Investigations in Sex Estimation: An Analysis of Methods Used for Assessment

By Donna L. Harrison

Investigations in Sex Estimation: A Comparison of Morphological and Metrical Methods is a collective comparison of the many morphological and metric methods currently used on adult and juvenile human remains. This single comprehensive resource for sex estimation in skeletal material includes discussion on the evolution of sexual dimorphism in modern humans, how sexual dimorphism manifests itself in those bones, the growth development in juveniles and how sexual dimorphism can be measured in their bones. Data is presented which contradicts previously held postulates and some different uses of sex estimation methods are suggested, such as using the grading system to assess evolutionary change in skeletons or grouping juveniles into smaller groups by age to better estimate sex. New insights are offered for future research from the presentation of case studies on gender and a comparison of the sex differences between two African-American collections which suggests a correlation between occupation and evolutionary change. To achieve the objective, data on 294 adults from four diverse collections of known sex are obtained using a total of 67 morphological and metric methods. An additional 23 morphological and metric methods used on juveniles, aged 0 to 18, from two of the collections. The methods are specific to either the cranium, mandible, pelvis, humerus or femur. The compilation of assessment of sex provides definitive answers on which type of method, morphological or metric, is more accurate, which morphological and metric methods are the most accurate, and which bones are more reliable to estimate sex especially in the absence of a pelvis. Investigations in Sex Estimation is intended to be used in the field and the laboratory for the identification of sex in human remains and aims to provide a wealth of data for future research in sexual dimorphic studies.

• Provides the anthropological community with the results of the first collective comparison of the many morphological and metric methods currently used on adults and juveniles
• Introduces methods, reviews sexual dimorphism, and shares new insights for future research
• Highly illustrated, using testing data from four very diverse skeletal collections

• Language: English
• Publisher: Academic Press
• Release date: Sep 29, 2018
• ISBN: 9780128162170

Donna L. Harrison

As a Bioarchaeologist, Dr. Harrison analyses skeletons, collecting age, sex and pathology data as well as investigating population specific traits. All of these components, and more, provide a wealth of data as a researcher. She particularly studies collections that include juveniles as they provide early data on trends that are found in adults. Her research is on the estimation of sex using morphological and metric methods and will be the basis for her proposed work. Dr. Harrison has participated in archaeological digs in the UK, Belgium and Spain and is a Registered Professional Archaeologist in the United States. She is a member of the Paleopathology Association, the American Association of Physical Anthropologists, the Society of American Archaeology and the American Institute of Archaeology.

Book preview

Female

Chapter 1

Introduction

Abstract

This chapter offers a brief review of the historical attitudes and issues surrounding the discipline of sex estimation. There are currently many methods used in the field when estimating sex. Morphological methods are methods of observation which are used on areas of the skeleton that are sexually dimorphic. Metric methods are used to measure the differences between males and females to perform statistical analysis. Both types are used in the research and are discussed. Concepts of gender, ethnicity, and culture are also explored as they can have an impact on any study, such as affecting the observer’s assessment of sex.

Keywords

Sex estimation; methods; gender; ethnicity; culture

The determination of sex is a matter of first importance in all anthropological work if reliable records are to be made or accurate deductions drawn from the material employed.

Douglas E. Derry, 1909 (p. 266)

1.1 Overview

The identification of sex in skeletal material is one of the fundamental components necessary for any archeological study. Without knowing the true sex—the ipso facto biological sex—of each skeleton, data that is collected and analyzed can be subject to bias and misinterpretation (Cox, 1995; Harrington & Blakely, 1995; Stewart, 1954; Walker, 2005; Walrath, Turner, & Bruzek, 2004; Weiss, 1972). While known sex collections often reduce these errors, many collections are not documented well, if at all, with regard to the sex of the skeletal material; this limitation does not make it any less important to establish sex in these collections.

Sex estimation, along with age estimation, is vital for accurate statistical analysis as well as an accurate demographic interpretation of population data. Compounding the problem is the challenge of estimating sex differences between human males and females which are not as pronounced as they are in other primates. For instance, bone size is highly sexually dimorphic in gorillas as is canine teeth size in chimpanzees (White & Folkens, 2005). By contrast, sex differences in bone size and canine teeth are relatively small in humans. One suggestion is that the only way to establish a high degree of accuracy when sexing human skeletons is to seriate the population; the females will be smaller and the males larger (Ferembach, Schwidetzky, & Stoukal, 1980). While this technique may have some merit, it is not always applicable and raises other questions: just how many are needed in a series in order to accurately determine sex, and how would serialization compensate for the variability inherent in both sexes?

