aged, and elderly). I transformed the semilandmark data into a common coordinate system and used pr incipal component analysis to reduce the dimensionality. I then performed va rious multivariate statistical procedures to evaluate the presence and nature of interactions between age, sex, and race with respect to facial shape. Finally, I visualized th ese age-related shape differences in facial bone curvature using vector plots. In support of my hypotheses, the results reve al a complex pattern of curvature change wherein the facial skeleton displays age-relate d shape differences as a whole, regionally, and locally, as well as age*race or age*race*sex interac tions at these various le vels. Visually, these differences may appear as s uperoinferior compression or ex pansion, lateral expansion or recession, mediolateral compressi on, or anterior recession depending on the region analyzed and the age groups compared. Thus, my findings indicat e that facial bone curvat ure alters with age at different levels and these changes vary acro ss human populations. My results, then, not only help create an integrated model of facial ag ing, but emphasize the impor tance of considering social and biological factors such as race and sex in such analyses. 13

PAGE 14

CHAPTER 1 INTRODUCTION The human face serves as the main source of our sense of identity, personality, and perception of beauty. It can denote honesty or duplicity, arrogance or humility, health or sickness. This, in turn, has leant to such axioms as saving face, a face that only a mother could love, and two-faced. Moreover, the face can be temporarily altered th rough the application of cosmetics or permanently via plastic surgery. In both cases, we manipulate the face in order to present a specific visage. Our f aces, then, are the medium through which we often interact with others and the world around us. The tremendous social implications placed upon the human face rely heavily upon its topological complexity. This complexity is visu ally constructed by soft-tissue curvilinear forms such as the forehead, cheeks, eyes and jaw line; all of which alter with age. Over time, the smooth appearance of the youth is replaced with deepening lines of demarcation (e.g., crows-feet and laugh-lines) as the soft tissues of the face descend. However, despite these obvious visual and by consequence social cues, it is important to remember that the face as a whole is not simply the reflection of soft tissue interactions. Instead, th is soft tissue rests upon complex curvilinear bone, which directly influences its shape and position. Thus, biologically the face is constructed by both the hard and soft tissues. And, as interrelated entities, over time internal hard tissue modification may impact the manifestation of external soft tissue change. Yet, alterations in craniofacial bone structure with age are not as well understood as those witnessed on the soft tissue level. An appreciati on of the effect(s) of advancing age on adult bone in this region, however could dramatically impact the way(s) in which craniofacial surgery is approached among the middle-aged and elderly, as well as provide insight into the evolution of human facial form in different populations. We should ask 14

PAGE 15

ourselves, then, is aging only skin deep, or does alteration in bony architec ture also serve as a source of change in the face we present to the world. Several studies within the fi elds of biological anthropology, orthodontics, and craniofacial surgery have tangentially explored this skin de ep concept via an examination of macroscopic differences in craniofacial dimensions. These st udies have noted that over time many of the bones of the face display increased thickness, lengt h, and breadth (Chapter 3). Moreover, it has been suggested that the cranio facial skeleton remodels over tim e via a process of differential growth (e.g., Hellman, 1927; Bartlett et al ., 1992; Pessa, 2000; Shaw and Kahn, 2007). In general, however, these results rely h eavily upon techniques (e.g., craniometry and roentgenographic cephalometry) which suffer fr om certain methodological constraints. Customarily, invariant numerical values, such as measured lengths between landmarks or angles between pairs of such lines, ar e incapable of characterizing regi ons exhibiting complex curvature and cannot fully account for the three-dimens ional world in which the human body exists. Furthermore, documented structur al differences are often treate d as being reflective of some universal human condition, despite being derived from a specific subset of humanity (i.e., generally European/European-American). This bi ased level of analysis and its generalized application presumes that universality is embedded in the European/European-American experience and thus overlooks fine-grain phe nomenon which may signify change in other ancestral populations. As a result, many of the existing studies merely re-affirm knowledge that has existed for over 150 years. Thus, age-related a lteration in cranial contour, particularly in the facial skeleton, remains a relative mystery. In order to better address the three-dime nsional interaction between and among the curvilinear structures of the facial skeleton with age, I have taken a multi-tiered approach. First, 15

PAGE 16

I garnered a sample size of nearly 700 indivi duals, embodying a mix of socially-determined races and biologically-determined sexes. The in dividuals, originating from several skeletal collections, are representative of modern cran ia from Europeanand African-Americans born in the 19th and 20th centuries. By utilizing such a large a nd diverse research sample, I believe I can adequately capture morphologica l variation over time, while si multaneously evaluating this variation in terms of sex and race. Furthermore, each racial population is parsed into three overarching age groups that roughly correspond to soci etal norms (i.e., young adult, middle-age, and elderly) containing approximately equal numbers of males and females. By partitioning adulthood into three distinct phases, I am able to more fully capture the spectrum of human facial aging than many earlier studies. In the course of my study, I used a nove l geometric morphometric technique known as ridge curve analysis. Rather than simply meas uring unconnected, arbitrary points, as is often done, this method utilizes semilandm arks to record and analyze th e curvature of a structure. Ridge curve analysis is a particularly ideal tech nique by which to measure the facial skeleton as this region is a composite of nu merous contours and curvilinear fo rms. Together, these curves constitute the arches of, for example, the eye orbits, cheekbones, nasal aperture, and maxillary ridge. Employing a specialized digital caliper, I captured the three-dimensional curvature of the facial skeleton, in term s of these regions. Then, using various geometric morphometric procedures and multivariate statistics, I assessed the presence and nature of change am ongst and between these curves in disparate populations and the sexes over time. Most interesting, I discovered global, regional, and local age-related changes which intera cted significantly with race and sex (Chapter 6). Thus, this large sample size, group partitioning, and analytic technique allows me to better mathematically 16

PAGE 17

quantify potential shape changes in th e facial skeleton that can be so cially described as signs of age. Aging is a naturally occurring, post-developmental phenomenon (Kirkwood, 1995). Most simply put, it is the process of becoming olde r; a consequence of existing through time. However unlike senescence, aging does not direc tly increase the probability of death (Crews, 2003). Consequently, this study focu ses solely on aging rather than senes cence. As such, I seek to characterize facial skeleton shape as a neut ral reflection of alterati on over time. This is a natural process to be understood within the context of the human condition. Hence, my goal is to examine external patterns of adult life history (p ost-development to death) in the facial skeleton and the possible extrinsic and intrinsic factors which may account for documented change. Chapter organization. I set the stage for this research in Chapter 2 which explores the complex structure and properties of bone. Cellula r alterations in bone composition with growth (i.e., development) and age are discussed micr oscopically. Meanwhile, adult gross metric alterations with age are documented in Chapter 3. This chapter serves not just as literature review of age-related res earch in the craniofacial skeleton, but also as an exploration of apparent scientific disinterest in this subject matter. Moreover, I try to tease apart methodological shortcomings and conceptual biases which unde rlie and constrain the field of age-related craniofacial research. A brief history of the field of geometric mor phometrics as it relates to biological variation is provided in the opening of Chapter 4. I then focus specifically on ridge curve analysis. This section explores the nature of semilandmarks in terms of benefits and potential limitations. I describe my materials and outline my methods in the Chapter 5. I begin th e chapter by exploring the demographics and history of the skeletal coll ections I used in this project, as well as my 17

PAGE 18

18 research material criteria. I then present my research hypotheses and data collection procedure. Chapter 5 ends by discussing how I went a bout extracting the semilandmarks and what multivariate statistical techniques were performed to address the issue of age-related change. Results are provided in Chapter 6 where I have attempted to inco rporate both numeric and visual depictions of the complex changes discovere d throughout the facial skeleton. I present a discussion of these results in Ch apter 7 in terms of the dynamic in teractions between genetic and epigenetic factors. Finally, I conclude this research in Chap ter 8 by summarizing my findings and positing the directions in which this research can be expanded.

PAGE 19

CHAPTER 2 BONE Bone is a highly organized, specialized tissu e which performs a myriad of structural, mechanical, and biochemical functions. It prov ides mechanical support for the soft tissues, levers for muscle action, and serves as a stor ehouse for calcium. Bone also protects vital structures such as the brain, heart, and lungs. Furthermore, it houses the bone marrow, which provides hemopoietic cells in th e adult (Steinbock, 1976). These versatile properties stem from bones unique characteristics. It is only by unde rstanding bones complex structure that can we begin to comprehend how it alters with age. Bone Structure Befitting its complex nature, bone consists of both a matrix of calcified intercellular ground substance and three types of living cells. Of these thre e cell types, osteoblasts are responsible for synthesizing the organic com ponents of the bone matrix and regulate its mineralization. Over time, osteoblasts become embedded in tiny spaces (lacunae) within the surrounding matrix. At this point they are referred to as a second type of bone cell, the osteocyte. Small cylindrical canals (canaliculi) connect the lacunae with one another, allowing communication between neighboring osteocytes and circulation of tissue fluids (Junqueria and Carniero, 2003). Osteocytes ar e also responsible for mainta ining calcium homeostasis via osteocytic osteolysis, which releases calcium by re sorption of the lacunar wall. The third type of bone cell is the osteoclast. These large, multinuc leated cells resorb excess or inferior bone matrix. Osteoclasts form and are located in concavities known as Howship's lacunae, which represent areas of bone resorpti on. Meanwhile, the matrix com ponent of bone is primarily a composite of two types of materials: type I collagen and hydroxyapatite. Type I collagen constitutes most of the organic po rtion of the matrix and gives bone its elasticity. The collagen 19

PAGE 20

matrix is impregnated by hydroxyapatite crystals which serve as a dense inorganic filling. This gives bone its hardness and rigidity (White, 2000; Junque ria and Carniero, 2003). Histologically, two distinct patt erns can be observed in bone st ructure. In immature (or woven) bone, the collagenous fiber bundles are rela tively coarse, interweavi ng in a non-oriented, random pattern. Meanwhile, the collagen fibers of mature, or lamellar, bone run more or less parallel to one another and are arranged in or derly sheets called lamellae. Loosely structured lamellae, called trabeculae, make up the articul ar ends and interior of most bones (Steinbock, 1976; Junqueria and Carniero, 2003). Trabecular tissue is permeated with vesicular spaces containing bone marrow and blood vessels from which it receives nourishment. On the other hand, dense layers of lamellae, known as compact bone, characterize the outer layer of most bones and provide mechanical strength. Given its dense nature, compact tissue cannot be nourished by surface blood vessel diffusion and inst ead relies on specialized vascular channels known as Haversian systems which contain hollow canals through which blood, lymph, and nerve fibers can pass. Molecularly and cellula rly identical, the only difference between compact and trabecular tissue lies in their porosity (S teinbock, 1976; Junqueria and Carniero, 2003). The internal and external surface of bone is covered by connective tissues and bone forming cells. Periosteum is a dense connective tissue membrane that i nvests the outer surface of bone. Consisting of two layers, the outermost layer of periosteum is made up of collagen and fibroblasts, while the inner laye r contains osteoprogenitor cell capable of differentiating into osteoblasts. The internal surface of bone, meanwhile, is lined by endosteum, which is also composed of osteoprogenitor cel ls and a small amount of c onnective tissue (Junqueria and Carniero, 2003). 20

PAGE 21

Osteogenesis Embryologically, bone formation occurs via two different processes: intramembranous and endochondral ossification. In both processes, preexisting connectiv e tissue is replaced by bone. In intramembranous ossification, membrane is transformed into bone. Specifically, bone is developed from mesenchymal condens ation which gives rise to os teoblasts. Small groups of these osteoblasts synthesize non-mineralized matr ix, or osteoid, which subsequently undergoes calcification. Flat bones, such as the clavicle and most of the skull ossify by this process (Steinbock, 1976; Junqueria and Carniero, 2003). However, in endochondral ossification a hyaline cartilage model of the bone is ossified. Occurring primarily in the long and short bones of the skeleton, mesenchymal cells differentiate into chondroblasts which form a hyaline cartilage model roughly correspo nding to the shape of the bone in the adult skeleton. Chondrocytes w ithin the center of this cartilaginous model undergo hypertrophy and the cartilage matrix is resorbed (Steinbock, 1976; White, 2000; Junqueria and Carniero, 2003). The dying c hondrocytes are then invaded by primitive mesenchymal cells and blood vessels, which diffe rentiate into osteoblasts and blood-forming cells of bone marrow, thus forming a primary cente r of ossification. Later, secondary centers of ossification appear in the cartilag inous epiphyses. These centers continue to expand until only a thin plate of hyaline cartilage separates the epiphysis from the diaphysis. Known as the epiphyseal plate, continued cartilag e growth and bone replacement in this region allows for rapid longitudinal growth until skeletal maturity is attained (Ste inbock, 1976; White, 2000; Junqueria and Carniero, 2003). Age-Related Morphological Changes to Bone Following skeletal maturity, bone is far from static. Instead, stru ctural integrity is maintained by bone remodeling which constantly removes old bone and synthesizes new bone in 21

PAGE 22

its place. Approximately 5-10% of the adult huma n skeleton is replaced annually as a result of normal bone remodeling (Parfitt, 1982). This process is accomplished by the juxtaposition of osteoblasts and osteoclasts into discrete temporar y anatomic structures called basic multicellular units (BMUs). As the lifespan of these cells is relatively short compared to that of the BMU, osteoblasts and osteoclasts must be constantly replenished for normal bone homeostasis to be maintained (Chan and Duque, 2002; Jilka, 2003). As a person advances in age, the orderly balance of this system d eclines, leading to changes in bone architecture. With advancing age, bone becomes more porous; the number of empty osteocyte lacunae increases; and canaliculi and Haversian canals become obstructed (Chan and Duque, 2002). The sixth decade is marked by diminished osteoblast activ ity and number. This, in turn, is associated with an overall reduction in active osteoid su rfaces (percentage of to tal trabecular surfaces covered with osteoid with a borde r of osteoblasts) (Sharpe, 1979). Osteocyte density in the deep layers of trabecular bone also de creases with age (Qiu et al., 2002). Osteoclast function remains relatively stable in men with age. In contrast, postmenopausal women exhibit variation in osteoclast function, which is characterized by early postmenopausal increase followed by decrease in the succeeding years (Clarke et al., 1996). Both sexes reach peak bone mass around the thir d decade of life and then decrease in the following years (Kanis et al., 1997; Russo et al ., 2003). However, women begin life with substantially less bone th an men (Takahashi and Frost, 1966) Computer tomographic scans indicate that comparatively young women have 25-33% less bone cross-sectional area and 1821% less bone mass in their long bones than you ng men (Riggs et al., 2004). Trabecular bone loss begins in young adulthood and proceeds linearly throughout life in both sexes. This early bone loss accounts for between 1/3 and 1/2 of total trabecular bone loss over life in both sexes. 22

PAGE 23

23 However, substantial sex differences exist for co rtical bone loss over time. In women, cortical bone density does not begin to decrease until mid-life in association with menopause. Meanwhile, for men, a small amount of cortical b one loss begins in early adulthood. This bone loss progresses at a constant ra te until around 75 years after which point it ac celerates (Riggs et al., 2004; 2008). Increased localized stresses at the hard tissue leve l are also associated with agerelated bone loss and have been im plicated in the diminishment of bones elastic properties (i.e., toughness and strength) and increas ed fragility (Currey et al., 1996; Diab et al., 2006; Zioupos and Currey, 1998). Microcosm to Macrocosm Bone Change Bone, like any biological structure, is arranged hierarchically. Structural and functional complexity increases upward; from the subatomic particle to molecules, and on to cells, tissues, organs, and organisms. As this complexity in creases, new attributes manifest which are not present at the previous level (Moss, 1997a; Ruff, 2006). Despite the coalescence of building materials to create a complex stru cture, the presence or manifestation of these new attributes can not be predicted from an examination of the building materials, alone (Ruff, 2006). For instance, understanding the key attributes of an atom doe s not promote knowledge of a leaf, much less a tree. As such, no amount of understanding regard ing cellular bone aging can foretell how it will manifest macroscopically. Instead, bone cell biology can best be employed as a way of explaining an already established macroscopic phenomenon. Thus, we are left to our own devices to first establish gross sh ape change in the facial skeleton with age before we can attempt to exploit any aspect of bone cell aging to explain it.

PAGE 24

CHAPTER 3 METRIC CRANIOFACIAL AGING LITER ATURE: REVIEW AND CRITQUE A theme that commonly surfaces within metric studies of aging in the craniofacial skeleton is the suggestion that research on this subject is scarce (e.g., West and McNamara, 1999; Akgul and Toygar, 2002). In truth, such claims may be over-exaggerated, as dozens of studies do exist wherein age-related craniofacial change is explored. Althou gh authors claiming otherwise may be engaging in a bit of poetic hyperbole, there is a noticeable imbalance between craniofacial research dedicated to bone growth as compared to that dedicated to adult skeletal aging. For instance, a search for craniofacial growth on a ma jor scientific search engine such as Pubmed (www.pubmed.gov) produces 4032 article s, while a search for cranio facial aging produces a mere 169 (both as of Jan 2008). Additionally, these results do not account for issues such as level of analysis (microscopic versus macroscopic), material under investigation (hard tissue versus soft tissue), or the model em ployed (human versus nonhuman). In the case of clinical populations, I suspect that there are a multitude of reasons which contribute to this discrepancy. From a purely practical perspective, access to children is relatively convenient (i.e., schools, hospitals, etc.), while adults te nd to be more mobile and thus too scattered for easily manageable long-term inve stigation. Moreover, adu lts, and in particular the elderly, often only visit hospita ls in instances of severe illn ess, potentially making them poor research candidates. Also, I imagine many find th e study of children and adolescents to be more emotionally and professionally rewarding than studying the physical consequences of aging. This, in turn, may produce biased funding opportunities for growth rather than aging research. Another explanation for the appare nt lack of interest in cranio facial skeletal change with age is the suggestion that bone alteration e ffectively ceases upon complete bone fusion (e.g., Ranly, 1988), or is so inconsequen tial as to not warrant further i nvestigation. Indeed, in several 24

PAGE 25

recent craniofacial aging analyses (e.g., Akgul and Toygar, 2002), the idea of bone stability following skeletal maturity is cited as the basis fo r general disinterest in such research. I suspect that our own social biases play a promin ent role in this particular stance. As mentioned in the introduction, our world is shaped by perspective. We see the passage of time in our mirrors and photographs as our sma ll, indistinct childhood faces transform into the mature individuality of adulthood. This relatively rapid process is marked by dramatic shifts in shape and size. Yet with adulthood, the face appear s, for the most part, to hold its size, shape, and functional capacity; only altering in skin position and quality. As such, our eyes and by consequence our minds, are tuned to perceive this region, at least in terms of bone, as relatively static. This ingrained perspective, though, is refute d by a specific set of craniofacial studies dealing specifically with bone Although the following review is not cumulative, it is representative of an extensive revi ew of the craniofacial literature in terms of age-related skeletal change relative to my study. This research is partitioned into cross-s ectional and longitudinal studies and I have attempted to highlight ma jor findings, methods, and populations utilized. Cross-Sectional Studies In 1858, Humphrey observed that the European cranium became thicker with age and in 1865 he also described widening of the gonial angle as a conseque nce of age-related tooth loss. Some years later, Pfitzner (1899) reported augm entation in the head breadth (euryon to euryon) and bizygomatic breadth, (i.e., face width -zygio n to zygion) throughout life in an Alsatian population. However, these results we re not statistically significant. The 20th century was marked by further craniofaci al revelations. In 1924, Todd examined 484 European-American male skulls ranging in ag e from 20-84 years and noted an increase in cranial thickness up to the seventh decade with no discernible changes thereafter. Hellman 25

PAGE 26

(1927, 1932) expanded upon this research with his examination of an American Indian skull series spanning adolescence to old age. He foun d continued facial height (nasion to gnathion) and facial width expansion until old age, at which time degeneration ensued. Moreover, Hellman (1927) suggested that different craniofacial regions grew at vary ing rates. However, these results were met with staunch criticism and as Israel (1973a) noted in his review of the craniofacial aging literature, one respondent to Hellmans work proclaimed, I am practically convinced (of the) fallacythat the face goes on growing until old age (176). Nonetheless, similar findings were documented in subseque nt research. For instance, Saller (1930, 1931) observed significant cranial augmentation in term s of bizygomatic and head breadth, as well as head length (glabella to opi sthocranion) in a European popul ation. Similarly, Jarcho (1935) noted head size gain (i.e., head length, facial le ngth and width) into the fourth and possibly sixth decade for a Russian, Uzbek, Kirghiz, and Armenian sample. Meanwhile, in 1936 Hrdlicka concluded that cr aniofacial growth extends beyond the early twenties, but at a much slower rate. Summariz ing nearly 40 years of re search based on his own work (1925, 1935) involving American Indians and old American whites, as well as those of his predecessors (e.g., Pfitzer, 1899; Jarcho, 1935 etc.) Hrdlichka beli eved skull and face diameter continues to grow into the fifth or sixth decade. However, dissenting opinions of craniofacial change were presented by Goldst ein (1936) and deFroe (1936). For instance, Goldstein (1936) concluded from an anthropometri c investigation of Ame rican Jews that there is actually a slight reduction in the average size of all head diam eters with age. deFroe (1936) also noted diminished facial he ight in a European series of 205 male and 90 female skulls and suggested the human head actually got smaller over time. 26

PAGE 27

Utilizing lateral head films spanning seve n decades, Oldberg (1946) found no dramatic change in frontal and pariet al bone thickness after the age of 25 in European woman. Additionally, Randalls ( 1949) study of 17,341 young Eu ropean-American male soldiers (grouped 17-18 years and 19-26 year s) in the U.S. Army noted no change in head circumference between the two age groups. M eanwhile, Hooten and Dupertuis ( 1951) reported continued facial growth into the sixth decade for a sample of approximately 10,000 Irishman. Utilizing anthropometric data, they detail ed increases in head length an d breadth, bizygomatic dimension, head circumference, and total f acial height; all of which extended well into the third decade (Hooten and Dupertuis, 1951). A facial height study of 572 European me n and 430 women aged 20-91 years by Wunsche (1953) found among males a 1.1mm increase in total facial height from 20-40 years, and a 2.4 mm decrease in this parameter after age 60. Wo men, on the other hand, displayed the greatest facial height at 26-30 years, while facial height was marked by a 7.7mm decr ease after 71 years. Based on an adult male and female Mexican po pulation, Lasker (1953) noted older age groups displayed larger average facial measurements (i.e., bizygomatic and bi gonial diameters, and facial height). Addressing change in the cranial base, Zuckerman (1955) documented an increase in absolute distances, yet little alteration in angular values among a series of 109 skulls wherein no attempt was made to determine sex or race. In his investigations of hyperostosis cranii in skeletal collections, Moore (1955) found no trend toward craniofacial enlargement after the second decade, despite the suggestion of c ontinued cranial expans ion depicted in his published graphs. In 1957, Tallgren examined facial height in 165 European women aged 20-81 years using cephalometric methods. She found facial hei ght increased throughout the decades, with the 27

PAGE 28

highest rate of change in th e third decade. Horowitz and Thompson (1964) also noted an increase in facial components with age, as well as sex differences among a European-American population with Class I occlusion (i .e., dental normality). Furthermore, Heath (1966) considered the relationship between dental state and craniofacial cha nge in 166 European lateral cephalometric radiographs, and found no correla tion between upper faci al dimensions and number of years a patient had been edentulous In 1969, Israel examined the morphology of the mandible relative to age in a European-American male dentate populati on ranging from 20 to 69 years. He found trends suggestive of increased body thickness and alveol ar crest loss with age among this population. Boersma (1974) described significant enlargement of cranial base angle and nasal floor in an older Dutch population of 150 individuals (a ged 58-91 years). Only minor changes in craniofacial dimension were reported; however, dentition state (i.e dentulous vs edentulous) was not noted. Once again, total facial height was found to increase with age in a study conducted by Nasjletti and Kowalski (1975) examining 510 Europ ean-Americans ranging in age from 20 to 86 years. Moreover, while there was an increase in total facial height, upperto-total facial height was found to remain constant. Upper facial hei ght pertains to the distance from nasion to the anterior nasal spine, while lower facial height is the distance from the anterior nasal spine to gnathion. These result were corroborated by Broadbent (1937), Brodie (1940, 1941), Herzberg and Holic (1943), and Sarnas (1957). In c ontrast, Thompson and Kendrick (1964) found during the third and forth decades lower facial height in creased more than the upper facial height for European-American males. Kowalski and Nasje tti (1976) went on to examine vertical facial proportions in a sample of 255 African-American males aged 20-80 years and found relatively constant increases in facial hei ght over time. Using this same population in conjunction with a 28

PAGE 29

population of 1,039 European-American males si milarly aged, Harris and colleagues (1977) focused their attention on alveolar and mandibul ar characteristics within dentate individuals. They suggested craniofacial ch ange was not only age, but race specific. Ruff (1980), on the other hand, proposed increases in craniofacial skeleton size with age were independent of sex and ancestry. Based on prehistoric Amerindian skulls from Indian Knoll, he also believed these changes to be relatively impervious to genetic, environmental, or mechanical factors. Returning to the examination of dental stat e put forth by Heath (1966) and Israel (1969), Lestrel and colleagues (1980) ex amined craniofacial differences between European-American dentulous and edentulous indi viduals over time. Using cepha lograms, mandibular cortical thickness increased among dentat e individuals with age, mirro ring Israels (1969) results. Additionally, edentate individuals demonstrated a marked decr ease in cortical thickness (Lestral et al. 1980). While cross-sectional craniofacial aging research was relatively scarce in the 1980s, renewed interest in this region was evident in the 1990s. For inst ance, Bartlett et al. (1992) noted a modest increase in facial width and depth, a nd an appreciable reducti on of facial height strongly correlated with toot h loss over time in 160 dry European-American skulls. A cephalometric study of lateral patient radiograp hs by Doual et al. ( 1997) found no modification in the anteroposterior diameter of the calvarium with age; however, a highly significant increase of the thickness of this stru cture was noted. Utilizing 84 ma les and 102 females ranging in age from 21 to 101 (ancestry was not specified), the authors also noted a significant increase in posterior facial height (Doual et al., 1997). Meanwhile, relative maxillary retrusion was evaluated by Pessa and colleagues in 1998. Using distance measurements derived from computed tomographic cranial data, the author s analyzed young (18-24 years) and old (43-57 29

PAGE 30

years) European-American males and females with upper dentition. They found a tendency toward lower maxillary retrusion relative to the u pper face with age in both sexes. Furthermore, a slight increase in vertical maxillary dime nsion was also noted (Pessa et al., 1998). Pessa et al. (1999a) expanded upon this materi al a year later by l ooking at age-related change in the orbitomaxillary region. By deri ving vertical distance measures from computed tomographic scan data of European-American males representing infancy (1-12 months), youth (15-24 years), and old age (53-76 years), the author s discovered the ratio of maxillary height to orbital height was greatest at youth. This led them to suggest that the orbital shelf expanded downward while the pyriform aperture migrated upward with age (Pessa et al. 1999a). That same year, ocular globe to orbital rim position with age was evaluated by Pessa and colleagues (1999b), who ascertained posterior movement of the orbital rim re lative to the anterior cornea over time. Another study by Pessa in 2000 utilized angular measurements of three-dimensional stereolithographic images from 12 European-A merican males split equally into young (19-24 years) and old (45-68 years) age groups to evaluate regional shif ts in midfacial morphology. He found a significant decrease in maxillary (superior to inferior maxilla at the articulation of the inferior maxillary wall and alveolar arch) and pyriform angles (nasal bone to lateral inferior pyriform aperture), but no change in glabellar angle (maximal prominence of glabella to nasofrontal suture). In light of these findings and previous res earch, Pessa suggested the midface underwent clockwise rotation relativ e to the cranial base with age (2000). Pessa and colleagues were also the first to attempt to analyze age-related facial bon e curvature in terms of slope analysis (Zadoo and Pessa, 2000; Pessa and Che n, 2002). Specifically, orbital and maxillary arch height was measured at equidistant increments relative to an arbitrary baseline for skeletal material as well as patient-derived computed to mographic scans. While this technique does not 30

PAGE 31

recreate curvature form, it does capture slope cha nge between equidistant points. Both studies utilized European-American males. However, only two age groups were used in the maxilla study: (young: 18-24 years; n=6) and (old: 40-66 years; n=6). Meanwhile, three age groups totalling 30 individuals were provi ded in the orbit study (young: 18-30 years; middle-age: 46-50; elderly: 74-80 years). The points along the older medial maxillary arch existed farther from the defined baseline than those of the younger age group. Significant increases in superomedial height were noted in the superi or orbital rim between the young a nd old age groups. The inferior orbital rim was also characteri zed by significantly increased in ferolateral height between the young and middle-age group (Zadoo and Pessa, 2000; Pessa and Chen, 2002). In 2004, Farkas and colleagues focused their atte ntion specifically on alterations in facial framework in their anthropometric study of 600 European-Americans ranging in age from 16-90 years. Using vertical and hor izontal measurements, they disc overed no consistent pattern of significant change within the facial framework. However, measurement values between early and late adulthood for both sexes displayed a high degree of similarity, whereas no significant differences were seen in mi ddle adulthood when compared to the other two age groups. Most recently, Shaw and Kahn (2007) expande d upon Pessas (2000) angular measurement work by examining facial bone computed tom ographic scans of 60 European-American males and females separated into three age groups: young adult (25-44 years); middle age (45-64 years); and old age ( 65 years). Significant age-related decreases in glabellar angle were noted in both sexes, suggesting recession of nasion and the supraorbital bar. Constant decrease of maxillary angle and increase in pyriform aperture areas were also reported in both sexes with age (Shaw and Kahn, 2007). 31

PAGE 32

Longitudinal Studies In 1950, Buchi conducted a longitudinal study of 200 Swiss adults encompassing youth to old age. Partitioning this population into six ag e classes and examining them twice at nine-year intervals, Buchi found that facial height increased in the 20-28 year age group over the course of the study for both sexes (1950). Facial height re mained steady until age 70, at which point it was found to decrease. Furthermore, Buchi suggested the aging process was marked by a steady increase in head and face diameter up to the 56-64 year age group (1950). A serial cephalometric study of 71 European-American ma les aged 22-34 years measured at one year intervals by Thompson and Kendrick (1964) found a si gnificant increase in total skull height and facial height (total, upper, and lower). Nineteen-sixty-seven proved to be a particularly profitable year for longitudinal publications on craniofacial cha nge with age. Utilizing the same subjects as Thompson and Kendricks 1964 study, Kendrick an d Risinger (1967) reported significant increases in anteroposterior skull dimensions (anterior and po sterior cranial length; head length; and upper, middle and lower facial depth) In contrast, Carlsson and Persson (1967) and Carlsson and colleagues (1967) found no changes in anterior cr anial base length, upper f ace height, and gonial angle within the five-year span between initial and final readings for an adult Swiss population undergoing dental extraction. Furt hermore, Tallgren (1967) reporte d no changes in cranial base and upper face length, or mandibular base shape, height, or length in a series of Scandinavian lateral radiographs of mature denture wearers examined over the span of seven years. An additional series of radiographs taken of the same patients several years later revealed no indications of change (Tallgren, 1972). In 1973, Israel published two studi es derived from longitudinal female samples. In the first study (1973a), 26 European-American females aged 24 to 48 years had cephalograms taken 32

PAGE 33

13-28 years apart. Israel found a slight increase in several cr aniofacial dimensions, including anterior facial height, cranial thickness, and cr anial base length. He also noted increases in condyle-gnathion length and menton-gonion distance. The second study (Israel 1973b), utilizing European-American radiographs from the Fels Longitudinal Study of Growth and displaying age ranges of 26 to 90 years, reported similar increa ses in cranial thickness, skull size, and sella turcica size with age. Bergstrom et al. (1973) also demonstrated increased sellar volume in females. Based on his work, Israel suggested thes e craniofacial changes we re reflective of an overall magnification process later in life (1973a; 1973b). Howe ver, Israels results were questioned on the basis of unstandardized record gathering. Much of th is criticism came from Tallgren (1974), who expanded upon her aforementioned work (1967) with an additional Scandinavian longitudinal study measuring 32 wome n aged 20-73 years over a 15-16 year span. She concluded no significant change occurred in th e cranial vault or cranial base in relation to external or internal size or thickness. Thus, Tallgren (1974) s uggested the adult craniofacial skeleton was marked by dimensional stability with age. Israel responded to Tallgrens criticism of his cephalometric technique in 1977 with a study utilizi ng the female sample reported previously (Israel 1973a) combined with a Euro pean-American male sample. He once again reported age-related differential change which he believed could not be explained simply by improper radiographic pro cedures (Israel, 1977). In 1976, Forsberg examined changes in early adulthood via a cephalome tric study of Swiss men and women in their twenties, divided into tw o groups re-examined at different intervals. One group, re-examined 5 years later, had angular and absolute measures taken. Meanwhile, a second group, re-examined 10 years later, had only angular measurements taken. Increases in lower face height for both sexes were noted for ag es 24 to 29. No changes were witnessed after 33

PAGE 34

29 years of age. Furthermore, Forsberg suggest ed the reported changes were a consequence of posterior mandibular rotation with resultant uprig hting of the incisors, as opposed to continued structural remodeling. A follow-up cephalometr ic study performed by Forsberg (1979) on 25 men and 24 women from the original study produced similar results with the exception of increases in lower face height up to age 34 (0.56mm in 10 years), which the authors attributed to posterior mandibular rotation. An anthropometric study conducted by Susanne (1977) on 44 Belgian males measured at 22 year intervals between the ages of 25 and 60 y ears, described increases in head length and breadth. Facial height and bi zygomatic width increases were also noted. In 1978, Susanne concluded that facial expansion, particularly in terms of facial heights and diameters, occurs during the fourth and sixth decades. Perhaps the most extensive longi tudinal investigation into age-related craniofacial change was conducted by Behrents (1985) on 113 European -American individuals who had participated in the Bolton-Brush Growth Study as children and who had additional cephalograms created for them into adulthood (25-83 years). Behren ts (1985) found genera l stability of the pterygomaxillary fissure, but inferior movement of the posterior palate. Moreover, overall mandibular length increased with age, while gonial angle became more acute, particularly in males. Chin landmarks were found to move downward and forward in males, and simply downward in females (Behrents, 1985). Lewis and Roche (1988) evaluated 20 EuropeanAmerican individuals from the Fels Longitudina l Study of Growth to determine the ages at which certain structures ceased growth. Each participant had a series of cephalograms taken from late adolescence through 40-50 years of age. Focusing part icularly on cranial base and mandibular length, the authors found maximum length expression of these structures occurred 34

PAGE 35

between 29-39 years of age, followed by small, but measurable, length d ecreases in subsequent years. Bishara and colleagues (1994) reported anteropos terior and vertical changes in skeletal dimensions with age for a sample of Europ ean-American males and females ranging from 25 to 46 years of age. Both sexes were found to display an increase in skeletal profile convexity. However, this was attributed to increased max illary prominence in men and posterior mandibular rotation in women. Also in 1994, Formby and colleagues established that adult females displayed more hard tissue changes after the ag e of 25 than before, while most hard tissue changes for adult males occurred prior to age 25. This was based on a longitudinal sample of European-American males and females ranging from 18 to 42 years of age. In 1999, West and McNamara cephalometrically evaluated longitudina l craniofacial changes from late adolescence into adulthood (15 to 48 years) for 58 subjects and reporte d increases in mandibular and midfacial length as well as posteri or and lower anterior facial heig ht enlargement. However, the expression of these alterations va ried between the sexes, such that males displayed anterior rotation of the mandible while females showed posterior mandibular rotation. Continued tooth eruption was also noted in both sexes (West and McNamara 1999). Craniofacial changes occurring in the third d ecade of life were evaluated longitudinally by Akgul and Toygar in 2002. Utili zing lateral cephalometric film s and dental casts from 30 Turkish men and women observed over the span of 10 years beginning in their early twenties, the authors discerned small modifications in craniofacial parameters which were more significant in women. Changes (i.e., increase s) in vertical dimensions such as total anterior, lower, and posterior facial height were reported in both sexes. However, these changes were only significant in women. Movement in the dentoalv eolar region was also found to be primarily 35

PAGE 36

related to tooth eruption in both sexes, while females displaye d significant increases in degree of overbite. Finally, mandibular and maxillary arch measurements diminished in both sexes, but only mandibular arch length decrease was si gnificant in males (Akgul and Toygar 2002). In 2003, Levine et al. directly challenged Pessa s (2000) assertion of midfacial clockwise rotation relative to the cranial base with age as being potentially error prone and limited due to small sample sizes. Their longitudinal trigonom etric analysis of European-American patients from the Bolton Cephalometric study (17-83 years) focused specifically on anterior maxillary and orbital growth via angular changes. The auth ors reported anterior m ovement of the anterior maxillary wall and vertical length enlargement with increasing age (Levine et al., 2003). In 2007, Dager and associates analyzed longi tudinal dental casts from 40 EuropeanAmerican patients (20 males and 20 females) derived from the University of Michigan Elementary and Secondary School Growth Study Casts were created and measured at approximately ages 20, 47, and 55. Nearly universal decreases in maxillary and mandibular arch width, length, and depth were reco rded. Finally, Pessa and colleague s (2008) added to their past work by analyzing age-related changes mandibular curvature. Using 16 (8 males, 8 females) serial frontal radiographs of Bolton Brush Growth Study dentate European-American individuals taken during youth and maturity (age range 560.1 years), the authors located seven twodimensional landmark coordinates along the inferior mandibular border. With the application of geometric morphometric analyses the authors discovered sex di fferences in mandibular shape with age. In particular, the curvature of the mature male mandible was found to be less convex compared to that of the youthful mandible. Furthermore, this curvature loss was most pronounced at the latter th ird of the mandible. 36

PAGE 37

Literature Critique Population Dynamics Overall, the literature reviewed here indicates that the cran iofacial skeleton does change with age. For instance, many cross-sectiona l and longitudinal studies report age-related increases in cranial thickness, h ead length and breadth, cranial base length, and facial diameters. However, and perhaps more interesting, are the ways in which these results are compiled and presented. In general, human variation is de-emphasized as the research samples are predominantly individuals of European ancestry. The results and conclusions drawn from this relatively homogenous population ar e often applied globally to the condition of human aging. Thus, craniofacial aging is pres ented as a single thing, applic able to everyone, regardless of temporal or geographical location and genetic ancestry. Scientific racism likely informed this stance in craniofacial literature from the late 19th century and early 20th century. For instance, in A Tr eatise of the Human Skeleton (1858), G.M. Humphrey attributed the the ugliness of the Negros mouth (284) to the inclination and size of the maxillae and lips. Clearly, though, racism can little explain the reliance upon Europeans and European-Americans in recent work. Instead, issues of availability, tradition, and critical comparison likely underlie the use of European/European-American samples in contemporary research. Researchers interested in utilizing existing wi de-ranging longitudinal data are primarily restricted to data from su ch projects as the Bolton-Brush Growth Study, the Fels Longitudinal Study of Growth, and the Mi chigan Growth Study; all of which were originally derived from European-American school children. Moreover, the United States and Europe is com posed primarily of individuals of European ancestry. For instance, as of 2005, European-A mericans (i.e., non-Hispanic whites) accounted for 66.4% of the US population, followed by indivi duals of Hispanic or Latino origin (14.8%), 37

PAGE 38

African-Americans (12.8%), and Asian-American s (4.4%) (US Census Bureau, 2005). As such, acquisition of a European-American clinical sample is much eas ier than other ancestral groups due to sheer availability. In addition, researchers may attempt to mimic sample dynamics utilized in earlier works as a means of directly evaluating an d critiquing other study results. This, in turn, self-perpetuates the trend toward utilizing European/European American material. However, it is well-established that morphologi cal variation in the cranium exists between different ancestral groups (e.g., Buikstra and Ubelaker, 1994; Rhine, 1990; White, 2000). These ancestral differences may very well inform the way( s) in which age manifests in the craniofacial skeleton, as suggested by Harris and colleagues ( 1977). Thus, exploration of social and genetic variation in this region cannot help but better capture the spectrum of human variation. Methodological Constraints This literature also exploits, almost exclusively, craniometric and cephalometric techniques to address the question of craniofacial aging. Th e very nature of distan ces and angles, in terms of where they can be collected, constrains data capture to a small region of the entire structure under study, thereby limiting the amount of geometric information available for capture. This type of analysis serves as a road map, displaying major and some minor areas of interest, but fails to illuminate the interstitial spaces. Adva nces in modern statistics, however, offer a remedy to this shortcoming. Specifica lly, the field of geometric mor phometrics fuses together biology and geometry in the study of shape in twoand three-dimensional space (Bookstein, 1982). This field, then, can offer a finer scale description of local and global variatio n than current methods. Perhaps again owing to the tradit ion of the field, and a desire to produce comparable results, this method remains underutilized with the exception Pessa and colleagues (2007). Given the potential for age-related changes to manifest differentially betw een the sexes and in different ancestral populations, and the benefits of more fully capturing the aspects of the face upon which 38

PAGE 39

39 we socialize everyday, I posit that geometric morphometrics is a superior method with which to analyze the aging facial skeleton. The following chapter further explicates this method and its utilization during my data gathering and analysis.