There are other oversimplifications made with regard to sex estimation that do not consider the effect of external factors on skeletal material. For example, sex estimation is more accurate when maturity is reached (Krogman & Iscan, 1986). However, this postulate discounts factors such as diseases or nutritional deficiencies which can affect bone growth during early childhood. Another oversimplification is that the more robust, the more likely the skeleton is considered male; conversely, the more gracile, the more likely the skeleton is considered female (White & Folkens, 2005). This postulate discounts the variability that can exist within populations as well as the effects of age, disease, or nutritional factors on size and growth.

The prevalent view in the literature is that sexual traits do not appear in the skeleton until the age of puberty, usually around 15–18 years of age. Remains before 15 years of age have a 50/50 chance of being correctly sexed (Krogman & Iscan, 1986). This postulate provides an easy excuse to avoid the difficult, but not impossible, task of finding accurate methods for sexing juveniles. For example, Molleson, Cruse, and Mays (1998) introduced a method using the shape of the orbit and reported an accuracy of 89%. As early as 1980, Weaver introduced a method using the auricular surface elevation of the pelvis reporting 85.1% accuracy in males, but only 57.6% in females. The method was retested by Sutter (2003) who produced an overall 71.7%. While these figures may not be as accurate as some adult methods, the results are still better than the 50/50 chance method generally prescribed in the literature.

Estimating sex has inherent difficulties built into what should be a simple process. Foremost is the unintentional bias in favor of males when interpreting the degree of sexual expression. Weiss (1972) performed an extensive study looking at male bias in 43 populations whose sex was unknown; they consisted of Amerindians, Africans, Japanese, and Europeans. These populations covered various time ranges and cultures. He found that 33 of the populations studied contained 39%–80% more males than females; in a random sample, one would normally expect a 50/50 split. Testing these results further to determine if male bias was a factor, he then studied 31 preindustrial populations in which sex was known and verified by records. In this group, only 13 populations had more males. He concluded from the results that in any unknown sex population under study, classification of males will be on average 12% higher than females. His recommendation was to use population methods or methods that had been previously used on a related population. However, even this suggestion does not fully solve the complexities of sex estimation. Sexual dimorphism can vary between populations and be affected by growth changes as a result of environmental and nutritional factors (Buikstra & Ubelaker, 1994). Therefore, while two populations can have similar characteristics, the expression of these characteristics can be different.

Other difficulties in estimating sex arise from material that is fragmented and poorly preserved. For example, it is estimated that 30% of the pelves in any archeological sample are missing the pubic bone (Waldron, 1987). When this crucial component of the pelvis, the bone considered the most sexually dimorphic in the literature, is missing or incomplete sex estimation must be able to analyze other bones that are fully present. Additionally, poor preservation of skeletal material can result in a loss of certain age groups for study, such as juveniles and the elderly due to lower bone mass levels; in the former, the bone mass is building up but is still at low levels, and in the latter, bone mass is decreasing as a result of the ageing process. The end result is that often times only the middle-aged group, from 20 to 50, is present (Walker, 1995). Therefore, it is paramount to the researcher to have many tried and tested methods on a variety of bones to estimate sex which can be used on incomplete or poorly preserved skeletons.

1.2 Method

There are currently many methods in the field that are used to estimate sex ranging from simple observations to more complex statistical analysis; each method has its advocates and its critics. What they have in common is the objective of providing accurate data that can then be used to identify, e.g., population trends or ethnic variation. The more information that is available, the more comparative studies can be performed (Brothwell & Zakrzewski, 2004).

The simplest and quickest method to estimate sex is by observation; sexual characteristics in an area of bone are either present or absent, highly marked or minimally marked, or not clear enough to make a judgment. This method, known as the morphological method, is used today primarily on the pelvis, the cranium, the mandible, and some long bones. During the 19th century, as anatomists dissected and examined the bones of the human body, they noticed differences between males and females. For example, Derry (1909) an anatomist working at the Cairo Government School of Medicine in Egypt, published a paper on his research which examined four characteristics of the pelvis that seemed to exhibit distinctive sex differences. These were the iliac crest, greater sciatic notch, subpubic angle and preauricular groove. According to Derry, these features could be recognized at a glance with little effort.