PAGE 40

CHAPTER 4 GEOMETRIC MORPHOMETRICS Evolution is propelled forward by the raw ma terials provided by biological variability. Through a careful study of variability, we are able to gain a fuller understanding of both who we are now, and who we were in the past. Studi es of variability, though, have predominantly focused on comparing gross anatomical features w ithin and between organisms. This level of analysis served as the basis for typologies su ch as the Linnean taxonomic system. While that typology has proven durable even when viewed in genetic terms, most classic analyses of human anatomical variation problematically re ly solely on qualitative comparisons. Qualitative analyses, dependent upon terms su ch as round or coarse, are basally imprecise in that these sorts of adjectives lend themselves to subjective interpretation between observers. Whereas one observer ma y deem a structure to be round, another may perceive it as oval or even squarish. This reliance upon subjective English language terminology not only fosters confusion amongst native speakers, bu t likely inhibits prof essional understanding between native and non-native English speakers. Accordingly, results, conclusions, and interpretations advanced via qual itative studies ar e often suspect. In contrast, the language of math is objectivel y precise; numerical valu es are constant and thus invariant to personal inte rpretation. While, of course, the ways in which these numerical values are gathered are always subject to sele ction and measurement bi as, the definition of a numerical value such as is always the sa me. Assuming the absence of measurement and selection bias, this universality in numerical definition enables explicit comparison between subjects on a much finer scale than qualitativ e analyses. And, depending on the anatomical structure under investigation, metric analyses can offer insight into growth patterns, biomechanical design, and evolutionary processes. As scholars realized the objective power of 40

PAGE 41

quantitative analyses, they began to set aside techniques based upon descriptive labels. This endeavor was aided by advances in multivariate statistical methods (i.e., principal component analysis, correlation coefficients, etc.). By the mid 20th Century, researchers had fused their studies of quantitative shape an alysis with statistical methods creating the m odern field of morphometrics (Adams et al., 2003). Traditional Morphometrics Early work in the field of morphometrics during the 1960s and 1970s applied multivariate statistical methods to morphological variables us ually in the form of linear distance measures between well-defined points and/or angles. K nown as traditional morphometrics, this allowed for quantification of covariation in the measurements and assessmen t of variation pa tterns within and between specimens. Despite these dramatic advances in morphologi cal analyses, certain obstacles to the quantitative cap ture and representation of anat omy remained (Adams et al., 2003). For instance, it is difficult to access the hom ology of linear distances because homology presumes biological correspondence. More precis ely, the location of a feature must be present in the exact same place between organisms. Unfortunately du e to biological va riability, this is rarely the case. Moreover, endpoi nts of linear distances may not ex ist in visually obvious areas such as the intersection of sutures, thus increa sing the likelihood of measurement error. Due to these two complications, the precis e capture of linear distance ma y be grossly imprecise, and the results incomplete (Adams et al., 2003; Slice, 2005). Additionally, a measurement cannot be taken at all if one of the two endpoints is dist orted and/or absent, thus reducing the amount of shape information one can capture. The final, and perhaps most important i ssue here, is the lack of context which exists between such measuremen ts. The spatial (and th erefore causal and/or functional) relationship between and among the endpoi nts is lost when using linear measures and 41

PAGE 42

angles (Adams et al., 2003; Slice, 2005). Wit hout this information we are unable to fully visualize the morphological differe nces between specimens. We ar e thus left with incomplete information and, by extension, inadequate understanding of biol ogical variability. Geometric Morphometrics In the 1980s, geometric morphometrics (GM) was developed as a way to more fully capture and analyze morphological shape, than pr evious methods allowed. GM defines shape as the geometric properties of an object which are invariant to or ientation, location, and size (Slice et al., 1996). Shape is often captured within GM using anatom ical landmark coordinates, in either two or three dimensions, in lieu of trad itional distance or angle measures. Anatomical landmark coordinates are capable of reflecting biological homology in tissue and anatomical location across specimens (Bookstein, 1991). They are categorized into three groups: Type I discrete juxtapositions of tissue; Type II poin ts of maximal curvature; and Type III extremal points (Bookstein, 1991). Landmark morphometr ics often utilizes the same endpoints employed in linear distance measures and the construction of angles. As a consequence, collection of landmark coordinates automatically allows for the calculation of every possible distance and angle measure which employs them using basi c geometry and trigonometry. Moreover, the relative spatial configuration of landmark coordinates is pres erved throughout an analysis, thereby allowing for visualization of morphologi cal differences using interactive computer graphics (Rohlf and Marcus, 1993). In the process of preserving spatial configuration, though, th e effects of variation in position, orientation, and scale must be sequestered prior to the analysis of landmark coordinates as shape variables. This can be accomplished via superimposition methods which overlay landmark configurations according to some optim ization criterion. Generalized Procrustes analysis (GPA) superimposes landmark configurations via least-squares estimates for location 42

PAGE 43

and orientation that minimize the sum of squa red distances between corresponding points on two configurations. Simply put, GPA registers a se ries of forms onto a common coordinate system by sequestering rotational and translational differe nces and scaling them in such a way as to maximize their best fit ( Gower, 1975; Rohlf and Slice, 1990; Bookstein, 1991; Goodall, 1991 ). These transformed landmark coordinates can then undergo rigorous statistical analyses to assess significant group differences and/ or quantify group means or va riation depending on the nature of the research question. Visualization of the anatomical differences accounting for significant group variation can be performed using vector plots. In a two group comparison, difference vectors are drawn from the locations of the landmark coordinates of one configuration (reference) to locations of the landmark coordinates of the second configuration (target). These difference vectors (i.e., directional vectors of change) are represented as lines ex tending outwards from the landmark coordinates and can be exaggerated for visualization purposes by multiplying the displacement matrix by some factor. They illustrate the dire ction of movement required at each coordinate to transform the shape of the reference specimen in to the shape of the ta rget specimen (Slice, 2005). Ridge Curve Analysis While landmark coordinate data is vastly superior to measurement and angle data in characterizing biological shape, some anatom ical regions, such as boundaries and surface curvature, lack well-defined landmarks. GM addresses this issue via ridge curve analysis, which is capable of characterizing and analyzing curvat ure. Morphological curv ature can be captured by collecting a sequence of points (continuous stream data) along the margin of a structure, such as the superior orbital rim. The minute details of the structure, however, are often blurred in this process due to the large pool of coordinates collected, much of which contains redundant 43

PAGE 44

information (Rohlf, 1990). A solution to this issue is the conversion of data into a more compact form via the creation of semilandmarks. In this process, curve coordina te data are resampled by an algorithm into a user-defined number of evenly-distributed fixe d or sliding coordinate points, capable of preserving the visual shape of the or iginal curve. Semilandmarks made their first appearance in the literature in the appendix of Booksteins Morphometric Tools for Landmark Data: Geometry and Biology (1991) and were first formally applied to two-dimensional data by Bookstein in 1997. Equidistant Semilandmarks The simplest form of semilandmark data are represented by the placement of equallyspaced points on curves or surfaces. An algorithm resamples the large amount of raw curve data into a manageable number of evenly-distributed fixed points. The number of semilandmarks is user-defined and reflects just enough points to maintain th e original curves shape. For instance, while 100 raw points may have been collect ed along the margin of the superior orbital rim, 10 points spread evenly along the curve itself may adequately reflect the structures overall shape to the naked eye. To eff ectively analyze the data in a sta tistically meaningful manner, the researcher needs to compare the same number of semilandmarks for all of the contours measured. Once the semilandmarks are transfor med by GPA, multivariate statistics can be applied to them. There are, however, limitations in the applicability of fixed semilandmarks due to their inherent nature. For example, since they are generated by a si ngle algorithm, fixed semilandmarks are mutually-correlated. Consequen tly, they provide a somewhat biased estimate of shared diversity by diminishing variability. Moreover, Guntz et al. (2005) have shown that structural artifacts of equidist ancy may manifest as strong loca l shape differences when fixed semilandmarks are transformed and visualized. Unlike anatomical landmarks, whose relative 44

PAGE 45

position is stable regardless of the structural variation occurring around them, the final fixed position of a semilandmark is influenced by the complexity of the curvature from which it is derived. Thus, inter-specimen variation in the convolution of th e curve (i.e. degree of twists, turns, and bends) may result in variation in the final position of the same semilandmark between specimens as a consequence of the simplisti c nature of the algorithm from which the semilandmarks were derived. When analyzed, this statistical artifact can be misinterpreted as biological variation, thereby conflating perceived differences between groups. This issue can be somewhat mitigated by the use of true anatomical landmarks (Type I or II) as the beginning and endpoint of the curve. These points are anchored at strictly defined locations, thereby constr aining the position of the semilandmar ks between them. This is the equivalent of tying down a tarp during a storm. With only one side tied down, the tarp will flap in the wind, but with both sides ti ed down it will remain relatively stable. Moreover, regions displaying complex twists or sharp bends shoul d be avoided when appl ying equidistant fixed semilandmarks, as these points are derived from the overall length of the curve. For example, the passageway for the supraorbita l nerve and vessels may manifest as a hole above the superior orbital rim or as a notch along it. The presentation of a notch creates a small bubble imperfection in the overall curvature of the orbital rim. If this notch is captured when collecting curvature data of the rim, the overall distance me asure of the points will be greater for orbits which have notches than those which do not. By consequence, semilandmarks derived from orbits with notches will be positioned slightly di fferently than those derived from rims without notches. This inter-specimen variation in mo rphology may manifest falsely as variation in overall orbital rim shape. One solution to th is dilemma would be to simply bypass the notch 45

PAGE 46

when collecting the initial orbital rim data, thereb y maintaining structural consistency within the data set. Sliding Semilandmarks In some structures, such as the corpus callo sum which connects the two hemispheres of the brain (Bookstein, 1997; Bookstein et al., 2002), these sharp bends and twists cannot be avoided. Here, sliding semilandmarks serve as a viable alternative. The perp endicular projection or minimum Procrustes distance criteria is often em ployed to align semilandmarks (Andresen et al., 2000; Sheets et al., 2004). This criteria can be thought of as an extensio n of GPA, wherein the coordinates are allowed to slide along their curve until all the semilandmarks on a target specimen correspond as closely as possible in position to the equivalent se milandmark points in a reference specimen (Adams et al., 2004), thus mini mizing the Procrustes di stance (Sheets et al., 2004). As a result, the target forms semilandma rks rest along lines exis ting perpendicular to a curve traversing the corresponding semilandmarks on the reference form (Sampson et al., 1996; Sheets et al., 2004). The sliding movement of a semilandmark is also often constrained by the fixed positions of surrounding t rue anatomic landmarks (Bookste in, 1997; Guntz et al., 2005). Like equidistant fixed semilandmarks, sliding se milandmarks are not true landmarks. Instead, they exist relative to one another. Thus, specimens are compared by examining the entire sequence of sliding semilandmarks, which corresponds to the entire contour curvature (Weber et al., 2001). Both techniques allow for analysis of a wide spectrum of morphological regions by not restraining the resear cher to only areas cont aining true landmarks. Geometric morphometrics represents a mo rphological revolution in the capture and visualization of biological form Unlike the morphological vari ables (i.e., distance and angle measures) and statistical analysis techniques (i.e., univa riate) commonly proffe red in craniofacial studies of age, GM retains spatial relationships between variab les and, by consequence, allows 46

PAGE 47

47 for the depiction of subtle changes in shape. As such, the application of morphometric techniques, specifically ridge curve analysis, can not help but better reflect alterations in the craniofacial skeleton with age.

PAGE 48

CHAPTER 5 MATERIALS AND METHODS The data utilized in this project consist of fixed equidistant semilandmarks derived from three-dimensional continuous stream coordinate data of facial bone curvature. Specifically, I recorded continuous stream data obtained from contemporary African-American and EuropeanAmerican dry crania. Born in the 19th and 20th centuries, the biological ag e at death and sociallydetermined race of these individuals were obt ained from autopsy reports and associated paperwork. In order to maintain homogeneity w ith the studies previously mentioned, and to conform to present-day notions of social race, I treated documented distinctions of black or white as synonymous with African-American and European-American during the data gathering and analysis stag es of this project. To ensure a suitably sized sample, I collect ed data from several different skeletal collections. Specifically, I made use of cran ia from the Terry Collection, housed at the Smithsonian Institution (n = 464) ; the Bass Donated Collection, lo cated at the University of Tennessee, Knoxville (n = 101); the Hamann-Todd Osteological Collection at the Cleveland Museum of Natural History (n= 104); as well as the Maxwell Museums Documented Skeletal Collection at the University of New Mexico, Albuquerque (n= 17). Skeletal Collections Robert J. Terry Anatomical Collection The Robert J. Terry Anatomical Collection is one of the largest and most widely utilized skeletal collections in the world. Housing 1,728 documented skeletons, the Terry Collection was established by its namesake in the first decade of the 20th century. Originally housed at the Missouri Medical College, the collection currently resides at the Smithsonian Institution and represents primarily African-Americans and European-Americans born during the 19th and 20th 48

PAGE 49

centuries (Hunt and Albanese, 2005). Principall y derived from St. Louis hospitals and morgue cadavers intended for use in medical school anat omy classes, the skeletons amassed under Terry represent the lower socioeconomic classes of St. Louis and Missour i. The skeletons are mostly individuals unclaimed by relatives and thus became property of the state. With the exception of some excised calvarium for brain dissections, sa gittally-sectioned crania and sectioned sterna, the skeletons are more often than not comple te and intact. Moreover, Terry focused on representing the wide range of normal skelet al variation, and thus pathological specimens are rare (Hunt and Albanese, 2005). Following Terr ys retirement, skeletal collection continued under Mildred Trotter until her subsequent reti rement in 1967. During her tenure, Trotter attempted to fill perceived demographic gaps in the collection. In particular, Trotter increased the number of European-American females and younger individuals. Thanks to an economic surge in the decade following World War II and a so cial shift in regards to anatomical study and body donation, individuals collected under Trotter generally came from a higher socioeconomic station than those amassed by Te rry (Hunt and Albanese, 2005). Documentary materials are also quite prolific for the indivi duals in this collection, and while not all-encompassing, every attempt was made by Terry and Trotter to locate accurate age at death information for the collection. In in stances where definite age is unknown, estimated age categories in multiples of 5 are provide d. The Terry Collections age-at-death skews toward older individuals, with a mean age at death for males of 53 years and 58 years for females. Furthermore, years of birth ra nge from 1828 to 1943 (Hunt and Albanese, 2005). Hamann-Todd Collection The Hamann-Todd Collection is also one of the largest, most researched skeletal collections in the country. Totaling 3,592 skeletons it is actually the merger of two separate collections; the Hamann and the Todd Collect ions. Launched in 1893 by Carl August Hamann, 49

PAGE 50

the Hamann Collection was amassed primarily from anatomical teaching specimens. The Todd Collection began in 1912 and was assembled under the direction of T. Wingate Todd (Ayers et al., 1990). The two collections were later combined to form the Hamann-Todd Collection which currently resides at the Clevel and Museum of Natural History. Representing African-Americans and European-Americans of lower socioeconom ic status, the Hamann-Todd Collection ceased procuring skeletons after 1938. Thus, the indivi dual represented within were born almost exclusively in the 1800s. The African-American portion of the collection represents southern African-Americans who migrated to Cleveland primarily from Georgia, Alabama, and South Carolina. Meanwhile, 60% of the European-Ameri cans with known birth places originated from Europe, while the majority of the American-born individuals came from th e Northeast (Ayers et al., 1990). Despite prolific documen tation (i.e., autopsy records, etc.), age-at-dea th information has been found to be suspect among some of the individuals in this co llection (Lovejoy et al., 1985). Lovejoy and associates (1985) attempted to address this issue by ranking a subset of the specimens according to stated age and observed age. Furthermore, an internal reexamination of age-at-death was conducted on th e entire collection by Robert Me nsforth and is available to researchers along with stated age (unpublished data 2006). In order to ensure a reliable sample, I only chose those individuals whose stated age an d observed age fell within the particular age ranges of this study (discussed below). Maxwell Museums Documented Skeletal Collection Established in 1984, the Maxwell Museums collec tion consists entirely of individuals who passed away within the last 25 years. Repr esenting 235 individuals (as of September 2003), collection material was obtained by donation or by the Office of the Medical Examiner when no family could be located. The collection repres ents individuals primarily from New Mexico. 50

PAGE 51

Heath and occupational data have also been co llected since 1995. The Ma xwell Collec tion is the largest, most well documented co llection of its kind in the American West (Human Osteology LabMaxwell Museum, 2007: http://www.unm.edu/%7Eosteolab/coll_doc.html ). William M. Bass Donated Skeletal Collection Established in 1981 by William Bass, this collection currently contains over 400 individuals. These donated indi viduals are provided by the Tennessee State Medical Examiners Office; the University of Tennessee, Knoxville, Medical Center; and fune ral and nursing homes. Typically, donated individuals are accompanied by well-docum ented medical records. Demographically, the collection is heavily skewed towards European-American males, though African-Americans of both sexes are presented, as well as a handf ul of Hispanic individuals. Age ranges extend from fetal up to 101 years, but the collection as a whole mirrors mortality rates commonly associated with lower soci oeconomic status (Marks, 1995; Forensic Anthropology Center, 2007: http://web.utk.edu/~anthrop/FACresources.html ) Inclusion Criteria One of the primary analytical decisions I ma de during data collection was to exclude any crania displaying obvious pathology, since disease states may alter the natural curvature of the facial skeleton. In addition, given the norma l wear and tear incu rred on the human face throughout life and during curati on, many individuals did not have all 16 curves (see below) intact for measurement. As a result, I arbitr arily established a maximum threshold of trauma below which a cranium could be included. Accordingly, I gathered data on crania with regional trauma (ante/peri/postmortem) as long as: (1) data were not collected on these regions, and (2) said trauma did not compromise neighboring struct ures. Within this guideline, then, I included some sagittally-sectioned crania that could be re-articulated with little to no deformation. Nasal 51

PAGE 52

and maxillary curve data, however, could obviously not be collected from these crania due to damage during the sectioning process. Ideally, I would have only utili zed individuals with a full dentition in order to maintain morphological consistency between crania. Unfortunately, the c onfluence of factors such as trauma, dental disease, overall health, diet, and curation necessitated an additional adjustment of my selection criteria. Generally, young adult and middle-aged specimens, versus the elderly individuals, displayed only minor tooth loss with associated al veolar resorption. As such, I excluded those maxillary curves that exhibited al veolar alteration due to tooth loss, yet still captured curvature data from the remaining regions of the face. However, maxillary change subsequent to substantial tooth loss can be quite severe. Initially, the maxillary alveolus th ins in the labiopalatal directi on, while still retaining alveolar height. This manifests as a knife-edged ma xillary ridge (August and Kaban, 1999). Yet, as bone loss progresses, alveolar he ight is lost and the anteri or nasal spine becomes more prominent. The consequential superior and posterior movement of the maxilla with bone loss results in decreased lower facial height (Augus t and Kaban, 1999). During the data collection process, it became readily apparent that most of the crania over 60 years of age displayed this sort of dramatic maxillary bone resorption. In light of these pronounced bony changes, I became concerned that the inclusion of any curve data from such individuals coul d potentially reflect the long-term craniofacial effects of edentulism as opposed to the aging proces s in general. This concern prompted me to modify my criteria, agai n. In order to mitigate for potential structural alterations due to severe alveolar resorption, particularly in th e regions closest to the maxilla (i.e., zygomatic arch and nasal aperture), yet still allow for inclusion of elderly specimens, I excluded specimens who were nearly or complete ly edentulous. I only included curvature data 52

PAGE 53

(with the exception of the maxillary curve) from those older specimens with low to moderate alveolar resorption. I defined m oderate alveolar resorption as the bilateral presence of some posterior dentition (at minimum molars); maintenan ce of most of the anterior teeth and/or tooth sockets; and, in instances of concentrated regional edentulism, socket resorption which could not dramatically compromise the overall shape of the maxillae. The final requirement pertains to maintenance of several millimeters of bone belo w both the antra (sinus) and nasal cavities. Cranial Regions and Age Groups For the purposes of my study, each racial population was divided into three age groups, containing approximately equal numbers of males and females: young adult (18-39 years), middle-aged (40-59 years), and elderly (60+ years). Appendices A-1 and A-2 provide demographic specifics for the number of indivi duals in each age group. These age groups were organized as such in an attempt to broadly capture the span of human life. Though the age groups are socially-defined, I made the presumption that age-related biologic al change influences these social constructions. Age 18 was set as the beginning of adulthood in light of the fact that skeletal maturity (i.e., fusion of the spheno-oc cipital synchondrosis a nd eruption of the third molars) is typically reached by this age. Individuals di splaying delay in the completion of these two features were excluded regardless of documented age. In order to evaluate the presence of varia tion in facial bone contour within each age group a hierarchical approach was taken to the f ace. More precisely, the facial region was deconstructed into 16 curves (see below) which represented the contour of individual facial bones. These curves were then analyzed in diffe rent combinations to re flect the face globally, regionally, and locally. Facial re gions included the eye orbits, zygomatic arches (including the temporal portions), nasal aperture, and max illary ridge. Locally, the frontal, zygomatic, maxillary, nasal, and temporal bones were repres ented. The simplicity of each individual curve 53