Metric methods also have their roots in the 19th century. Some of the early studies attempted to record and understand how measured differences between males and females could be interpreted and used as a general rule of thumb applied to all human skeletons without regard to population affiliation. For example, Joseph Hyrtl (1871), an Austrian anatomist, studied the differences in sternum measurement between males and females. He concluded that there existed a proportional size difference in the sternum that clearly demonstrated sexual dimorphism. He stated that in females, the manubrium exceeded half the body length while in males, the body length was twice the length of the manubrium. This was later refuted by Dwight (1881) who repeated the study and found that while statistically the mean of his measurements did agree with Hyrtl’s law, individually the measurements did not. In fact, approximately one-half of the skeletons were exceptions to the law. Undeterred, metric measurements still continued to be recorded and analyzed by most anatomists well into the 20th century, most notably by the German anthropologist Rudolf Martin (1928). The second volume of his book was a compendium of measurement instructions for the body, many of which are still referred to and used today. Martin believed that measurements were critical to the study of variation within the human body, both individually and as part of the individual’s population affiliation.

The latter half of the 20th century has seen the introduction of statistical analysis incorporated into anthropological studies; metric data is not just compared, but subjected to correlation testing and fitted into data curves to demonstrate statistical acceptance. While the results may be more accurate when applied to sex estimation studies, accuracy is only as certain as the accuracy of the data being analyzed. If true sex is not known, then statistical analysis is subject to the same observational error as either morphological or metric methods. The use of statistical analysis has been questioned by several authors, including Howells (1966). Krogman and Iscan note that elaborate statistical analysis does not raise the averages appreciably (1986, p. 259).

In any study that uses a method or methods to estimate sex, the concepts of gender, culture, and ethnicity must also be considered. Specifically, how much of an impact do they have on any given research?

1.3 Gender

Biologically, we are born to be either male or female; we are taught that chromosomes determine biological sex and the resulting development of sexual characteristics. The typical male has a chromosome configuration of "XY that results in the creation of a scrotal sac containing the testes and a penis while the typical female has an XX chromosome configuration that results in the creation of a vagina and clitoris. The term gender used to express one’s own psychosexual concept is a construct of the 20th century; an individual perceives themselves as being either male or female regardless of their external sexual organs. Normally, there is agreement between the mind’s concept and the physical presentation; sometimes there is disagreement. When this happens, difficulties can arise for the individual, especially when they do not conform to cultural norms. Additionally, children can be born with a mixture of external sexual characteristics. For instance, the chromosome can be XX but a penis and scrotum are present; similarly, the chromosome can by XY but large breasts develop. Individuals who experienced this sexual ambiguity were historically labeled hermaphrodites. Today, they are labeled intersex" and their condition is known as DSD, or disorders of sexual development (Fausto-Sterling, 2012). These individuals are usually seen by a multidisciplinary team consisting of specialists trained in the latest surgical advances, as well as psychiatrists who provide information, counseling, and support. Often decisions concerning surgery are deferred until such time as the child is old enough to decide for themselves what their psychological gender is (Me, My Sex, and I, BBC, 2011).

This practice was not always the approach taken. In the 1960s and 1970s, the treatment for children who were diagnosed with any kind of sexual abnormality was often surgical alteration depending on how their sexual organs appeared and regardless of how their internal sexual structure was formed. Dr. Money, a renowned sexologist working out of Johns Hopkins University, dominated the field with his views that children were born gender neutral and remained so up to 2 years of age (Money & Ehrhardt, 1972). He was an advocate of nurture over nature; the primary factor in determining one’s own gender was not biology but the way in which a child was raised. He devoutly believed that through surgery and hormonal treatment an intersex child could be made into whichever sex seemed best and could be raised as such with no detriment to their mental wellbeing. His most famous case concerned a twin boy, David Reimer, who as an infant of 8 months underwent a circumcision that resulted in the total loss of his penis. When the boy’s parents sought Dr. Money’s counsel, he was given the opportunity to prove his radical new theory by changing the gender identity of a perfectly normal boy. He convinced the child’s parents that the baby boy would be best served by his treatment program, and for the next 12 years, the baby boy was raised as a girl while his twin was raised as a boy. It was the perfect controlled experiment. Dr. Money reported no problems, claiming his treatment a huge success, and published his results in 1972 in his book "Man & Woman Boy & Girl." He was proclaimed a pioneer in gender transformations by the medical community which resulted in his approach becoming standard practice when treating intersex children (Colapinto, 2000).

However, there were some persons in the medical field who disagreed with his positive prognosis. One of them was Dr. Keith Sigmundson, chief psychiatrist at the Child Guidance Clinic Winnipeg, Canada, currently in charge of the ‘now’ girl’s case. While attempting to continue with the treatment advocated by Dr. Money, he did not fail to notice how characteristically male the child acted and behaved (Colapinto, 2000). Another was Dr. Milton Diamond, Professor of anatomy and reproductive biology at the University of Hawaii at Manoa. His early experiments while at university involved manipulating the hormones in guinea pig fetuses. He injected testosterone into the mother; after birth, the females in the litter emulated masculine behavior (Diamond & Young, 1963). The experiment demonstrated that sex was determined before birth while still in the womb; it was nature, not nurture, that determined gender identity. The results prompted him to investigate cases involving gender surgery; he was particularly interested in Dr. Money’s test case. In a 1997 paper, "Sex Reassignment at Birth: A Long Term Review and Clinical Implications," Diamond and Sigmundson collaborated on a follow-up study on David Reimer who in the interim years had reverted back to being a boy, something he had always felt he was (Diamond & Sigmundson, 1997). Based on the interview with David, the history of the case and other follow-up interviews with children who had been surgically altered, Diamond stated that normal humans were not born psychosexually neutral at birth but with an inclination to interact with their environs as either male or female.

More recent research has postulated that the brain also plays a role in determining gender identity; there are significant differences in areas of the brain between males and females. For example, the hypothalamus secretes hormones in the pituitary gland; this area is nearly twice as large in males as it is in females (Gorski, 1998; Swaab, 2007). Gender, then, is more than the sum parts of the physical sexual expression, the chromosomes and now the brain. Blackless et al. (2000) stated:

If, however, one relinquishes an a priori belief in complete genital dimorphism, one can examine sexual development with an eye toward variability rather than bimodality. (p. 151)

If we accept that variability can exist in the physical expression of sexuality in the living human body, how do we interpret the variability of sexual expression seen in bones archeologically? Can we identify a skeleton’s gender or are we relegated to using the limited categories of male and female when estimating sex? Between 1964 and 1991, the number of cases of DSD recorded in the literature ranged between 1 and 2 for every 1000 live births (Blackless et al., 2000). Small enough as this number of intersex cases is, there is no known number historically. Additionally, there has been no study ever performed on the effects of hermaphroditism on bone which leaves the archeologist no choice but to assess skeletal material as either male or female.

1.4 Ethnicity and Culture

Ethnicity is based on a group’s shared biological ancestry; culture is an expression of the group’s socially shared beliefs and practices. The human body develops and adapts to its environment and varies based on the culture of the ethnic group. For instance, repetitive activity performed by an individual can increase muscle size which results in the creation of entheses, or muscle insertions sites. Entheses stimulate increased bone formation which in turn forms musculoskeletal stress markers, or MSM (Weiss, 2003). Studies of activity patterns that investigate MSM, such as Dutour (1986), have helped to identify sex-specific activity in various cultures. Environmental influences can also affect the expression of sexual characteristics and account for the variation present in bones (Hawkey & Merbs, 1995). For instance, males in one ethnic group can express well-developed muscle attachments on their bones as a result of diet and activity while males in another ethnic group can express no significant development in the same areas.

Why is it important to test methods against different ethnic and regional groups? When geographical barriers occur between populations, the resulting isolation leads to speciation creating distinct morphological variation in the population (Campbell, 2009). It is often the case that a new method proposed for estimation of sex is developed and tested on a specific ethnic group that produces remarkable results. However, when this method is repeated on a different ethnic group the results can be contrastingly poor. For example, the study by Loth and Henneberg (1996) which looked at the mandibular ramus flexure in a South African population reported 99.1% accuracy in estimating sex. Donnelly, Hens, Rogers, and Schneider (1998) repeated their test using two sample groups. The first subset consisted of 16 known sex specimens, 2 females and 14 males. The second subset consisted of 80 specimens whose sex was estimated by pelvic morphology. The results from both samples ranged between 63% and 69%. Is the feature being evaluated affected by some type of environmental or cultural influence which presents itself in one group but not another? This is why it is important to test methods on different ethnic groups, both geographically and culturally.