PAGE 54

not only minimized the potential for statisti cal artifacts during se milandmarks extraction (Chapter 4), but when combined these curves were capable of reflecting topologically complex regions such as the zygomatic ar ch. Furthermore, when these regions were combined they, in turn, captured the complexity of th e entire face. I believed this transition from global to local in the facial skeleton would allow complex interact ions to surface statistically. Thus, during my analysis the question of facial shape variation between age groups was posed at smaller and smaller scales until no significant differences were detected. Hypotheses Based on changes documented in other areas of the craniofacial skeleton (Chapter 3) with age, I posited the following hypotheses: (H1) Statistically significant changes in shap e occur in the curvature of the facial skeleton with age. o (H1-A) These changes can be detected globally. o (H1-B) These changes can be detected regionally. o (H1-C) These changes can be detected locally. (H2) Statistically significant two-way inter actions occur between age and sex, and age and race in the curvature of the facial skeleton. o (H2-A) These interactions can be detected globally. o (H2-B) These interactions can be detected regionally. o (H2-C) These interactions can be detected locally. (H3) Statistically significant three-way inte ractions occur between age and sex and race in the curvature of the facial skeleton. o (H3-A) These interactions can be detected globally. o (H3-B) These interactions can be detected regionally. o (H3-C) These interactions can be detected locally. 54

PAGE 55

Data Collection Procedure I used a Microscribe 3DX digitizer (Imme rsion Corp., San Jose, CA) interfaced to a personal computer to collect cont inuous stream data representing 16 curves (8 from each side) from the facial region of the cranium (Table 5-1, Figure 5-1). This hardware translates data from the region of interest, via a sensor in the stylus point, into x y and z coordinates. Data points were recorded in millimeters to two decimal places, collected every 0.5mm, and saved as a database file (.db) within 3Skull a craniometric software program written by Stephen D. Ousley (2002). To ensure stability during the data collecti on process, I affixed each cranium to a metal ring stand covered with dental wa x prior to digitization. I then placed the tip of the handheld stylus on the designated anatomical landmark, tracing the regional border of interest until I reached the terminal anatomical landmark. Prior to actual data collection, I clearly marked the beginning and endpoints of each curve. By movi ng the stylus in a single smooth motion from beginning to endpoint, I was able to maintain continuous contact with the bone. This continuous contact guaranteed proper data collection throughout the process. By tapping a foot pedal attached to the Microscribe input device I could indicate the begi nning and end of data recording to the computer. Ta ble 5-1 and Figure 5-1 detail the 16 curves measured, as well as their respective beginnin g and endpoints. Curve numbers 1 th rough 8 represent the left side of the facial skeleton, while numbers 9 through 16 are presented on the right side. Definitions of the anatomical landmarks can be found in Appendix B-1. Special Curve Considerations Since my goal was to extract equidistant se milandmarks from the raw curvature data, I made all attempts to avoid regions within a gi ven curve that exhibited sharp infolds. As mentioned in Chapter 4, these regions of distorti on (e.g., superior orbital notch: see below) can 55

PAGE 56

create a suggestive pattern of biol ogical variation, but are in realit y nothing more than statistical artifacts of equidistant semilandmark crea tion. As such, I empl oyed certain regionallydetermined techniques to maintain smooth cont our collection. The following section outlines any special methods I employed in the data capture of each curve. Alveolar Curve Beginning at prosthion, I used th e stylus to trace the alveolar bone just above the tooth line where a smooth line of bone could be captured. This prevented sudden spikes in stylus movement at the alveolar bone/tooth margin. I filled in small areas of chipped alveolar bone with dental wax prior to data collection. Nasal Curve Some specimens displayed sharp infoldings of bone along the nasal aperture. As a result, the region appears jagged, rather than smoothly ar ched. These regions obviously distort overall curvature. Thus, I applied dental wax to the underside of the nasal bone to fill in these regions. This allowed the data gathering proce ss along this region to be uninterrupted. Superior Orbital Curve The superior orbital curve contains a short region between maxillofrontale and the superior rim which lacks easily perceivable curvature. This presented a challenge in terms of making consistent tracings in this zone. Thus, I drew a straight pencil line, with the help of a flexible ruler, through this region. In so doing, any di screpancy in data collection was mitigated, while consistency was maintained among the tracings. Some individuals also displayed supraorbital notches unilaterally or bilaterall y. These regions appear as distortions within the larger curve of the orbit. Such distortions can lead to misr epresentation in the subsequent calculation of semilandmarks between specimens. As such, I re moved these regions by filling the notches with dental wax, thereby allowing smooth continuation of the orbital rim. 56

PAGE 57

Inferior Orbital Curve The border of this curve is particularly well-defined, allowing for easy data collection across specimens. Outer Orbital, Superior Temporal, Inferior Temporal, and Inferior Zygomatic Curves Given easily located landmarks and smooth, we ll-defined borders, no special adjustments were required during data collection. Intra-Observer Error In an effort to evaluate my degree of intra-ob server error in recording the orbital curves, I digitized three crania, three times, over a peri od of several days. I maintained consistent orientation for each cranium for all measurement sessions. I then derived semilandmarks (see below) from the raw data for each facial region (i.e., maxillary ridge, nasal aperture, zygomatic arches, eye orbits) usin g a beta version of Morpheus et al. (Slice, 2005). Finally, I produced a matrix estimating the general variation local to each landmark without reference to the x y, or z axes for the pooled regional facial data as outlin ed by Valeri and associates (1998). The average error (across crania and semilandmarks) for each region was as follows: eye orbits= 0.35mm; zygomatic arches= 0.32mm; nasal aperture= 0.31mm; and, maxillary ridge= 0.27mm. The Microscribe 3DX has a reported accuracy of .23mm (Immersion Corporation 1998). Based on the accuracy of the digiti zer, the transformation proce ss required to generate the semilandmarks (see below), and their threedimensional nature, I concluded that my measurement error was negligible. Data Manipulation and Analysis Semilandmark Extraction As I mentioned previously, the raw continuous stream data represent lines of closelyspaced points bounded by fixed anatomical landmarks (Figure 5-1). I checked for any random 57

PAGE 58

points generated accidentally by the 3Skull soft ware or a sudden hand jerk by graphing each individual within the beta version Morpheus et. al (Slice, 2005). When found graphically, these points were located in the raw data set and re moved prior to the creat ion of semilandmarks. Once I established a clean data set, I crea ted semilandmarks within the beta version Morpheus et. al (Slice, 2005). Specifical ly, I applied an algorithm to the data which individually resamples each curve into a user-defined number of evenly-d istributed fixed semilandmarks. This userdefined number is meant to best preserve the vi sual shape of the orig inal curve (n= 126 for entire face) (Figure 5-2 and Table 5-2). Admittedly, the process of selecting semiland mark number was arbitrary, but I visualized several different configurations, with varying nu mbers of semilandmarks for each curve, before deciding on the values which I believed best captur e curve shape. Using the classic version of Morpheus et al. (Slice, 1998), I fit the semilandmarks into a common coordinate system via a generalized Procrustes analysis (GPA), which filters out the effects of location, scale and rotation. Finally, I prepared th ese coordinates to be imported into Microsoft Excel 2002 for subsequent multivariate statistical analysis. Multivariate Statistical Analysis The following multivariate statistical techni ques were applied to the facial skeleton curvature globally (all curves at once), regionally (curves making up the maxilla, nasal aperture, orbits, and zygomatic arches), a nd locally (individual curves). As the number of semilandmark shape variables generally exceeded the number of individuals collected for any one racial population, sex, or age group, I attempted to redu ce dimensionality by performing a principal component analysis (PCA) on th e covariance matrix of the GPA-aligned coordinates and utilizing the resulting principal component (PC) scores which accounted for 95% of the total sample variance. 58

PAGE 59

I next conducted a non-paramet ric, multivariate analysis of variation (MANOVA) using the aforementioned PC scores. MANOVA allows me to test for statistically significant mean shape differences between the age groups as well as their interaction with race and sex. All MANOVA results were reported in terms of the Wilks Lambda test statistic. The multivariate Fvalue obtained is based on a comparison of the e rror variance/covariance matrix and the effect variance/covariance matrix. Although I only men tion Wilks' lambda in this research, other statistics, including Hotelling's trace and Pillai's criterion, we re given during the analysis. Degrees of freedom (df) were also provided in the MANOVA results and describe the number of values in the final calculation of a statistic that are free to vary. Finally, the significance level (alpha) used in all my multivariate analyses was set at 5% (i.e., p-value = 0.05 threshold). The pvalue represents the probability of getting some thing more extreme than your result, when there is no effect in the population. T hus, I rejected the nu ll hypothesis (no difference/effect) in favor of my own alternative hypotheses (see above) when the pvalue was less than 0.05. Both the PCA and MANOVA were conducted within SAS 9.1.3. However, neither Windows Excel 2002 nor SAS 9.1.3 is capable of handling the shear number of characters created when examining th e face in its entirety (i.e., 126 semilandmarks representing 378 coordinates). As such, for my global examination of facial curvature change I exported the 378 coordinates from Morpheus et al. (Slice, 1998) into a text file and cut and pasted them into PAST ( http://folk.uio.no/ohammer/past/download.html ), a free multivariate software program designed to analyze massive data se ts. PC scores for the entire face were then generated in PAST (Hammer et al ., 2008). These scores, containing far fewer characters than the original data set, were finally pasted into SAS 9.1.3 and underwent subsequent MANOVA. 59

PAGE 60

To better understand the significant age group differences established by the overall F-tests (i.e., which age groups were differe nt from one another), I perfor med multivariate contrast tests on the regional PC scores within MANOVA. A contrast test is a comparison of means among some or all of the groups. In this particular re search, the contrast test s tested the null hypothesis of no significant difference between the means of young and middle-aged adults; middle-aged and elderly adults; and young and el derly adults, respectively. A si gnificant contrast test result indicates that the two age group m eans under investigation differ. I further characterized these regional ag e group differences by conducting a canonical variates analysis (CVA) of the regional PC sc ores for each respective population (i.e., AfricanAmerican or European-American) or subpopul ation (e.g., African-American female or European-American male). Canonical variates analysis is a dimension-reduction technique which finds linear combinations of the princi pal components that provide maximal separation between the age groups. The en suing canonical variate scores pr oduced by this analysis were then plotted within Windows Excel 2002. Both the contrast tests and the canonical variates analyses were conducted in SAS 9.1.3. Finally, I visualized my contrast test resu lts in terms of differences in semilandmark location by using the vector plot feature in Morpheus et al. (Slice, 1998). The vector plots reflect the directional change (in terms of the x, y, and z axes) required to relo cate the semilandmark locations of one configuration (reference) onto the co rresponding semilandmarks locations in a second configuration (target). Th e reference and target specimens used in this research are the mean shapes (consensus configurations) of the age groups found to be different from one another. Red lines extending from the sem ilandmark coordinates repr esent the directional vectors of change from refe rence to target points. 60

PAGE 61

In conclusion, the facial skeleton was pres ented globally, regionally, and locally using equidistant semilandmarks. These semilandmark s, which characterize facial bone curvature, were subjected to multivariate statistical analys is to search for statistically significant shape differences between age groups and their rela tionship to race and sex. By utilizing semilandmarks and various multivariate statistic t ools, I believe I am the first to capture the complexity of facial bone contour with age. This particular approach not only allows examination of facial change at various levels, but enables th ese changes to be visualized spatially. As such, this approach is distinctly novel within the realm of age-related craniofacial skeleton research. 61

PAGE 62

Figure 5-1. Curve tracings of the face. Landmar k abbreviations are as follows: prosthion (pr); ectomolare (ecm); subspinale (ss); rhini on (rh); maxillofrontale (mf); frontomalare orbitale (fmo); frontomalare temporale (fmt); zygotemporal superior (zts); auriculare (au); porion (po); zygotemporale inferior (zti); zygomaxillare (zm). [Drawing rendered by Shanna Williams.] Table 5-1. Curve definitions Curve Number Curve Name Definition 1, 9 Alveolar prosthion to ectomalare 2, 10 Nasal rhinion to subspinale 3, 11 Superior Orbital maxillofrontale to frontomalare orbitale 4, 12 Inferior Orbital maxillofrontale to frontomalare orbitale 5, 13 Outer Orbital frontomalare temporale to zygotemporale superior 6, 14 Superior Temporal zygotemporale superior to auriculare 7, 15 Inferior Temporal porion to zygotemporale inferior 8, 16 Inferior Zygomatic zygo temporale inferior to zygomaxillare 62

PAGE 63

63 Figure 5-2. Extracted semilandmark s for individual with all curves present. Image created in Morpheus et al. (Slice, 1998). Table 5-2. Number of semiland marks extracted for each curve Curve Number Curve Name Number of Semilandmarks Extracted 1, 9 Alveolar 7 2, 10 Nasal 7 3, 11 Superior Orbital 10 4, 12 Inferior Orbital 10 5, 13 Outer Orbital 6 6, 14 Superior Temporal 7 7, 15 Inferior Temporal 9 8, 16 Inferior Zygomatic 7

PAGE 64

CHAPTER 6 RESULTS Preliminary Data Analysis Pooling the Data In advance of delving into the main data anal ysis for this study, I a ddressed two conceptual issues. Foremost, I evaluated the appropriateness of pooling the data into aggregate groups. This concern arose from the differing temporal range s of the data set. Th at is, some collections crania represented individuals born throughout the 19th century, while others encompassed individuals born during the later pa rt of the 20th century. Ordinar ily, this slight time difference would have been negated as all the crania repr esented contemporary human s. However, a recent study noted secular changes in human cranial morphology in terms of height and facial width (Jantz, 2000). Since my study focuses on issues related to subtle faci al changes over a short amount of evolutionary time (i.e., the lifespa n of a specimen), the presence of secular transformations over the last 100 years could ha ve unduly affected my results. Therefore, I deemed it necessary to establish whether a co llection effect (i.e., temporal differences) manifested in my data. I tested this hypot hesis via a MANOVA on all 16 curves measured in this study for each racial population in terms of co llection effect and its in teraction with sex and age. I examined collection effect for each racial group because, in addition to temporal differences, each skeletal collect ion was racially unbalanced. Sp ecifically, the African-American specimens came almost entirely from the Terry Collection due to their prevalence across all the age groups. Meanwhile, I had to cull the European-American specimens, particularly in terms of females, from several different colle ctions due to the widespread issu e of edentulism (Chapter 5). Once divided racially, I found it was unnecessary for me to conduct a MANOVA on the African64

PAGE 65

American individuals since with one exception; they all originat ed from the Terry Collection. More importantly, though, I found there were no t enough degrees of freedom to perform the analysis, particularly for the interaction e ffects. The MANOVA results for the EuropeanAmerican specimens, however, indicated no signi ficant collection effect. Furthermore, no collection*age or collection*sex interactions were found in this population (Table 6-1). This would suggest that the secular changes noted by Ja ntz (2000) do not apply to facial curvature. This apparent contradiction may be linked to diff erences in the type of measurements taken (i.e., linear distances versus semilandmarks). Sp ecifically, semilandmarks retain the spatial relationship between shape variab les and capture more of a structures geometric shape than linear distance measures. Perhaps, then, th e secular changes noted by Jantz (2000) are eliminated when craniofacial skelet on shape is examined at a finer scale. In light of these results, I felt confident in poolin g the collections for each respectiv e age group and racial population. Dental State in the Elderly Age Group The second issue I addressed prio r to my main data analysis was the dental state of my elderly population. During refinement of my speci men inclusion criteria (C hapter 5), I decided to exclude edentulous elderly specimens on the premise that dramatic alveolar remodeling may alter surrounding curvature. To evaluate the valid ity of this premise, I collected curvature data on an additional 40 completely edentulous el derly African-American and European-American crania (10 males and 10 females within each group) and compared this data set with 40 randomly selected African-American and European-American elderly crania used in the main data analysis. A MANOVA was performed on the PC scores for the GPA-adjusted semilandmark coordinates for these two groups to test whether or not they differed in curvature shape. Only the curves representing semilandmarks from the eye orbits and zygomatic arches were utilized. When the regions were analyzed both together and separately (i.e., ey e orbits and zygomatic 65

PAGE 66

arches), statistically significant differences were detected between the ed entulous and dentulous groups. However, no significant two-way (dental st ate*race and dental st ate*sex) or three-way interactions (dental state*race*sex) were detected for any of the regions (Tables 6-2 through 64). These results not only validated to my decisi on to avoid mixing dental states for the elderly age group, but suggest tooth loss dramatically alte rs not only the shape of the maxillary ridge, but the shape of surroun ding bony structures. Main Data AnalysisMANOVA As indicated in Chapter 5, the facial skelet on was examined at three different levels: globally, regionally, and locall y. A MANOVA was conducted on the PC scores for the GPAadjusted semilandmark coordinate s at all three levels. Regardless of the level of analysis, sex and race main effects were always significant. I have highlighted the significant p-value results in the tables at the end of th e chapter. Significant p-values pertaining to age differences are highlighted in blue, while those involving only ra ce and/or sex differences are highlighted in yellow. As the goal of this research is to expl ore changes in facial skeleton shape with age, I only discuss results involving age differences and/or its interac tion with race and/or sex. Global Effects The African-American and European-American populations were divi ded into three age groups, containing approximately equal numbers of males and females: young adult (18-39 years), middle-aged (40-59 years), and elderly (60+ years). Appendices A-1 and A-2 provide demographic specifics for the number of indivi duals in each age group. Of the 684 individuals used in this research, 193 had all 16 curves available for analysis. A MANOVA conducted on the PC scores for these individuals revealed a si gnificant age effect, as well as a significant age and race interaction (Table 6-5). As these results validated my overarching hypothesis, I 66

PAGE 67

continued performing multivariate analyses in order to evaluate my hypotheses relating to regional and local age effect s on the facial skeleton. Regional Effects When viewed independently of the overall face, I found that the eye orbits (n=664; curves 3-4 and 11-12) also display a significant age effect, as well as significant sex*age and race*sex*age interactions (Table 6-6). The zygomatic arches (n=641), represented by curves 5-8 and 13-16, demonstrate significant age and race*age effects (Table 6-7), while the maxillary ridge (n=303; curves 1 and 9) exhibits a significa nt race*age interaction (Table 6-8). Finally, I detected no significant main or interaction effects involving age in the nasal aperture (n=393; curves 2 and 10; Table 6-9). Local Effects The curves constituting the maxillary ridge and nasal aperture represent these areas both regionally and locally. The eye orbits and z ygomatic arches, on the other hand, are created by multiple curves encompassing multiple bones. As such, I deconstructed these areas further into the individual curves and, by consequence, indi vidual bones which associate together to form each region. To determine whether or not the individual bone s which constitute the eye orbits (frontal bone =curves 3 and 11; maxillary bone =curves 4 and 12) display age-related shape differences, I analyzed the superior and inferior orbital curves independently. The supe rior orbital rim (n=667) exhibits a significant age effect, as well as a race*s ex*age interaction (Table 6-10). The inferior orbital rim (n=669) also demonstr ates a significant age effect, but no age-related interactions (Table 6-11). In an effort to determine whether or not ag e-related effects could be gleaned locally in zygomatic arches, I broke this region down into individual curves (i.e., curves 5-8 and 13-16) 67

PAGE 68

representing the zygomatic bone ( outer orbital and infe rior zygomatic curves) and the temporal bone (superior and inferior temporal curves ). Interestingly, I only found significant age differences in the inferior temporal curve shap e in terms of a race*age interaction (Tables 6-12 through 6-15). Main Data AnalysisPost-Hoc Tests The MANOVA results discussed above support my hypotheses outlined in Chapter 5. However, a unique pattern of age-related effects ma nifests at each level of analysis. Not all the areas under investigation displayed age-related di fferences in shape and of the ones that did, no two exhibited exactly the same se quence of main and interaction effects. In light of these complex patterns, I attempted to tease apart th e age-related changes farther by examining which age groups were contributing to th e aforementioned age effects at th e regional level (i.e., in terms of the eye orbits, zygomatic arch es, and maxillary ridge). Each regional data set was split in terms of the age effect detected. For instance, if a significant race*age interaction was found, the data would be separated into Europeanand African-Americans. Multiv ariate contrast tests within MANOVA were then performed on the PC scores to compare the means between the age groups (i.e., young, middle-aged, and elderly adult means) for each population. To better appreciate the relationship(s) between the age groups, I also performed a canonical variates analysis (CVA) on each regions PC scores and plotted the canonical vari ate scores. Finally, these regional differences were visualized in te rms of spatial differences in bony curvature using vector plots which compared the mean shapes between age pairings which were found to be significantly different. Eye Orbits As a significant race*sex*age interaction was detected in the eye or bits, I separated the data set for this region into European-American females, European-American males, African68

PAGE 69

American females, and European males. Contrast tests found several of th e age pairings to be significantly different from one another for al l of the subpopulations except African-American females (Tables 6-16 through 6-19). The following explores the age group differences detected within the other three subpopulations. African-American male. The young adult eye orbits we re different from the elderly orbits in African-American males (Table 6-17). To visualize this relationship, I performed a canonical variates analysis on the PC scores of this data set an d plotted the resultant canonical variates scores (Figure 6-1). Seventy-one percent of the variance was accounted for on the first canonical variate axis, while the second canonical variate axis repr esented the remaining 29% of variance. It is clear from the plot that the middle-aged (red square) mean lies between the elderly (blue diamond) and young adult (yellow triangle) means on the first canonical variate (x-axis). Thus, the young and elderly data points exist farther apart from one another spatially, than any of the other age comparisons. This is consistent with the contrast test results, wherein the young and elderly African-American male orbits ar e the most different from one another. Figure 6-2 shows vector plots of the mean orbital configurations for young (yellow spheres) and elderly (gray spheres) African-American males in the anterior and left lateral views. As a visualization device, the vector plot take s the landmark coordinates of one configuration (reference), in this case the young adult mean, and draws vectors from the landmark locations of that configuration onto the homologous landmark lo cations of a second configuration (e.g., target = elderly adult mean). The red arrows extending from the points are the difference vectors (i.e., directional vectors of ch ange) from the reference to the target To aid in the visualization of these subtle differences, the vectors have been ma gnified by a factor of 14 for all subsequent eye orbit plots. 69

PAGE 70

This overlay illustrates a complex pattern of shape change between young and elderly African-American males in terms of the eye orbits Visible yellow sphere s indicate that these particular semilandmarks are pos itioned more anteriorly in the young adult than the elderly configuration. In instances where only the elde rly gray sphere is vi sible, the corresponding young yellow sphere is located behind it and is thus obstructed from view. While admittedly some asymmetry is present betwee n the orbits, overall it appears as if superoinferior compression is occurring along the medial portion of the eye orbits along with some la teral expansion at the lateral orbital borders in the tran sition from young to elderly. Furthermore, the left lateral view of the left orbit indicates that the inferior orbit in oriented more posteriorly in the elderly adults. European-American male. Contrast test results for the European-American male eye orbits indicate that the young a dult orbit is significantly different from both the middle-aged and elderly eye orbit (Table 6-18). A CVA plot of this data set mirrors this relationship (Figure 6-3). The first canonical axis, representi ng 65% of the variance, separa tes the young adult orbits from the middle-aged and elderly orbits. The means for both the middle-aged and elderly eye orbits group closely together at the exclusion of the young adult mean. Figure 6-4 shows vector plots of the mean orbital configurations for young (yellow spheres) and middle-aged (blue spheres) individuals, in the anterior and left lateral views. The young adults serve as the reference and middle-aged adults as the target. Overall, there again appears to be superoinferior compression of th e medial orbits. However, unlike the pattern witnessed in the African-American male orbits, the lateral margins of the eye orbits are oriented more posteriorly with middle-age, while the medi al margins are oriented more anteriorly. In addition to the superoinferior compression, the superi or rim is oriented more anteriorly, while the inferior rim is oriented more pos teriorly with middle-age. Figure 6-5 displays vector plots of the 70

PAGE 71

mean configurations for young and elderly in the anterior and left lateral views. The young adults serve as the reference and elderly adults as the target. Instead of superoinferior compression, this transition is marked by a superoinferior expansion. Furthermore, the medial border is oriented more anteriorly, while the lateral border is posi tioned more posteriorly in the elderly adults compared to the young adults. European-American female. Finally, contrast tests performed on the EuropeanAmerican female orbits indicate significant di fferences exist between the young and middle-aged and middle-aged and elderly group means (Table 6-19). A CVA plot (Figur e 6-6) of the data mirrors this pattern wherein more point overlap is evident between the young and elderly groups along the first canonical variate ax is (61% of variance) than the middle-aged groups. Thus, these two groups are positioned closer to one another than either is to the middle-aged group. A vector plot of the transition from young (refe rence) to middle-age (t arget) in EuropeanAmerican female orbits displays a relatively random pattern of directional change in terms of the vectors (Figure 6-7). However, a vector plot of the transition from middle-age (reference) to elderly (target) reveals a mediolateral compression of the eye orbits (Figure 6-8). The medial and inferior orbital rims are al so positioned more posteriorly, wh ile the lateral orbital rim is positioned more anteriorly in the elderl y adults compared to the young adults. Zygomatic Arches As the interaction between age and race was found to be significant for the zygomatic arches, this data set was partitioned into Eur opeanand African-American individuals for all subsequent data analyses. My investigation of the relationship between the age groups via contrast tests revealed no si gnificant differences between the age groups within the EuropeanAmerican zygomatic arches (Table 6-20). 71

PAGE 72

African-Americans. However, mean differences betw een age pairings were detected among the African-American zygomatic arches. In particular, sign ificant differences were found between the young and elderly adult means, as well as between the middle-aged and elderly adult means (Table 6-21). As such, the CVA was onl y performed on the African-American data set (Figure 6-8). Along the first canonical axis, wh ich accounts for 75% of the variance, the middleaged and young data points overlap with the m eans for each group lying relatively close to one another. The elderly data points, meanwhile, exhibit far less overlap with the other two age groups. This pattern concurs with th at found in the contrast tests. I visualized the regional differen ces noted in the contrast test s via vector plots. While I conducted the vector plot on both arches, I only depict the left arch in the left lateral and anterolateral views (Figure 6-9) to simplif y the visual interpretation. The young adult configuration is the reference, wh ile the elderly configuration is th e target. In this instance, the vectors have been magnified by a factor of 10. Based on the directional vectors of change, the anterior zygomatic arch exhibits superoinferior compression in the elderly adults compared to the young adults. In addition to this compression the an terior zygomatic arch is also positioned more laterally in the elderly adults while many of the points making up the superior and inferior borders of the posterior arch are positioned more medially. In Figure 6-10, I visualized the direction of change created by transforming the mi ddle-aged (reference) zygomatic arch into the elderly (target) zygomatic arch. Again, the vector s have been magnified by a factor of 10. This age transition is marked by overall disorganization, with no clear directional pattern manifesting along any of the curves. Maxillary Ridge In light of the significant in teraction between age and race in this region, the data was partitioned into Europeanand African-American populations (Table 6-22 and 6-23). Contrast 72

PAGE 73

tests on these two populati ons only detected significant mean age group differences within the European-American maxillary ridge. European-Americans. Specifically, the mean shape of the middle-age maxillary ridge is significantly different from both the young and elderly European-American maxillary ridge (Table 6-23). A CVA plot of the European-American maxillary ridge data set indicated that the first canonical variate, which captured 63% of the variance, al so separated the young and elderly data points from the middle-a ged points (Figure 6-11). Vector plots of the young (reference) and middle-aged (target) maxillary ridge semilandmarks magnified by a factor of 14 found posterior recession of th e anterior ridge with associated lateral expans ion of the lateral borders in the su perior and anterior views (Figure 612). In the vector plots of th e middle-aged (target) and elderly (reference) maxillary ridge, the midpoint (i.e., prosthion) and the endpoints (i.e., ectomolare) of the maxillary ridge project more outwards and slightly down in the elderly c onfiguration. The remaining semilandmarks are positioned more medially in the elderly than the middle-aged population (Figure 6-13). Shape Patterns in the Facial Curvature The aforementioned MANOVA results confirm my hypotheses wherein the facial skeleton does display age-related shape changes globall y, regionally, and locally. These age-related shape changes involve various combinations of age as a main effect or as part of a two and threeway interaction with race and sex. Regiona lly, I observed significan t age-related shape differences in the eye orbits, zygomatic arches, and maxillary ridge, but no t the nasal aperture. Locally, both the superior and inferior orbital rims exhibited significant age-related differences in shape. Of the four curves compromising the zygomatic arch, only the in ferior temporal curve (i.e., inferior border of the zygomatic process of the temporal bone) disp layed a significant age effect. 73

PAGE 74

The age groups contributing to these significant age effects were evaluated at the regional level. In the eye orbits these age group di fferences existed between subpopulations (i.e., African-American male and European-American male and female). No significant age group differences were detected in the African-American female orbits data set. When I graphed these significant differences in terms of spatial change for the other three s ubpopulations an overall pattern of superoinferior compression often manifest ed in the eye orbits w ith age, but instance of superoinferior expansion and mediolater al compression were also detected. The zygomatic arch exhibited age group differences in shape at the leve l of race. While no significant age-related differences were detected in the European-American zygomatic arches, age group differences were found in the African-Ame rican zygomatic arches. The superoinferior compression pattern witnessed in many of the age comparisons for the eye orbits was also found in the transition from young to elderly in the Af rican-American anterior zygomatic arch, but was not present in the transition from middle-aged to elderly. In the maxillary ridge, age-related change al so existed at the racial level with only the European-American maxillary ridge exhibiti ng significant age group differences. In the transition from young to middle-aged adult this change was marked by recession anteriorly and expansion laterally, while in the transition from middle-age to elderly this change was primarily represented by inward recession alon g the maxillary border. In orde r to contextualize the results I observed here, I will now explore the impli cations and underlying factors which may be responsible for these gross external changes in facial bone morphology over time. 74

PAGE 75

Table 6-1. MANOVA output for the evaluation of collection effect Effect F-Value DF P-value collection 1.42 108 0.1034 sex 2.75 36 0.0267 age 0.92 72 0.6238 collection*sex 1.55 72 0.1048 collection*age 0.88 180 0.758 sex*age 0.95 72 0.5857 Table 6-2. MANOVA output for the evaluation of dental state in the eye orbits and zygomatics arches Effect F-value DF P-value race 5.21 36 <0.0001 sex 4.42 36 <0.0001 teeth 4.41 36 <0.0001 race*sex 2.17 18 0.0126 race*teeth 0.76 36 0.793 sex*teeth 1.18 36 0.3166 race*sex*teeth 0.77 36 0.7764 Table 6-3. MANOVA output for the evaluation of dental st ate in the eye orbits Effect F-value DF P-value race 2.19 22 0.0119 sex 2.88 22 0.001 teeth 2.3 22 0.0075 race*sex 1.13 22 0.3481 race*teeth 0.86 22 0.6358 sex*teeth 1.07 22 0.4031 race*sex*teeth 0.15 22 0.3291 Table 6-4. MANOVA output for th e evaluation of dental stat e in the zygomatic arches Effect F-value DF P-value race 5.35 30 <0.0001 sex 3.61 30 <0.0001 teeth 2.81 30 0.0012 race*sex 3.68 30 <0.0001 race*teeth 0.74 30 0.804 sex*teeth 1.07 30 0.4119 race*sex*teeth 0.72 30 0.8185 75

PAGE 76

Table 6-5. MANOVA output for the entire face Effect F-value DF P-value race 7.95 57 <0.0001 sex 3.81 57 <0.0001 age 1.32 114 0.0364 race*sex 1.61 57 0.014 race*age 1.45 114 0.0083 sex*age 1.07 114 0.3211 race*sex*age 0.95 114 0.607 Table 6-6. MANOVA output for the eye orbits Effect F-value DF P-value race 13.18 32 <0.0001 sex 7.83 32 <0.0001 age 1.98 64 <0.0001 race*sex 2.57 32 <0.0001 race*age 1.06 64 0.351 sex*age 1.43 64 0.0156 race*sex*age 1.65 64 0.0012 Table 6-7. MANOVA output for the zygomatic arches Effect F-Value DF P-value race 19.2 48 <0.0001 sex 11.54 48 <0.0001 age 1.51 96 0.0015 race*sex 2.67 48 <0.0001 race*age 1.39 96 0.0096 sex*age 0.94 96 0.7385 race*sex*age 1.05 96 0.3593 Table 6-8. MANOVA output for the maxillary arch Effect F-value DF P-value race 1.65 18 0.0489 sex 2 18 0.013 age 1.22 36 0.1841 race*sex 1.13 18 0.3254 race*age 1.56 36 0.0222 sex*age 1.05 36 0.3908 race*sex*age 0.69 36 0.9161 76

PAGE 77

Table 6-9. MANOVA output for the nasal aperture Effect F-value DF P-value race 30.98 15 <0.0001 sex 5.09 15 <0.0001 age 1.16 30 0.2524 race*sex 1.36 15 0.1628 race*age 0.88 30 0.6568 sex*age 1.29 30 0.1378 race*sex*age 0.87 30 0.6661 Table 6-10. MANOVA output for the superior orbital rim Effect F-value DF P-value race 10.64 21 <0.0001 sex 10.47 21 <0.0001 age 1.74 42 0.0027 race*sex 2.41 21 0.0002 race*age 0.95 42 0.4548 sex*age 1.37 42 0.0586 race*sex*age 1.46 42 0.0306 Table 6-11. MANOVA output for the inferior orbital rim Effect F-value DF P-value race 17.13 19 <0.0001 sex 7.41 19 <0.0001 age 2.1 38 0.0001 race*sex 2.73 19 0.0001 race*age 0.87 38 0.6894 sex*age 1.28 38 0.118 race*sex*age 1.22 38 0.1724 Table 6-12. MANOVA output fo r the outer orbital curve of the zygomatic arches Effect F-value DF P-value race 14.88 17 <0.0001 sex 11.09 17 <0.0001 age 1.43 34 0.0541 race*sex 1.01 17 0.4472 race*age 1.37 34 0.0787 sex*age 0.69 34 0.9076 race*sex*age 0.93 34 0.5879 77

PAGE 78

Table 6-13. MANOVA output for the superior temporal curv e of the zygomatic arches Effect F-value DF P-value race 28.31 13 <0.0001 sex 12.35 13 <0.0001 age 1.4 26 0.0891 race*sex 2.17 13 0.0096 race*age 1.5 26 0.0505 sex*age 1.25 26 0.2547 race*sex*age 0.71 26 0.8536 Table 6-14. MANOVA output for the inferior zygomatic curve of the zygomatic arches Effect F-value DF P-value race 17.14 16 <0.0001 sex 6.16 16 <0.0001 age 1.27 32 0.1429 race*sex 2.64 16 0.0005 race*age 1.07 32 0.3676 sex*age 0.91 32 0.6106 race*sex*age 1.03 32 0.4287 Table 6-15. MANOVA output for the inferior temporal curve of the zygomatic arches Effect F-value DF P-value race 25.96 20 <0.0001 sex 9.94 20 <0.0001 age 1.3 40 0.1028 race*sex 2.21 20 0.0018 race*age 1.81 40 0.0016 sex*age 0.57 40 0.9852 race*sex*age 1.33 40 0.0818 Table 6-16. Age group contra st tests for African-American female eye orbits Effect F-value DF P-value Young vs Middle-aged 1.55 28 0.0547 Young vs Elderly 1.42 28 0.1003 Middle-aged vs. Elderly 1.37 28 0.1237 78

PAGE 79

Table 6-17. Age group contra st tests for African-American male eye orbits Effect F-value DF P-value Young vs Middle-aged 1.42 30 0.0884 Young vs Elderly 2.59 30 <0.0001 Middle-aged vs. Elderly 1.42 30 0.0585 African-American Male Eye Orbits -4 -2 0 2 4 -4 -2 0 2 4 Can1Can2 Elderly Middle-aged Young Elderly Mean Middle-aged Mean Young Mean Figure 6-1. Canonical variates plot of the African-American male eye orbits. Figure 6-2. Vector plot of African-American male eye orbit: young to elderly vectors of change. Yellow spheres= Young adult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Anterior view B) Left lateral view. 79

PAGE 80

Table 6-18. Age group contra st tests for European-American male eye orbits Effect F-value DF P-value Young vs Middle-aged 1.94 29 0.0058 Young vs Elderly 1.9 29 0.0078 Middle-aged vs. Elderly 1.16 29 0.2778 European-American Male Eye Orbits-4 -2 0 2 4 -4 -2 0 2 4 Can1Can2 Elderly Middle-age Young Elderly Mean Middle-age Mean Young Mean Figure 6-3. Canonical variates plot of the European-American male eye orbits. Figure 6-4. Vector plot of European-American male eye orbit: young to middle-aged vectors of change. Yellow spheres= Young adult; Blue sphe res =Middle-aged adult. Red arrows indicate direction of change. A) Anterior view. B) Left lateral view. 80

PAGE 81

Figure 6-5. Vector plot of European-American male eye orbit: young to elderly vectors of change. Yellow spheres= Young adult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Anterior view. B) Left lateral view. Table 6-19. Age group contra st tests for European-Ame rican female eye orbits Effect F-value DF P-value Young vs Middle-aged 1.63 27 0.0415 Young vs Elderly 1.5 27 0.0731 Middle-aged vs. Elderly 2.1 27 0.0038 European-American Female Eye Orbits -4 -2 0 2 4 -4 -2 0 2 4 Can1Can2 Elderly Middle-aged Young Elderly Mean Middle-aged Mean Young Mean Figure 6-6. Canonical variates plot of the European-American female eye orbits. 81

PAGE 82

Figure 6-7. Vector plot of European-American female eye orbit: young to middle-aged vectors of change. Yellow spheres= Young adult; Blue spheres =Middle-aged adult. Red arrows indicate direction of change. A) Anterior view. B) Left lateral view. Figure 6-8. Vector plot of Eur opean-American female eye orbit: middle-aged to elderly vectors of change. Blue spheres= Middle-aged a dult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Anterior view. B) Left lateral view. 82

PAGE 83

Table 6-20. Age group contra st tests for European-American zygomatic arches Effect F-value DF P-value Young vs Middle-aged 1.21 47 0.1779 Young vs Elderly 0.76 47 0.8671 Middle-aged vs. Elderly 1.37 47 0.0661 Table 6-21. Age group contra st tests for African-Ame rican zygomatic arches Effect F-value DF P-value Young vs Middle-aged 1.34 47 0.0807 Young vs Elderly 2.81 47 <0.0001 Middle-aged vs. Elderly 1.59 47 0.0123 African-American Zygomatic Arches-4 -2 0 2 4 -4-3-2-101234 Can1Can2 Elderly Middle-aged Young Elderly Mean Middle-aged Mean Young Mean Figure 6-8. Canonical variat es plot of the African-A merican zygomatic arches. 83

PAGE 84

Figure 6-9. Vector pl ot of African-American zygomatic arch: young to elderly vectors of change. Yellow spheres= Young adult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Anterior view. B) Anterolateral view. Figure 6-10. Vector plot of African-American z ygomatic arch: middle-aged to elderly vectors of change. Blue spheres= Middleaged adult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Left lateral view. B) Anterolateral view. 84

PAGE 85

Table 6-22. Age group contrast tests for African-American maxillary ridge Effect F-value DF P-value Young vs Middle-aged 0.67 18 0.8335 Young vs Elderly 1.11 18 0.3512 Middle-aged vs. Elderly 1.24 18 0.236 Table 6-23. Age group contrast tests fo r European-American maxillary ridge Effect F-value DF P-value Young vs Middle-aged 1.83 17 0.0352 Young vs Elderly 1.45 17 0.1307 Middle-aged vs. Elderly 2.08 17 0.0135 European-American Maxillary Ridge-4 -2 0 2 4 -3-2-1 0 1 2 3 Can1Can2 Elderly Middle-aged Young Elderly Mean Middle-aged Mean Young Figure 6-11. Canonical variates plot of the European-American maxillary ridge. 85

PAGE 86

86 Figure 6-12. Vector plot of European-Ameri can maxillary ridge: young to middle-aged vectors of change. Yellow spheres= Young adult; Blue spheres =Middle-aged adult. Red arrows indicate direction of change. A) Superior view. B) Anterior view. Figure 6-13. Vector plot of European-American maxillary ridge: middle-aged to elderly vectors of change. Blue spheres= Middle-aged a dult; Gray spheres =Elderly adult. Red arrows indicate direction of change. A) Superior view. B) Anterior view.