This leads to the basic question that was posited earlier and that every researcher faces: how much does gender, culture, or ethnicity impact the skeletal system? In archeological material, the answer is not usually clear even when there are associated gender-specific grave goods. For instance, an on-going excavation in Russia’s southern Ural steppes has recently uncovered a skeleton in a burial mound that appears to be morphologically male but is currently waiting for the DNA analysis results. He is surrounded by an abundance of artifacts and jewelry that are normally associated with a female grave. Other items contained within the burial are a small wicker chest thought to be a vanity case, a large silver mirror, garments with plaques portraying flowers and rosettes, and horse teeth filled with red pigment. There is even a pair of earrings that have been found, one on either side of the skull near the temporal bone. The grave is one of many burials currently under investigation by Professor Yablonsky (2013) of the Institute of Archeology at the Russian Academy of Sciences. This burial and those related to it are part of the Sarmatian culture that occupied the area during the Early Iron Age extending into the Bronze Age. The Sarmatians were militant horse warriors, both the males and females; horse harnesses and skeletons are frequently found in male graves while quivers of arrows are found in female graves (Yablonsky, 2010). However, none of these goods was found within this particular burial. Is this new burial an example of cultural gender or is it something else as yet unknown?

When estimating sex, it is not enough to decide only on one method to use. Other variables must be considered: which bones are the best preserved and demonstrate sexual dimorphism, should juveniles be included, should the concept of gender be considered, what will be the influences of culture and variation from ethnicity. These variables have an impact on the results of any study.

1.5 Research Parameters

The focus of this book is to assess which morphological and metric methods used to estimate sex are better when tested against the cranium, mandible, pelvis, humerus, and femur. These are the bones most often used in sex estimation studies because they are the most sexually dimorphic. From an archeological perspective, they are also the bones more likely to be preserved in a grave because of the high density of their cortical bone (Mays, 2005). Therefore, new and promising methods that are developed for forensic interest, such as using the hyoid bone (Kindschuh, Dupras, & Cowgill, 2010) or the bony labyrinth of the ear canal (Osipov et al., 2013) to determine sex, are not included.

Both morphological and metric types of methods will be applied to multiple ethnic adult and juvenile samples. All evaluations will be performed on stored collections, not skeletons in situ.

Biological sex, not gender, will be assessed on skeletons of known sex. Unfortunately, there is a paucity of gender studies in the archeological field which makes interpretation of ambiguous skeletal material difficult without a frame of reference. However, if a skeleton so completely digresses from its recorded sex, further discussion will occur.

Historically, the literature is full of the terms black and negro to differentiate material in a mixed sample from white. Therefore, when discussing these past studies, the term used by these authors will be used to reduce the amount of tautology associated with this ethnic group. However, since the new research data is specific to the Hamann–Todd and Terry Collections the term African–American will be used in discussion related to their samples.

The word estimation—and not determination—is used. Both words have been used in the litany of studies on the subject; estimation is a more recent usage. According to dictionaries, determination is defined as an exact judgment; estimation is an educated guess. I can make an exact judgment of sex by looking at death records. I can only guess the sex of an undocumented skeleton "in situ" by looking at the available features on sexually dimorphic bones. In this book, since my background is more field work than lab work, I learned to estimate sex and use the term prolifically and without apology.

Normally when analyzing a skeletal collection the usual practice is to discuss the results within the same chapter. However, since multiple skeletal collections are being analyzed, the collection results will be presented without in-depth discussion. Instead, a literature comparison will be provided which compares the results of the most accurate morphological and metric methods for each collection to the results reported from the original source. In the main study results chapters, the literature comparison will be followed by supplemental observations. The supplemental observations section contains any additional pertinent sex estimation methods, as well as any unusual features that may have impacted assessment.

Chapter 2 provides background information on sexual dimorphism in adults which includes a discussion on the evolutionary changes of locomotion and parturition and their effect on the human skeleton. Chapter 3 provides similar background information for juveniles but focuses on the interpretation of growth patterns and variations from various longitudinal studies which identified sex differences at early ages. Chapter 4 provides information on the collections used in this research, such as their historical context, sample size, and demographics. Chapter 5 details the adult and juvenile methods used in this research. Other sections discussed are excluded methods, inter and intraobserver error, and statistics. Chapters 6 through 9 present the results of sex estimation on the collections. Chapter 6 is specific to pilot study results. Each subsequent chapter is specific to the results of a single collection. Chapter 10 compares and discusses the results of all four collections and provides summarized charts—one for morphological methods and one for metric methods—on the reliability of each method across the four samples. Chapter 11 is a summation of the research, as well as its limitations, implications, and future research considerations. All appendices are available from the author upon written request to the publisher.

Chapter 2

Background in Adult Sexual Dimorphism

Abstract

Evolutionary influences from locomotion and parturition and the associated structural changes to the human skeleton are discussed. A review of the investigations into sexual dimorphism which has resulted in the establishment of diverse methods used today to estimate sex in the cranium, mandible, pelvis, humerus, and femur