PAGE 87

CHAPTER 7 DISCUSSION Humans represent approximately six million years of hominid, and more than 65 million years of primate/mammalian evolution. We owe our current existence to both our ancestors failures and successes. Unlike our genetic c ousins, however, much more of our daily understanding of humanity is situ ated in the intersection between biology and culture. That is, our mate selection, child-rearing, and quality of li fe have been, and continue to be, influenced not only by a biological need to survive and reproduce, but also by social taboos and norms. In many respects, our outward physical appearance crafted as it is by genetic expression is appraised by the social world in which we exist. As a result, the determin ation of what genes we pass onto future generations is affected as much by Darwinian notions of evolution as it is by the social rules to which we comport ourselves. Un til very recently, the judgment of a suitable mate (both culturally and biologically) was made more often than not, on the appearance of a young body as most humans did not live long enough for old age (i.e., > 70 years) to manifest. This recent longevity of Homo sapiens is a unique manifestation in our evolutionary tree. For instance, it is estimated that ov er the span of 280,000350,000 generations, Australopithecines and early Homo species lived an average of 15-20 years (Cutler, 1976; Weiss, 1984; 1989). And, more recently, early agriculturists and nomadic pastoralists lived only about 25 years (Weiss, 1984; 1989; Crews, 2003). Yet w ith technological and medical advances, the stress of daily life began to recede; harsh environments were controlled and manipulated; nutrition improved; and disease loads diminished. Slowly but surely the choke hold of limited life was loosened. 87

PAGE 88

Since the Industrial Revolution (late 1800s), the trend in U.S. life expectancy has been one of gradual improvement. At the turn of the tw entieth century, European-Americans lived into their late forties/early fifties, while non-Eur opean-Americans lived into their early thirties (Erhardt and Berlin, 1974). By the middle of the 20th century, roughly 88% of Americans across multiple ethnic/racial categories survived to see th eir fiftieth birthday. And, by the close of the last century, only an unlucky few did not celebra te their seventieth birthday (Anderson, 2001; Crews, 2003). For the first time in human history, then, we ar e able to witness the widespread effects of long life in the faces around us. No longer is th e face of advancing age a societal aberration, but instead a typical part of our day to day existen ce. Moreover, the age-re lated shape changes we witness in this region are not simply the product of soft tissue interactions, but, as my research and others indicate, encompass the hard tissue layer as well. Indee d, facial aging is far more than skin deep. Level of Analysis By applying novel geometric mo rphometric techniques to the facial skeleton at various gross levels, I was able to quant itatively detect subtle morphological changes in facial curvature with age. In addition to an e xpansion of the work by Pessa et al. (2007) via extending curvature alteration upwards into the facial skeleton, the results presented here support the long standing, but often overlooked belief in age-related craniof acial research that morphological change occurs throughout life. More precisely, I found a complex model of curvature cha nge wherein the facial skeleton displays age-related shape differences as a whole, regionally, and locally. These results support my research hypotheses outlined in Ch apter 5. Yet at each level, a unique pattern manifests. For instance, the entire adult face (e ye orbits, zygomatic arch es, nasal aperture, and maxillary ridge) exhibits statistically significant sh ape differences in relationship to age. But, 88

PAGE 89

when broken down into individual regions, the zygomatics arches, eye orbits, and maxillary ridge display shape variation with age, while th e nasal aperture does not. Furthermore, when viewed locally, the individual curv es characterizing the superior a nd inferior orbital rims show shape divergence with age, but onl y the inferior temporal curve (i nferior border of the posterior zygomatic arches) exhibits such change. The divergent nature of mo rphological change in certain regions of the facial skeleton serves as corroborating evidence of Hellmans (1927) concept of aging, wherein different areas of the face display different patterns of change with age. Indeed, the shifting contours we associate with the transition from yo uth to maturity, then, are not uniform, but in fact the coales cence of regional and local changes. Racial Differences in Facial Morphology with Age My research also demonstrates the importan ce of considering racial differences in any analysis of facial aging. At each structural level, I observed raci al variation. Globally, I detected a race and age interaction in the facial skeleton. I also noticed this pattern regionally in the maxillary ridge and zygomatics arches. Concomitantly, I observed a three-way interaction between age, race and sex in the eye orbits. Lo cally, I also found this th ree-way interaction in the superior orbital rim. Finally, of the curv es constituting the zygomatic arches, I detected a race and age interaction the inferior temporal curv e. As such, these results indicate there is no such thing as a common, and thus universal, ag ing process in the facial skeleton. Instead, the global, regional and local shape alterations I document here are racial averages, wherein human variation is condensed to a single form for each age group within each population/subpopulation. These abstractions, then, are applicable to ever yone within the re spective population (e.g., African-Americans) or subpopulat ion (e.g., African-American male s) yet reflective of no one being. While it would be inappropriate to use th ese results are predictors of individual facial 89

PAGE 90

shape change with age, they do highlight the complexity of the human condition across populations. As I further explored racial difference in facial bone curvature, comparisons between the age groups (e.g., young vs. middle-aged, young vs. elderly, and middle-aged vs. elderly) produced an interesting array of differences for e ach facial region. In pa rticular, I was able to discern different age pairing patterns between European-American males and females, and African-American males for the eye orbits. In the zygomatic arches, I only detected age group differences in shape within the African-American s. Finally, in the maxillary ridge I only found shape differences between the age groups in the European-Americans. These racial findings are in line with th e work of Harris and colleagues (1977) who suggested that craniofacial change was not only age, but race specifi c. This is no t surprising as the skull is one of the most diagnostic regions of the skeleton from which to derive race, which I treat here as a proxy for ancestry. Primarily, thes e determinations are co nfined to the facial skeleton. The morphology of the zygomatic arches, ey e orbits, and maxillary ridge, in particular, vary between individuals of African and European ancestry. In general, the European face is characterized by retreating, narrow zygomatics, wherein the cheek bones do not jut forward. This is in contrast to the Af rican face, which is marked by broader, more anteriorly-positioned zygomatic arches. The vault and associated faci al region is also broa der in individuals of African ancestry, thereby orienting the zygomatic arches farthe r apart from one another than in the European face (Rhine, 1990; Bass, 1998; Gill, 1998). Meanwhile, racial shape differences in the ey e orbits are qualitatively characterized by such adjectives as rectangular for individuals of African ancestry or sloped for individuals of European ancestry (Rhine, 1990). This region is also highly diagnostic of sexual dimorphism in 90

PAGE 91

terms of the prominence of the brow ridge (Buikstra and Ubelaker, 1994) and orbital rim shape (Rosas and Bastir, 2002; Pretoriu s et al., 2006), which may account for the three-wa y interaction between age, race, and sex detected in this region. Finally, the maxillary region is more prognathic among individuals of African than Eu ropean ancestry (Rhine, 1990; Bass, 1998). This genetic and, by consequence, morphologi cal durability in terms of ancestry/race predates advanced age in the evolution of huma ns, which perhaps explains why these differences are not masked by the relatively new evolutiona ry manifestation of long life. Thus, the morphological trajectory of the face and specifi cally the zygomatic arches, eye orbits and maxillary ridge varies between different ancestral groups (i.e., African-Americans and EuropeanAmericans) with age. While I do not explor e the underlying genetic and/or environmental factors responsible for the particul ar racial shape patterns witnessed with age in this analysis, the ability to quantitatively detect such relations hips highlights the importance of considering not only sexual dimorphism (which has been done previ ously), but also racial differences in facial aging. Indeed with every new ancestral population examined, we further our ability to characterize the human c ondition of facial aging. Visualizing Regional Shape Differences with Age Another major advancement achieved in this analysis is the transition from linear and angular measurement data to coordinate (x, y, an d z) data. Not only is shape information more fully captured during statistical an alyses, but the use of geometri c morphometrics allows me to illustrate the aforementioned age-related regional differences on the coordinate data points themselves using vector plots. This method is di rectly linked to the analysis of overall variation and is clearly sensitive to aspects of facial bone curvature, as it is capable of reflecting agerelated morphological change in the bony margins of the eye orbits, zygomatic arches, and maxillary ridge. While a specific pattern did not al ways manifest in these regional plots, the eye 91

PAGE 92

orbits did exhibit such shape changes as superoinferior expa nsion, superoinferior compression, and mediolateral compression depending on th e subpopulation and age groups analyzed. The zygomatic arches also display some superoinfe rior compression in the transition from young to elderly adult in African-American s. Finally, the European-Ameri can maxillary ridge generally exhibits expansion or recession of its border depending upon the age pairing under investigation. These alterations in facial form clear ly correspond to bone remodeling, as bone morphology is modified by bone cell deposition a nd resorption. However, such a simplistic answer is hardly intellectually satisfying as it does not speak to the underlying processes and mechanisms which stimulate this modification. Instead, I posit that th e principles of the functional matrix hypothesis may serve as a useful tool with which to hierarchically tease apart why this phenomenon occurs in the first place. Functional Matrix Hypothesis First introduced by Melvin Moss and Ri chard Young in 1960, the functional matrix hypothesis (FMH) offers a theory of craniofacial growth and devel opment which is epigenetic, as opposed to genetic, in origin. More precisely, extrinsic (e.g., mechanical loading and electromagnetic fields) and intrinsic (e.g., in tra-organismal, biophysical, and biomechanical events occurring within/between cells) factors alter skeletal morphology. In the most general sense, FMH treats the head and neck as a composite structure cons isting of independent functions including: digestion, vi sion, olfaction, and respiration. Each of these functions is carried out by its own respective functional cran ial component composed of a functional matrix (related cells, muscles, organs, spaces, etc.) wh ich performs the functi on; and a skeletal unit (bone, cartilage and tendinous tissue) which prot ects and supports the functional matrix (Moss and Young, 1960; Moss, 1962; 1968; 1969; Moss and Greenberg, 1967; Moss and Salentijn, 1969; Moss 1995; 1997a). 92

PAGE 93

Skeletal Units In FMH, skeletal units do not necessarily correspond directly to the bones of formal osteology. In reality, terms such as frontal or maxillary bone ar e biologically meaningless, as they are created by arbitrary divisions. Instead, Moss proposes that each bone should be thought of as a composite of contiguous skeletal units, kn own as micro-skeletal units. For instance, the maxilla is composed of nasal, pneumatic, orbita l, palatal and basal micro-skeletal units, each with their own unique functional role. Furtherm ore, macro-skeletal units consist of several adjacent bones, such as the endocranial surface of the calvarium which supports and protects the neural mass (Moss, 1962; 1968; 1972). This principle may explain the unique pattern of age-related change seen at each structural level of my study. For instance, both the zygoma tic arches and eye orb its display age-related shape alteration. Yet, when analyzed locally, th e individual curves (i.e., superior and inferior rims) of the eye orbits exhibit modification with age, while only the inferior temporal curve of the zygomatic arch displays age-related change. Given the complexity of the eye orbits in terms of protecting the brain internally, the eyes externally, and constituting a portion of the ethmomaxillary complex, it is not surprising that se veral skeletal units exist at various levels. The zygomatic arches, meanwhile, do not directly contribute to the prot ection of the brain and eyes and thus do not need to display the same degree of skeletal unit intricacy and complexity. Functional Matrices Size, shape, and positional changes in skelet al units are directly regulated by functional matrices under FMH. As such, growth and mainte nance of a skeletal unit depend almost entirely on its associated functional matrix (e.g., rela ted soft tissues and spaces) (Moss, 1962; 1968; 1971; 1995). For example, frontal bone shape on the outer table (i.e., cranial crest and ridges) is functionally related to the muscle s of the scalp, while inner tabl e morphology is sensitive to the 93

PAGE 94

dura mater and neural mass. Thus, growth in su pporting tissues, such as bone, is a secondary response to the primary growth of associated f unctional matrices (soft tissue and volumes) (Moss and Young, 1960). Furthermore, skeletal unit (bone) growth is accomplished in large part by bony resorption and deposition (i.e., remodeling), which produces large and small scale changes in size and shape. As such, it is not restricted to simply the period encompassing birt h to skeletal maturity, but is applicable to any phase of human life marked by craniofaci al alteration (Moss, 1962). For instance, the populations/subpopulations examined for each region in th is analysis often exhibited similar shape gradients in terms of the chronological positio ning of the age groups (Figures 6-1, 6-3, and 6-8). However, this pa ttern did not hold true to the European-American maxilla and European-American female eye orbits (Figures 6-7 and 6-11). These trends speak to the independence of the functional matrices encompassing the maxilla and eye orbits for European-Americans/ European-American females in terms of chronological shape change. Yet, they also highlight the fact that in general the independent regi onal functional matrices and their associated skeletal units act in unison chronologically to preserve functional unity/integrity in the human face. Thus, the aging process in the facial skeleton is typically characterized as gradual, persistent alteration over time. These results comp lement a multitude of age-related craniofacial research which suggests dimensional and structur al changes continue well into late life (e.g., Hrdlichka, 1936; Israel, 1969; Susanne, 1977; Be hrents, 1985; Pessa et al., 2007). Indeed, having established in this study that change in facial bone curvature (i.e., shape) extends well into adulthood, so too does craniofacial growth. Loading as an Epigenetic Factor Bone reacts to changes in epig enetic factors such as mechan ical stimuli (e.g., Moss, 1962; Rubin and Lanyon, 1985; Wolff, 1986; Rubin et al., 1990; Moss 1997a; 1997b; 1997c). As such, 94

PAGE 95

the morphology of each skeletal component is of tentimes influenced by its respective load history. Unlike the limb bones, however, the skul l does not display a single, predictable pattern of deformation. During stride, limb bones are compressed by body weight. Thanks to their curvature, they generally deform in bending, with one cortex in tension and the other in compression (Biewener, 2002; Herring and Ochare on, 2005). The predicability of this postcranial model allows for clear-cut modeling. In the skull, on the other hand, topological complexity in bone shape and thickness, multiple fibrous joint attachments, and varied muscle and occlusal loads prohibit creati on of a general model of deforma tion. Regional strain patterns are responsive to particular cont racting muscle combinations such as temporalis versus masseter and right versus left muscle (Herring and Oc hareon, 2005). Anteriorly, deformation is influenced by occlusal forces from the dental arcades, as well as forces arising from nearby muscles and joints. And, unlike the strain pattern noted in the limb bones, where bending commonly presents as longitudinal compression on one side and longitudinal tension on the other (Rafferty et al., 2000), shear or torsion is the most common pattern of strain produced in the craniofacial skeleton (H erring and Ochareon, 2005). Moreover, the regions and occasionally individu al bones of the skull vary in terms of the strain magnitude. For instance, the areas of the skull which serve as anchorage for the masticatory musculature (e.g., zygomatic arches, mandible) bear significant loads during biting or mastication (Behrents et al., 1978). On the other hand, areas such as the upper face and brow ridges experience comparably low levels of st rain during these same activities (e.g., Endo, 1966; 1970; Hylander et al., 1991; Ravosa et al., 2000; Peterson an d Dechow, 2003). Given the structural and functional complexity of the human skull, no single mechanical parameter is fully capable of predicting bone modification patterns. Large strains, for example, such as those 95

PAGE 96

initiated by rigorous physical activity (e.g., masti cation) produce microcracks and other damage to bone. This, in turn, stimulates osteoclasts and osteoblasts to remove and repair damaged tissue. While these high magnitude mechanical signals produce a bony response, the resulting microdamage also leaves these areas more prone to fracture (Rubin et al., 2006; Judex et al., 2007). Extremely low-magnitude, high frequency mech anical signals, on the other hand, are also capable of contributing to bone formation/maintena nce, with less risk of structural damage. Specially, small-scale, persistent mechanical sign als associated with pass ive muscle contraction (vibration) as one would produce with standin g, can serve also as an abolic signals to bone tissue, thereby controlling bone mass and mo rphology (Rubin and Lanyon, 1984; 1985; Rubin et al., 2001; 2006; Judex et al., 2007). Bone Responsiveness to Loading Though observed macroscopically, it has been post ulated that loadings are also imposed at other structural levels. In connected cellular network theory, cells are organized into layers: an initial input layer, intermediate la yer, and a final output layer. Each cell, in any layer, is capable of receiving several weighted input s. In the initial layer, thes e inputs represent loadings. The stimuli are summed and compared against a thre shold value (Wasserman, 1989; Parfitt, 1994). If exceeded, an intracellular signal is generated (i.e., mechanotransduction occurs) (French, 1992; Goldsmith, 1994; Parfitt, 1994; Hamill and Mc Bride, 1995; Moss, 1997a; 1997b; 1997c). Mechanotransductively activat ed bone cells (i.e., osteocytes) initiate membrane action potentials via interconnected gap junctions thereby allowing the signal to be transmitted to all the intermediate layers. This process of signal reception and transm ission continues hierarchically upward until the final layer of cel ls (i.e., osteoblasts) is reac hed. Osteoblast output, in turn, directs the rate, duration, and magnitude of the sp ecific response of associated osteoblast groups. 96

PAGE 97

Depending on the informational signal, this bony response may be deposition, resorption, and/or maintenance (Parfitt, 1994; Moss, 1997a; 1997b; 1997c). Macroscopic and potentially microscopic bone cell responsiveness, th en, is closely linked to extrinsic physical loading. Thus, regional changes in preexisting high and/or low stain patterns may explain bone shape alteration with ag e. In the craniofacial skeleton high strain magnitudes are primarily produced by masticati on and its associated muscles (masseter, temporalis, medial and lateral pterygoids). As the maxilla is a key component of the masticatory apparatus, it is directly subject to these high st rain occlusal forces (Herring and Ochareon, 2005). Furthermore, the masseter muscle originates from the lower border of the medial surface of the zygomatic arch (Moore and Agur 2002). Thus the maxillary ridge and zygomatic arches experience elevated loading patterns related to their resp ective roles in masticat ion, either as part of the masticatory complex or as a site for mas ticatory muscular attachme nt. Meanwhile, the eye orbits endure very low strains during biting or ma stication compared to these other two regions, but are nonetheless influenced by muscle contra ction. Several muscles of facial expression originate from the margins of the eye orbits (i.e ., occipitofrontalis, corr ugator supercilii, and orbicularis oculi) (Moore a nd Agur, 2002). Consequently, the underlying orbital bone experiences low strain from both passive muscle vibration, as well as facial expression-related contraction. As mentioned earlier, muscles serve as functi onal matrices under FMH. As such, dynamic and passive loadings produced by their contractio n can serve as epigenetic factors for bone cell responsiveness. Simply put, skeletal units are tu ned to their functional matrices. Thus, changes in muscle quality (functional matrix) would alte r loading patterns (epigenetic information), 97

PAGE 98

thereby modifying skeletal response (resorp tion, deposition, maintenance levels) and changing morphological form. Aging Muscle and the Functional Matrix Hypothesis Reductions in muscle quality have been positivel y linked to the aging process. In general, the human body is marked by a progressive di minishment of striated musculature and mechanical performance with age (Borken et al ., 1983; Porter et al., 1995 ; Gallagher et al., 1997; Rosenburg 1997; Hatch et al., 2001; Peyron et al., 2004). In the ma sticatory apparatus, both the masseter and medial pterygoid muscles exhibit ag e-related changes in the cross-sectional area and density. This musculature variation persists even when contributing factors such as systemic diseases, muscle pathologies and tooth absence are controlled (Newton et al., 1987; 1993; Galo et al., 2007). Age-related deterioration in the fa st and slow fibers of these striated muscles results in impaired muscle force. Consequently, alteration in muscle stru cture affects the pattern and effectiveness of masticatory movements. When analyzed against young subjects, elderly individuals show hyperactivity during dental posture maintenance and hypoactivity during chewing (Galo et al., 2006; 2007). Moreover, muscular activity of the masseter and temporalis muscles in various mandibular positions is hi gher in elderly dentate individuals than young dentates (based on percentages of maximum EMG values). This indicates that greater energy expenditure is required of the elderly to accomp lish the same effect (Alajbeg et al., 2006). Thus, the age-related shape changes in the ma xillary ridge and zygomatic arches may be explained by changes in high and possibly low strain patterns associated with masticatory muscle deterioration. Similarly, age-related muscle wa sting may also contribute to the shape changes witnessed in the eye orbits, yet in this region the b ony response is a conseq uence of changes in an exclusively low strain mechanical environmen t. The orbicularis oculi muscle, for example, exhibits hypertrophy with age which presents as hammock-like folds (festoons) of the muscle in 98

PAGE 99

the lower eyelid (Perkins and Batniji, 2005). This hypertrophy presumably alters the underlying bone. As the aging process invariably leads to muscle wasting, generalized reduction in muscle activity and by consequence strain magnitude may dilute or remove essential regulatory bone signals which stave off bone eros ion (Haung et al., 1999; Rubin et al., 2006). Thus, changes in strain environments, regardless of their initial magnitudes, th roughout the craniofacial skeleton likely play an important role in age-re lated change in facial bone shape. Other Contributing Functional Matrices However, it is important to remember that many different functional matrices and related epigenetic processes other than musculature ar e capable of evoking cellular and consequently gross modification (Jorgensen, 1994; Moehrle and Paro, 1994). For instance, spaces and fibrous tissues also constitute functional matrices (Moss and Young, 1960; Moss, 1962; 1995; 1997a; 1997b). As such, dimensional change in spaces su ch as the sinusal cavitie s may also contribute to the bony alterations witnessed in this study. The frontal si nuses, for example, abut the superomedial portion of the eye or bits internally. These cavitie s enlarge with age and exhibit dramatic individual variability (Fatu, 2006). Thus, dimensional sinusal changes may also contribute to the age -related alterations in eye orbit shape. Furthermore, tooth loss, in particular loss of the fibrous periodontal ligament, may play a role in the maxillary change detected herein, as disruption at the pe riodontal-alveolar bone interface serves as local stimulus for bone remo deling. While this disruption is harnessed by odontology to stimulate directed tooth movement (B ourauel et al., 2000), w ith tooth loss, this remodeling goes unchecked, leading to alveolar resorption. While admittedly I attempted to mitigate for tooth loss in terms of partial/completely edentulous individuals, I was not able to consistently utilize individuals with a full dent ition across the age groups. Thus, the middle-aged and elderly populations inevitably displayed more tooth loss than the young adult population. As 99

PAGE 100

such, changes in maxillary ridge shape with age ma y also be related completely or in part to destruction of the periodontal ligament with even the small levels of tooth loss documented in this analysis. In summation, with age, muscle structure diminishes. This, in turn, compromises the muscle forces generated by dynamic and passive muscle contraction. When this concept is interwoven with the principles of FMH, muscle deteriorati on (functional matrix) leads to alteration in craniofacial loading patterns (epigenetic fa ctors). As bone is t uned to the loading dynamics of muscle contraction, such deterior ation would predictably modify bone response leading to microscopic and macroscopic morphol ogical change. Furthermore, sinusal cavity expansion and periodontal ligament destruc tion with tooth loss may also alter the microenvironment of various skeletal units. Thus I speculate that muscle deterioration, sinus expansion, and/or periodontal ligament loss with age may serve as a driving force(s) in the remodeling patterns detected in this research. Absence of Nasal Aperture Change The absence of detectable age-related change in the nasal aperture is also intriguing. This phenomenon may be linked to the lack of muscle attachment sites along the nasal aperture margins. While the nasal aperture may indeed experience loading, it is not related to muscle contraction. Thus, this region experiences none of the bony consequences associated with agerelated muscle wasting seen elsewh ere in the facial skeleton. Howe ver, it is equally possible that not enough data was collected on th ese regions to detect change. The sample size for the nasal aperture was several hundred less than that for the eye orbits and zygomatic arches, given curation constraints. Increased sample sizes ma y then produce measurable shape difference with age. Another issue may be that my nasal aper ture measurements were simply not sensitive enough to detect age-related change. Perhaps th e addition of more semilandmarks would have 100

PAGE 101

101 identified measurable alteration. It is also c onceivable that the area se lected for each region is not the portion which undergoes ch ange. My nasal aperture curves did not actually capture the inferior nasal aperture margin, as a viable endpoint landmark (e.g., nasospinale) was not consistently present in this region due to curatio n damage and/or sectioning. Change may then be occurring along the nasal aperture, but is concentrated to the inferior margin. Limitations Finally, we cannot move forward without appreciating the gaps we have left in our wake. Most importantly, the cross-sectiona l nature of this research can re flect general changes in facial morphology, yet it is incapable of detecting subtle individual a lterations with time. Only longitudinal data can provide insight into th is phenomenon. Moreover, my large age groups, which represent roughly 20 years each, may potentia lly mask more fine grain alterations. The statistical effects of procedures such as GPA may also play a role in the results documented herein. For instance, when the zygomatic arches are deconstructed into individual curves much of the initial size information pertaining to the sp atial distance between the right and left arches and superior and inferior borders is lost. Loss of this size information, which is distinct from the scale information sequestered duri ng GPA, may have contributed to the lack of significant age effects when the zygomatic arch curves are analy zed individually. I beli eve, however, that these limitations in no way undermine the novelty of my research. Instead, th is study reflects how advances in data collection and analysis can take classic scientific enquiries into new and exciting directions. Finally, I conclude this study in the ne xt section by discussing potential directions in which this research can be expanded.

PAGE 102

CHAPTER 8 CONCLUSION The life expectancy of humans has dramati cally increased over the last 100 years, particularly in the industrialized world. This increase has been facili tated by better access to nutrition and overall advances in health care. Along with the benefit of extended life, though, we are faced with the social reality of a sharp increase in the number of elderly persons. The facial visage of the elderly will s oon become as prevalent in our society as young adult and middleaged faces. Adult facial aging, particularly in terms of the ha rd tissue, however, is a little understood process. This is problematic as it w ill impact both the medical community, as well as our daily social world. Gi ven this intertwined phenom enon, my study expands upon our understanding of if and how the bony architecture of the face ch anges. In particular, my study investigates the intera ction between chronological age and shape alterations in the facial skeleton. To achieve this goal, I inspected globa l, regional, and local sh ape differences within and between populations (i.e., Af rican-American and European-American) and the sexes with age via analysis of the bony curves of the face. These curves not only reflect the contours of the facial skeleton, but also allow for the applica tion of powerful three-dimensional morphometric techniques. In the course of my study, I found that facial skeleton curvature does alter over time at various levels. In addition to global shape change (i.e., encompassing the en tire facial skeleton), various regions of the facial skeleton (i.e., eye orbits, zygomatic arches, and maxillary ridge) also exhibit shape alterations with ag e. Locally, the superior and in ferior orbital curves and the inferior temporal curve also show significant age-related shape ch ange. When I visualized these changes spatially, diverse patterns manifested (e .g., superoinferior or mediolateral compression, lateral expansion, etc.) depending on the region analyzed and the age groups compared. In all 102

PAGE 103

likelihood the bone remodeling (i.e ., resorption and depos ition) responsible for this curvature change is guided by the conflu ence of various extrinsic and in trinsic factors. However, alterations in mechanical envir onment resulting from diminished musculature quality and sinusal cavity expansion with age, and/ or periodontal ligament destruction with tooth loss may be of upmost importance in stimulating this change While not the objec tive of this study, this research does open the door for further inves tigation into the source(s) of this change. Beyond serving as a foundation upon which to visu alize and quantify change in the facial architecture over time, my research also explores the ways in wh ich ancestry (i.e., race) interacts with age. Specifically, I found racial differences in terms of the shape of the zygomatic arches and maxillary ridge with age. Furthermore, the eye orbits displayed a three-way interaction between race, sex, and age. These results highl ight the important role biologically-influenced social differences such as race play in the manifestation of facial age and underscore how important it is for researchers to be cognizant of their study sample when attempting to speak about human aging. This is particularly apropos as classic research, especially w ithin the medical realm, often overlooks ancestral differences in age studies. Instead, documented change in relatively homogenous populations (i.e., European/European-American) is treated as reflective of the human condition, thereby reducing diversity to a single, universal entity. This reductionist view preemptively discounts biological variation, while inadvertently reasserting social Darwinism. Aging, to be sure, is by no means id entically manifested across human populations. Since this biased level of anal ysis overlooks fine-grain phenomenon which signify the norm for other racial populations, a single study group cannot serve as a pr oxy for the entirety of human existence. By expanding the notion of a single scientific norm into a diversity of norms, my 103

PAGE 104

research adds to an integrated model of f acial aging that incorporates multiple human populations. Future work. The real power of my research, however, lies in its methodological flexibility, particularly in terms of the que stions which can be posed. The method is bound neither to a particular society nor time period, and thus not c onstrained to contemporary groups originating within the c ontiguous United States. As such, my research is capable of conducting between group comparisons, as well as within group comparisons, from around the globe. For instance, this method can be used to explore the relationship betw een whiteness or blackness, and aging. Quite often terms such as white or European and black or African are employed when discussing the inhabitants of Eu rope and Africa respectiv ely. However, these terms often serve to figuratively flatten, or mi nimize, geographic diversity, which is especially problematic given the numerous nationalities and ethnicities encompassing these regions. Thus, my methods could be implemented on additional co llections to tease apart the degree to which such societal classifications of ethnicity and nationality are based in biological reality. A potential question to ask would be, Do subtle skeletal shape differences exist between individuals culturally-defined as say French and Austrian, or are these terms simply descriptors of origin of birth and thus biologically irrelevant? Cl osely related to this question is the notion that if such shape differences among these classifications exist, what role does the aging process play in terms of maximizing or minimizing said differences? Additionally, how does sexual dimorphism factor into this phenomenon? Furthermore, this research could be expanded to identify the shapes associated with youthfulness, and by extension, beauty. Mate se lection, both biologica lly and socially, for humans is linked to the notion of attractiveness. However, the definition of attractiveness 104

PAGE 105

varies both culturally and temporally. By expl oring our response to various regional and global facial curvature patterns, we might be able to quantify the combination(s) of facial shapes to which attractiveness is culturally and individually attributed. While such work would obviously be a tremendous contribution to the field of plastic surgery, the creation of averages which are populationally and chronologically appropriate w ould also greatly assist in reconstructive surgeries. By applying these techniques to patient CT scans, the phe nomenon of facial aging could also be extended into living popul ations instead of dry crania. Such contemporary material provides the most concise repres entation of present day facial form. And, given the sheer quantity and availability of such patient scans compared to skeletal material, not only could sample sizes be substantially increased, but so too could the diversity of the samples collected. Moreover, the forensic potential of this research is vast. By divers ifying the populations under analysis, classification norms could be generated capable of establishing age, as well as sex and ancestry for unknown skeletal material. And, as th e curves utilized in this study incorporate various regions of the facial sk eleton, classification criteria c ould be constructed using various combinations of available facial curvature. This, in turn, could be applied to fragmentary material lacking the large regi ons typically required for clas sic forensic anthropology linear analyses. Additionally, an exploration of the aging process in ar chaeological terms could be addressed. In particular, by examining modern humans through time, I could investigate the impact of life history on the manifestation of a ge. This would allow me to utilize not only contemporary human material, but archaeological ma terial from such sources as ancient Roman, early colonial, or slave cemeteries. This is a pa rticularly exciting prospe ct, given the dearth of 105

PAGE 106

historical documents available pe rtaining to living conditions of the non-elite class in Europe and North America. By incorporating the historical record, contemporary skelet al material could be compared to historic cemetery museum material to examine the impact of such factors as migration patterns, gene flow, subsistence strategies, famine periods, and disease on the morphology of the face at different ages. Finally, my work on facial aging strongly bene fits from being firmly rooted within the field of biological anthropology. Biological anthropology provides a perfect forum within which to address the aging process, because it inte grates notions of human evolution, adaptation, development, and variability. The bulk of re search examining age-related facial changes, particularly in recent years, originates primarily within th e clinical realm (i.e., odontology, reconstructive surgery, etc.). This field is principally tied to soft tissue alterations, and reductionist by nature, as its pr imary research objective is linked to improving patient surgical outcome. Within the field of biological anth ropology, though, I can address the relationship(s) between time, genetics, and the soci al world as it is manifested in facial bone arch itecture. This more holistic approach allows me to not only assess, but expand upon our current understanding of the aging process in the facial skeleton. In conclusion, the differences between human age groups in facial skeleton morphology documented herein highlight major shifts in the human condition throughout life. This information has always been encoded in our bodi es, yet until recently rarely had an opportunity to manifest. Throughout much of our existence, old age was the exception to the rule. As such, we have yet to witness widespread socio-cultural effects of protracted biological age on society. As increasing members of our society live far longer th an humans have previously, we are compelled to reevaluate not only what const itutes as old, but the cultural meaning behind 106

PAGE 107

107 the concept of old age. This presents a challe nge to us both socially and biologically. As such, biological anthropology stands at the forefront of how we examine the cultural impact(s) on our skeletons. This study, with its expansive da ta collection, novel data manipulation, and multivariate statistic analyses clearly demonstrat es how productive biological anthropology is at helping to unravel the tangled threads that exis t between human biology and human society.

PAGE 108

APPENDIX A AGE GROUP POPULATION DEMOGRAPHICS Table A-1. African-American age groups Age Male Female TOTAL Young Adult 88 79 167 Middle-Aged Adult 68 51 119 Elderly Adult 49 25 74 TOTAL 205 155 360 Table A-2. European-American age groups Age Male Female TOTAL Young Adult 50 51 101 Middle-Aged Adult 81 38 119 Elderly Adult 52 52 104 TOTAL 183 141 324 108

PAGE 109

APPENDIX B LANDMARK DEFINTIONS Table B-1. Landmark definitions Landmark Abbrev. Definition auriculare au the point vertically above the cen ter of the external auditory meatus at the root of the zygomatic process, a few millimeters above porion ectomolare ecm the most lateral point on the outer surface of the alveolar margins of the maxilla, often at the second molar position frontomalare orbitale fmo the point where the frontozygomatic suture cross the inner orbital rim frontomalare temporale fmt the most laterally positioned point on the frontozygomatic suture porion po the uppermost point on the margin of the external auditory meatus prosthion pr midline point at the most anterior point on the alveolar process of the maxillae subspinale ss the deepest point seen in the pr ofile below the anterior nasal spine zygomaxillare zm the most inferior point on the zygomaticomaxillary suture zygotemporal inferior zti the most superior point on the zygomaticotemporal suture zygotemporal superior zts the most inferior point on the zygomaticotemporal suture 109

PAGE 110

LIST OF REFERENCES Adams D, Rohlf F, Slice D. 2004. Geometric mor phometrics: ten years of progress following the 'revolution'. Ital J Zool 71:5-16. Akgul A, Toygar T. 2002. Natural craniofacial change s in the third decade of life: a longitudinal study. Am J Orthod Dent ofacial Orthop 122:512-22. Alajbeg I, Valentic-Peruzovic M, Alajbeg I, Cifrek M. 2006. The influence of age and dental status on elevator and depressor muscle activity. J Oral Rehabil 33:94-101. Anderson R. 2001. United States Life Tables, 199 8: National Vital Stat istics Report, 48. Hyattesville: NCHS. Andresen P, Bookstein F, Conradsen K, Ersboll B, Marsh J, Kreiborg S. 2000. Surface-bounded growth modeling applied to human mandi bles. IEEE Trans Med Imaging 19:1053-1063. August M, Kaban L. 1999. The aging maxillofaci al skeleton. In: Rosen C, Glowacki J, Bilezikian J, editors. The aging skelet on. San Diego: Academic Press. p 359-372. Ayers HG, Jantz RL, Moore-Jansen PH. 1990. G iles and Elliot race discriminant functions revisited: a test using rece nt forensic cases. In: Gill GW, Rhine S, editors. Skeletal attribution of race: methods for forensic anthropology. Maxwell Museum of Anthropology, Anthropological Papers 4. p 65-71. Bartlett S, Grossman R, Whitaker L, Ingervall B. 1992. Age-related changes of the craniofacial skeleton: an anthropometric and histologic analysis. Plas t Reconstr Surg 90:592-600. Bass W. 1998. Human osteology : a laboratory and field manual of the human skeleton. Columbia: University of Missouri Press. Behrents R. 1985. Growth in the aging craniofacial skeleton. Ann Ar bor: Center for Human Growth and Development: The University of Michigan. Behrents RG, Carlson DS, Abdelnour T. 1978. In vivo analysis of bone stra in about the sagittal suture in macaca mulatta during masticatory movements. J Dent Res 57:904-8. Bergstrom K, Lundberg P, Sjovall A. 1973. The volum e of the sella turcic a in patients with anorexia nervosa. Eur Neurol 9:183-189. Biewener A. 2002 Future directions for the anal ysis of musculoskeleta l design and locomotor performance. J Morphol 252:38-51. Bishara S, Treder J, Jakobsen J. 1994. Facial and dental changes in adulthood. Am J Orthod Dentofacial Orthop 106:175-186. Boersma J. 1974. Alterations of skull dimens ions in aged persons. J Dent Res 53:678-682. 110

PAGE 111

Bookstein F. 1982. Foundations of morphom etrics. Annu Rev Ecol Syst 13:451-470. Bookstein F. 1989. Principal warps: thin-plate splines and the decomposition of deformations. IEEE Trans Pattern Anal Mach Intell 11:567-585. Bookstein F. 1991. Morphometric tools for la ndmark data: geometry and biology. Cambridge: Cambridge University Press. Bookstein F. 1996. Standard formula for the uniform shape component in landmark data. In: Marcus L, Corti M, Loy A, Naylor G, Slice D, editors. Advances in Morphometrics. New York: Plenum Press. p 153-168. Bookstein F. 1997. Landmark methods for forms without landmarks: mo rphometrics of group differences in outline shape. Med Image Anal 1:225-243. Bookstein F, Schafer K, Prossinger H, Seidler H, Fieder M, Stringer C, Weber G, Arsuaga J, Slice D, Rohlf F, Recheis W, Mariam A, Marcus L. 1999. Comparing frontal cranial profiles in archaic and modern homo by mo rphometric analysis. Anat Rec 257:217-24. Bookstein F, Streissguth A, Sampson P, C onnor P, Barr H. 2002. Corpus callosum shape and neuropsychological deficits in adult males with hea vy fetal alcohol exposure. Neuroimage 15:233-51. Borkan G, Hults D, Gerzof D, Robbins A, Silbert C. 1983. Age changes in body composition revealed by computed tomography. J Ge rontol A Biol Sci Med Sci 38:673-677. Bourauel C, Vollmer D, Jaeger A. 2000. App lication of bone remodeling theories in the simulation of orthodontic tooth movements, J Orofac Orthop 61: 266. Broadbent B. 1937. The face of the normal child. Angle Orthod 7:209-233. Brodie A. 1940. Some recent observations on the growth of the mandible. Angle Orthod 10:6377. Brodie A. 1941. On the growth of the human head form the third month to the eighth year of life. Am J Anat 68:209-262. Buchi E. 1950. Anderungen der Korperform beim erwachsenen Menschen. Anthropologische Forschungen 1:1-44. Buikstra J, and Ubelaker D, eds. 1994. Standa rds for data collection from human skeletal remains. Fayetteville: Arka nsas Archaeological Survey. Carlsson G, Bergman R, Hedegard B. 1967. Changes in contour of the maxillary alveolar process under immediate dentures: a l ongitudinal clinical and x-ray cephalometric study covering 5 years. Acta Odontol Scand 25:45-75. 111

PAGE 112

Carlsson G, Persson G. 1967. Morphologic changes of the mandible after extraction and wearing of dentures. A longitudinal, clinical, and x-ray cephalometric study covering 5 years. Odontologisk revy 18:27-54. Chan G, Duque G. 2002. Age-related bone lo ss: old bone, new fact s. Gerontology 48:62-71. Clarke B, Ebeling P, Jones J, Wahner H, O Fallon W, Riggs B. 1996. Changes in quantitative bone histomorphometry in aging healthy men. J Clin Endocrinol Metab 81:2264-2270. Crews D. 2003. Human senescence: evolutionary and biocultural perspectives. United Kingdom: Cambridge University Press. Currey J, Brear K, Zioupos P. 1996. The effects of aging and changes in mineral contend in degrading the toughness of the human femora. J Biomech 29: 257-260. Cutler R. 1976. Evolution of huma n longevity. J Hum Evol 5:169-202. Dager M, McNamara JJ, Baccetti T, Franchic L. 2007. Aging in the craniofacial complex longitudinal dental arch changes through the sixth decade. Angle Orthod 78:440-444. deFroe A. 1936. Meetbare variabelen van de n menschelijken Schedel en hun onderlinge Correlaties in Verband met Leeftijd en Geslacht, Universiteit van Amsterdam. Diab T, Condon K, Burr D, Vashishth D. 2006. Ag e-related change in the damage morphology of human cortical bone and its ro le in bone fragility. Bone 38:427-431. Doual J, Ferri J, Laude M. 1997. The influence of senescence on craniofacial and cervical morphology in humans. Surg Radiol Anat 19:175-183. Endo B. 1966. Experimental studies on the mechan ical significance of the form of the human facial skeleton. J Fac Univ Tokyo III: 1-101. Endo B. 1970. Analysis of stre sses around the orbit due to masse ter and temporalis muscles respectively. Am J Anthropol Soc Nippon 78: 251-266. Erhardt C, Berlin J. 1974. Mortality and Morb idity in the United States. Cambridge: Harvard University Press. Farkas L, Eiben O, Sivkov S, Tompson B, Katic M, Forrest C. 2004. Anthropometric measurements of the facial framework in a dulthood: age-related changes in eight age categories in 600 healthy white North Ameri cans of European ancestry from 16 to 90 years of age. J Cran iofac Surg 15:288-98. Formby W, Nanda R, Frans Currier G. 1994. Longitudinal changes in the adult facial profile. Am J Orthod Dentofacial Orthop 105:464-476. 112

PAGE 113

Forsberg C. 1976. Growth changes in the adult face: a longitudinal roentgen cephalometric investigation on men and women in early a dulthood. Karolinska Institutet, Stockholm. Forsberg C. 1979. Facial morphology and ageing: a longitudinal cephalometric investigation of young adults. Eur J Orthod 1: 15-23. French A. 1992. Mechanotransduction. Annu Rev Physiol 54. Gallagher D, Visser M, De Meersman R. 1997. Appe ndicular skeletal muscle mass: effects of age, gender, and ethnicity. J Appl Physiol 83:229-239. Galo R, Vitti M, Mattos Mda G, Regalo SC. 2007. Masticatory muscular activation in elderly individuals during chewi ng. Gerodontology. 24:244-248. Galo R, Vitti M, Santos C, Hallak J, Regalo S. 2006. The effect of age on the function of the masticatory system-an electromyographi cal analysis. Gerodontology 23:177-182. Gill GW. 1998. Craniofacial criteria in the skelet al attribution of race. In: Reichs KJ editor. Forensic Osteology. Advances in Identification of Hu man Remains. 2nd ed. Springfield, IL: Charles C. Thomas. P 293-317. Goldsmith P. 1994. Plant stems: a possible model system for the transduction of mechanical information in bone modeling. Bone 15:249-250. Goldstein M. 1936. Changes in dimensions and form of the face and head with age. Am J Phys Anthropol 22:37-89. Goodall C. 1991. Procrustes methods and the statis tical analysis of shap e (with discussion). J Royal Statistical Soc B 53:285-340. Gower J. 1975. Generalized Procrustes analysis. Psychometrika 40:33-50. Guntz P, Mitteroecker, P Bookstein F. 2005. Sem ilandmarks in three dimensions. In: Slice D editor. Modern morphometrics in physical anthropology. New York: Kluwer Academic. p 73-97. Hamill O, McBride DJ. 1995. Mechanoreceptive membrane channels. Am Sci 83:30-37. Hammer O, Harper DAT, Ryan PD. 2008. PAST PAlaeontological ST atistics, Version 1.77. Free Software: http://folk.uio.no/ohammer/past Harris J, Kowalski C, LeVasseur F, Nasjleti C, Walker G. 1977. Age and race as factors in craniofacial growth and deve lopment. J Dent Res 56:266-274. Hatch J, Shinkai R, Sakai S, Rugh J, Paunovich E. 2001. Determinants of masticatory performance in dentate adults Arch Oral Biol 46:641-648. 113

PAGE 114

Heath M. 1966. A cephalometric study of the middle third of the adult human face related to age and the loss of dental tissue Arch Oral Biol 11:677-686. Hellman M. 1927. A preliminary study in development as it affects the human face. Dental Cosmos 69:250-269. Hellman M. 1932. Introduction of growth of the human face from infancy to adulthood. Int J Orthod 18:777-798. Herring S, Ochareon P. 2005. Bone-special prob lems of the craniofacial region. Orthod Craniofac Res. 8:174-82. Herring S, Teng S. 2000. Strain in the braincas e and its sutures during function. Am J Phys Anthropol 112:575-593. Herzberg F, Holic R. 1943. An anthropologic st udy of facial height Am J Orthod Dentofac Orthop 29:90-100. Hooten E, Dupertuis C. 1951. Age changes and sele ctive survival in Irish males. Studies in Physical Anthropology, Am Assoc of Phys ical Anthropol and Wenner-Gren Foundation 2:1-130. Horowitz S, Thompson R. 1964. Variations of th e craniofacial skeleton in postadolescent males and females. Angle Orthod 34:97-102. Hrdlicka A. 1925. The old American. Baltimore: Williams and Wilkins Co. Hrdlicka A. 1936. Growth during adu lt life. Proc Am Phil Soc 76:847-897. Huang RP, Rubin CT, McLeod KJ. 1999. Changes in postural muscle dynamics as a function of age. J Gerontol A Biol Sci Med Sci 54:352-357. Humphrey G. 1858. A treatise on the human skeleton. Cambridge: Macmillion and Company. Humphry G. 1864. On the growth of the jaws. Trans. of the Cambridge Phil. Society, Cambridge; zit. In: Brodie A. 1941. On the gr owth pattern of the human head. Am. J. Anat. 68: 209-262. Hunt DR, Albanese J. 2005. History and dem ographic composition of the Robert J. Terry anatomical collection. Am J Phys Anthropol. 127:406-417. Hylander W. 1975. The human mandible: lever or link? Am J Phys Anthropol 43:227-242. Hylander WL, Picq PG, Johnson KR. 1991. Masticat ory-stress hypotheses a nd the supraorbital region of primates. Am J Phys Anthropol 86:1-36. 114

PAGE 115

Israel H. 1969. Pubertal influence upon the gr owth and sexual differentiation of the human mandible. Arch Oral Biol 14:583-590. Israel H. 1973a. Recent knowledge concerning craniofacial aging. Angle Orthod 43:176-184. Israel H. 1973b. Age factor and the pattern of change in craniofacial structures. Am J Phys Anthropol 39:111-128. Israel H. 1977. The dichotomous pattern of cran iofacial expansion during aging. Am J Phys Anthropol 47: 47-52. Jantz R. 2001. Cranial change in Amer icans: 1850-1975. J Forensic Sci 46:784-7. Jarcho A. 1935. Die altersveranderungen de r rassenmerkmale bei den erwachsenen. Anthropologischer Anzeiger 12:173-179. Jilka R. 2003. Biology of the basic multicellula r unit and the pathophysiology of osteoporosis. Med Pediatr Oncol 41:182-185. Jorgensen R. 1994. Developmental significance of epigenetic impositions on the plant genome: a paragenetic function for chromo somes. Dev Genet 15:523-532. Judex S, Lei X, Han D, Rubin C. 2007. Low-magn itude mechanical signals that stimulate bone formation in the ovariectomized rat are depe ndent on the applied frequency but not on the strain magnitude. J Biomech 40: 1333-13339. Junqueira L, Carneiro JE. 2003. Basic Histology. New York: Lange Medical Books McGrawHill. Kanis J, Delmas P, Burckhardt P, Cooper C, Torgerson D. 1997. Guidelines for the diagnosis and management of osteoporosis. Osteo Int 7:390-406. Kendrick G, Risinger H. 1967. Changes in the ante roposterior dimensions of the human male skull during the third and fourth decades of life Anat Rec 159:77-81. Kirkwood, TBL. 1995. What can evol ution theory tell us about the mechanisms of aging? In: Esser K, Martin GM, editor s. Molecular Aspects of Ag ing. New York: John Wiley and Sons. p 5-14. Kowalski C, Nasjletti C. 1976. Upper face height-tot al face height ratio in adult American black males. J Dent Res 55: 913. Lasker G. 1953. The age factor in bodily measurements of adult male and female Mexicans. Hum Biol 25:50-63. 115

PAGE 116

Lestrel P, Kapur K, Chauncey H. 1980. A cepha lometric study of ma ndibular cortical bone thickness in dentulous persons and dentur e wearers. J Prosth Dent 43:89-94. Levine R, Garza J, Wang P, Hurst C, Dev V. 2003. Adult facial growth: app lications to aesthetic surgery. Aesthetic Plast Surg 27:265-8. Lewis A, Roche A. 1988. Late growth changes in the craniofacial skeleton. Angle Orthod 58:127-135. Lovejoy CO, Meindl RS, Mensforth RP, Bart on TJ. 1985. Multifactorial determination of skeletal age at death: a met hod and blind tests of its accuracy. Am J Phys Anthropol 68:114. Marks M. 1995. William M. Bass and the developm ent of forensic anthropology in Tennessee. J Foren Sci 40:741-750. Moehrle A, Paro R. 1994. Spreading the silence: epigenetic transcrip tional regulation during Drosophila development. Dev Genet 15:478-484. Moore S. 1955. Hyperostosis cr anii. Springfield: CC Thomas. Moore K, Agur A. 2002. Essential clinical anatomy: 2nd Edition. Philadelphia: Lippincott Williams and Wilkens. Moss M. 1962. The functional matrix. In: Kraus B, Reidel R, editors. Vistas in orthodontics. Philadelphia: Lea & Febiger: 85. Moss M. 1968. The primacy of functional matrices in orofacial growth. Dent Pract Dent Rec 19:65-73. Moss M. 1969. Functional cranial an alysis of the mandibular angular cartilage in the rat. Angle Orthod 39:209-14. Moss M. 1972. Twenty years of functional cranial analysis. Am J Orthod Dentofacial Orthop 61:479-85. Moss M. 1979. Functional cranial analysis and the functional matrix. Int J Orthod 17:21-31. Moss M. 1995. Growth and development of the cran iofacial complex: an epigenetic viewpoint. In: Goodrich J, Hall C, editors. Craniofacial anomalies: growth and development from a surgical perspective. New York: Thieme Medical Publishers, Inc. p 1-7. Moss M. 1997a. The functional matrix hypothesis re visited. 1. The role of mechanotransduction. Am J Orthod Dentofacial Orthop 112:8-11. 116

PAGE 117

Moss M. 1997b. The functional matrix hypothesis revi sited. 2. The role of an osseous connected cellular network.: Am J Ort hod Dentofacial Orthop 112: 221-226. Moss M. 1997c. The functional matrix hypothesis revisited. 4. The epigenetic antithesis and the resolving synthesis. Am J Orthod Dentofac Orthop 112:410-7. Moss M, Greenberg S. 1967. Functio nal cranial analysis of the human maxillary bone: I, basal bone. Angle Orthod 37:151-64. Moss M, Salentijn L. 1969. The primary role of f unctional matrices in facial growth. Am J Orthod 55:566-77. Moss M, Young R. 1960. A functional approach to craniology. Am J Phys Anthropol. 18: 28192. Nasjletti C, Kowalski C. 1975. Stability of uppe r face height total f ace height ratio with increasing age. J Dent Res 54:1241. Oldberg S. 1945. Uber die Bedeutung der Hyperostos is frontalis interna u nd einiger verwandter Skelettveranderungen unter besonderer Berucksichti gung der Verhaltnisse bei Altersdiabetes. Upsala Lakareforen, Forhandl 51:1-116. Parfitt A. 1982. The coupling of bone formation to bone resorption: a critical analysis of the concept and its relevance to the pathogenesis of osteoporosis. Meta Bone Dis Relat Res 4:1-6. Parfitt A. 1994. Osteonal and hemi-osteonal remodeling: the spatial and temporal framework for signal traffic in adult human bone. J Cell Bio 55:273-288. Perkins SW, Batniji RK. 2005. Rejuvenation of the lower eyelid complex. Facial Plast Surg 21: 279-85. Pessa J. 2000. An algorithm of facial aging: verification of Lambros's theory by threedimensional stereolithography, w ith reference to the pathogen esis of midfacial aging, scleral show, and the lateral suborbital trough deformity. Plast Reconstr Surg 106:479-88. Pessa J, Chen Y. 2002. Curve analysis of the aging orbital aperture. Plast Reconstr Surg 109:751-5. Pessa J, Desvigne L, Lambros V, Nimerick J, Zadoo V. 1999a. Changes in ocular globe-toorbital rim position with age: Implications for aesthetic blepharoplasty of the lower eyelids Aesthetic Plast Surg 23:337-342. Pessa J, Slice D, Hanz K, Broadbent TJ, Rohr ich R. 2008. Aging and the shape of the mandible. Plast Reconstr Surg 121:196-200. 117

PAGE 118

Pessa J, Zadoo V, Mutimer K, Haffner C, Chen Y, DeWitt A, Garza J. 1998. Relative maxillary retrusion as a natural consequence of aging: combining skeletal and soft-tissue changes into an integrated model of midfaci al aging. Plast Reconstr Surg 102:205-12. Pessa J, Zadoo V, Yuan C, Ayedelotte J, Cuellar F, Cochran C, Mutimer K, Garza J. 1999b. Concertina effect and facial aging: nonlinear aspects of youthfulness and skeletal remodeling, and why, perhaps, infants have jowls. Plast Reconstr Surg. 103:635-44. Peterson J, Dechow PC. 2003. Material properties of the human cranial vault and zygoma. Anat Rec A Discov Mol Cell Evol Biol. 274:785-97. Peyron M, Blanc O, Lund J, Woda A. 2004. In fluence of age on adap tability of human mastication. J Neurophysiol 92:773-779. Pfitzner W. 1899. Social-anthropolg ische studien, I. Der cinfluss des lebens alters auf die anthropologischen characktere. Zeitschr ift fr Morphologie und Anthropologie 1:325377. Porter M, Vandervoort A, Lexell J. 1995. Aging of human muscle: structure, function and adaptability. Scand J Med Sci Sports 5:129-142. Pretorius E, Steyn M, Scholtz Y. 2006. Inve stigation into the us ability of geometric morphometric analysis in assessment of sexual dimorphism. Am J Phys Anthropol 129:64-70. Ousley SD. 2002. 3Skull, Museum of Natural History, Smiths onian Institute, Washington D.C. Qiu S, Rao D, Palnitkar S, Parfitt A. 2002. Age and distance from the surface but not menopause reduce osteocyte density in human cancellous bone. Bone 31:313-318. Rafferty K, Herring S, Artese F. 2000 Three-dimensional loading and growth of the zygomatic arch. J Exp Biol 203:2093-2104. Randall F. 1949. Age changes in young a dult Army males. Hum Biol 21:187-198. Ranly D. 1988. A synopsis of craniofacial growth. Norwalk: Appleton & Lange. Ravosa M, Noble V, Hylander WL, Johnson K, Ko walski E. 2000. Masticatory stress, orbital orientation and the evolution of the prim ate postorbital bar. J Hum Evol 83: 667-693. Rawlinson S, Mosley J, Suswill R, Pitsillides A, Lanyon L. 1995. Calvarial and limb bone cells in organ and monolayer culture do not s how the same early responses to dynamic mechanical strain. J Bone Miner Res 10:1225-1232. 118

PAGE 119

Rhine S. 1990. Non-metric skull racing. In: Gill G, Rhine S, editors. Skeletal attribution of race. Maxwell Museum of Anthropology. Anth ropological Papers No. 4. Albuquerque: University of New Mexico. p 9-20. Riggs B, Melton L, Robb R, Camp J, Atkinson E, McDaniel L, Amin S, Rouleau P, Khosla S. 2008. A population-based assessment of rates of bone loss at multiple skeletal sites: Evidence for substantial trabecular bone loss in young adult women and men. J Bone Miner Res 23:205-214. Riggs B, Melton L, Robb R, Camp J, Atkins on E, Peterson J, Rouleau P, McCollough C, Bouxsein M, Khosla S. 2004. Population-based study of age and sex differences in bone volumetric density, size, geometr y, and structure at different sk eletal sites. J Bone Miner Res 19:1945-1954. Rohlf F. 1990. Fitting curves to outlines. In: Rohlf F, Bookstein F, editors. Proceedings of the Michigan morphometrics workshop. Ann Arbor, Proceedings of the Michigan Morphometrics Workshop. Special Publication 2. p 167-76. Rohlf F, Marcus L. 1993. A revolution in mo rphometrics. Trends Ecol Evol 8:129-132. Rohlf F, Slice D. 1990. Extensions of the Proc rustes method for the optimal superimposition of landmarks. Syst Zool 39:40-59. Rosas A, Bastir M. 2002. Thin-plate spline analys is of allometry and sexual dimorphism in the human craniofacial complex. Am J Phys Anthropol 117:236-245. Ross C, Hylander W. 1996. In vivo and in vitro bone strain in the ow l monkey circumorbital region and the function of the postorbital septum. Am J Phys Anthropol 101:183-215. Rubin CT, Lanyon LE. 1985. Regulation of bone mass by mechanical strain magnitude. Calcif Tissue Int 37: 411-417. Rubin C, Judex S, Qin YX. 2006. Low-level mech anical signals and their potential as a nonpharmacological intervention for os teoporosis. Age Ageing 35:ii32-ii36. Rubin CT, McLeod KJ, Bain SD. 1990. Functional strains and cortical bone adaptation: epigenetic assurance of skeletal integrity. J Biomech23 Suppl 1:43-54. Rubin C, Turner AS, Bain S, Mallinckrodt C, McLeod K. 2001. Anabolism: Low mechanical signals strengthen long b ones. Nature 412: 603-604. Ruff C, Hayes W. 1988. Sex differences in agerelated remodeling of the femur and tibia. J Orthop Res 6:886-896. Ruff C, Holt B, Trinkaus E. 2006. Who's afraid of the big bad Wolff?: "Wolff's law" and bone functional adaptation. Am J Phys Anthropol 129:484-498. 119

PAGE 120

Russo C, Lauretani F, Bandinelli S, Bartali B, Di Iorio A, Volpato S, Guralnik J, Harris T, Ferrucci L. 2003. Aging bone in men and wo men: beyond changes in bone mineral density. Osteo Int 14:531-538. Saller K. 1930. Die Fehmaraner. Deutsche Rassenkunde 4:1. Saller K. 1931. Sddithmarsische Geestb evlkerung. Deutsche Rassenkunde 7:1. Sampson PD, Bookstein FL, Sheehan FH, Bolson EL, 1996. Eigenshape analysis of left ventricular outlines from contrast ventriculograms. In: Marcus LF, Corti M, Loy A, Naylor GJP, Slice DE. editors. Advances in Morphometrics. New York: Plenum. p. 131152. Sarnas K. 1957. Growth changes in skulls of anci ent man in North America. Acta Odontol Scand 15:213-271. SAS Institute Inc. 2000-2004. SAS 9.1.3. Help and Do cumentation. Cary, NC: SAS Institute Inc. Sharpe W. 1979. Age changes in human bone: an overview. Bulletin of the New York Academy of Medicine 55:757-773. Shaw RJ, Kahn D. 2007. Aging of the midface b ony elements: a three-dimensional computed tomographic study. Plast Reconstr Surg 119:675-681. Sheets HD, Keonho K, Mitchell CE, 2004. A combined landmark and outline-based approach to ontogenetic shape change in the Ordovician Trilobite Triarthrus becki. In: Eelewa A. editor. Applications of Morphometrics in Paleontology and Biology. New York: Springer. p. 67-81. Slice D. 1998. Morpheus et al.: software for morphometric res earch. Revision 01-31-00. Department of Ecology and Evolution, Stat e University of New York, Stony Brook, New York. Slice D. 2005. Modern morphometr ics. In: Slice D, editor. Mode rn morphometrics in physical anthropology. New York: Kluwer Academic. p 1-40. Slice D, Bookstein FL, Marcus LF, Rohlf FJ. 1996. A glossary for morphometrics. In: Marcus LF, Corti M, Loy A, Naylor GJP, Slice DE editors. Advances in Morphometrics. New York: Plenum. p. 531-551. Steinbock R. 1976. Bone as a living tissue: Pale oopathological Diagnos is and Identification. Springfield, IL: Charles C. Thomas, pp. 3-16. Susanne C. 1977. Individual age changes of th e morphological characteristics. J Hum Evol 6:181-189. 120

PAGE 121

Susanne C. 1978. Aging, continuous changes of a dulthood. In: Johnston F, Roche F, Susanne C, editors. Human Physical Growth and Maturation. New York: Plenum Press. Takahashi H, Frost H. 1966. Age and sex related changes in the amount of cortex of normal human ribs. Acta Odontol Scand 37:122-130. Tallgren A. 1957. Changes in adult face height due to aging, wear and loss of teeth and prosthetic treatment. Ac ta Odontol Scand 15:24. Tallgren A. 1967. The effect of denture wearing on facial morphology: a 7-year longitudinal study. Acta Odontol Scand 25:563-92. Tallgren A. 1972. The continuing reduction of the re sidual alveolar ridges in complete denture wearers: a mixed-longitudinal study cove ring 25 years. J Prosthet Dent 27:120-132. Tallgren A. 1974. Neurocranial morphology and aging: a longitudinal roentgen cephalometric study of adult Finnish women. Am J Phys Anthropol 41: 285-293. Thompson J, Kendrick G. 1964. Changes in the vert ical dimensions of the human male skull during the third and fourth decade of life. Anat Rec 150:209-214. Todd T. 1924. Thickness of the male white cranium. Anat Rec 27. U.S. Census Bureau. 2005. American Community Survey. Valeri CJ, Cole TM 3rd, Lele S, Richtsmeier JT. 1998. Capturing data from three-dimensional surfaces using fuzzy landmarks. Am J Phys Anthropol. 107:113-124. Wasserman P. 1989. Neural computation: Theory and practice. New York: Nostrand Reinhold. Weber G, Schafer K, Prossinger H, Gunz P, Mitterocker P, Seidler H. 2001. Virtual anthropology: the digital evolution in anthr opological sciences. J P hysiol Anthropol Appl Human Sci 20:69-80. Weiss K. 1984. On the number of members of ge nus of Homo who have ever lived, and some evolutionary interpreta tions. Hum Biol 56:637-649. Weiss K. 1989. Are known chronic disease related to the human lifespan and its evolution? Am J Hum Biol 1:307-319. West K, McNamara JJ. 1999. Changes in the craniofacial complex from adolescence to midadulthood: a cephalometric study. Am J Orthod Dentofacial Orthop 115:521-532. White T. 2000. Human osteology. Sa n Diego: Academic Press. 121

PAGE 122

122 Witzel U, Preuschoft H. 2002. The functional shap e of the human skull, as documented by threedimensional FEM studies. Anat Anz 60:113-135. Wolff J. 1986. The law of bone remodeling (transla tion Maquet P, Furlong R.) Berlin: Springer. Wunsche H. 1953. Altersceranderungen netris cher Merkmale in der Kindheitund beim Erwachsenen, sowie ihre Beziehungen zum sozi alen Mileu und zum Ha bitus. Zeitschrift fr Morphologie und Anthropologie 45:368-415. Zadoo V, Pessa J. 2000. Biological arches and cha nges to the curvilinea r form of the aging maxilla. Plast Reconstr Surg 106:460-466. Zioupos P, Currey J. 1998. Changes in the stiffne ss, strength, and toughness of human cortical bone with age. Bone 22:57-66. Zuckerman S. 1955 Age changes in the basicranial axis of the human skull. Am J Phys Anthropol 13:521-539.

PAGE 123

BIOGRAPHICAL SKETCH In 2001, Shanna E. Williams received her Bach elor of Arts degrees in anthropology and biology from the University of Virginia, Charlo ttesville, Virginia. That same year, she was accepted into the anthropology graduate program at the University of Florida. In 2004, Shanna received her masters degree and continued on fo r her doctoral degree. Upon completion of her doctoral degree, she hopes to pursue a career encompassing both research and teaching. 123