The Bioarchaeology of Childhood

In biological anthropology, non-adult remains (i.e., developmentally immature skeletal remains)Footnote ¹ were described as having “little anthropological value” by Earnest Hooton in 1930 (Hooton Reference Hooton1930, 15). This sentiment was not unique at the time but reflected the state of biological anthropology in the early 20th century. At this point, research methods focused on metrics or measurements of the body to try to find differences between populations. Meanwhile, publications utilized descriptive writing, describing the disease or condition visible, rather than trying to understand where it came from or why some people were more affected than others, or what the presence of this disease in the past might mean about life in the past, more broadly. Hooton (Reference Hooton1930) continued that, in this world of metrics and descriptive writing, the physical remains of non-adults were fragile and often poorly preserved, so they did not offer the same academic value or meaning that a fully articulated and mature skeleton could.

When biological anthropologists did consider non-adult remains, it was often with the intent of establishing methods to estimate age and sex, rather than asking questions about childhood health or agency (Lewis Reference Lewis2007, 10). While these studies (e.g., Schour and Massler Reference Schour and Massler1941; Olivier and Pineau Reference Olivier and Pineau1960) now form foundational pieces for the study of childhood, they were ultimately limited in scope, listing and describing conditions rather than implementing theoretical frameworks to better gain a more nuanced view of life in the past.

The pattern of metrics and descriptive writing started to change some thirty years later, when researchers began to link non-adult growth patterns with poor health during life (e.g., Johnston Reference Johnston1962). Shortly afterwards, biological anthropologists began to examine measurements of long bones to understand growth and stunting in past populations. Once again, these studies were not particularly focused on learning more about childhood experiences, but rather used conditions of childhood as a proxy for the overall health of past populations (Lewis Reference Lewis2007; Saunders Reference Saunders, Katzenberg and Saunders2008). It is important to note, that many of these studies were conducted on Indigenous remains, often without the informed consent of descent groups; this is an approach that should not be replicated today.

Progress in the study of childhood, however, or at least the study of non-adult skeletal remains, came to a pause in the 1980s when Buikstra and Cook (Reference Buikstra and Cook.1980) stated that studies of children were hindered by poor preservation, lack of recovery, and small sample sizes. Subsequent papers focused on this lack of preservation of non-adults, in some ways, as an explanatory factor as to why they were not investigating non-adults more closely (e.g., Von Endt and Ortner Reference Von Endt and Ortner1984). Lewis (Reference Lewis2007, 12) states that this assumption continues today, limiting further investigations.

Then, in the same year that Grete Lillehammer wrote “A Child Is Born,” Goodman and Armelagos (Reference Goodman and Armelagos1989) suggested that children under the age of five are particularly sensitive to environmental and cultural insults. So, once again, researchers started looking at non-adult remains to learn about their growth and health, in order to learn more about the communities in which they lived. However, more than just using biological data from non-adult remains as a proxy for adult health and well-being, Goodman and Armelagos (Reference Goodman and Armelagos1989) established the biocultural approach (discussed later in the Element) and incorporated this theoretical framework into their studies, pushing beyond descriptive research into learning more about life in the past.

A rise in research incorporating non-adult skeletal remains prompted the publication of several foundational pieces of work, including Scheuer and Black’s (Reference Scheuer and Black2000) Developmental Juvenile Osteology and Lewis’ (Reference Lewis2007) The Bioarchaeology of Children. There are also specific associations (e.g., the Society for the Study of Childhood in the Past [SSCIP]) and academic journals (e.g., Childhood in the Past) that promote and publish newly developing research on the topic of infants and children in the past. Today, studies incorporating non-adult remains employ questions that are more socially driven, attempting to understand how children were viewed and socially treated in the past, how they interacted with (and related to) their own world, and how they contributed to the communities around them (Halcrow and Tayles Reference Halcrow and Tayles2008; Baker 2018 in Beauchesne and Agarwal Reference Beauchesne and Agarwal2018; Thompson et al. Reference Thompson, Alfonso-Durruty and Crandall2014). As a result, children are being included in a wider range of studies than ever before.

This Element incorporates some of the foundational methods and approaches used in the bioarchaeological study of infants and children, with associated recording sheets available at the end of the Element (fillable forms also are available for download online).Footnote ²

Theory in Childhood Bioarchaeology

Theory is an analytical tool used to help explain certain phenomena or processes. Within biological anthropology, the intentional use of theoretical frameworks can allow a researcher to move beyond descriptive reports to contextualized and informed conclusions. In many ways, theory is the lens through which you can view your data. Hypothetically, this means you can use the same dataset with different theoretical frameworks and, ultimately, get different perspectives or insights from the data.

Bioarchaeological research today relies on the incorporation of theoretical frameworks, like the biocultural approach. Although the application of theory is not always explicit in research, it is often present. However, I would argue that being explicit with which theoretical framework you are using enables others to understand where you are coming from, and where you are going with your research question and analysis of data. Identifying and describing which theoretical framework underpins your analyses can also help reveal potential biases in your research, which enables the reader to understand your perspective and why you interpreted or considered some items in particular ways. With this point in mind, key theoretical frameworks are summarized in what follows; but texts also are available that delve deeply into this topic. For further discussions regarding use of theory in bioarchaeology, see Cheverko and colleagues’ (Reference Cheverko, Prince-Buitenhuys and Hubbe.2021) Theoretical Approaches in Bioarchaeology or Geller’s (Reference Geller and Martin2021) Theorizing Bioarchaeology.

Cultural Burial Practices

The recovery of non-adult skeletal remains may be beyond the control of archaeologists. Rather, the burial of infants and children can be influenced by social determinants and culturally relevant burial practices, as many cultures treat human infant bodies in different ways than those of other community members (Lally Reference Lally2008, 29). For example, Maria Liston and Susan Rotroff (Reference Liston and Rotroff2013, 2) examined infant remains recovered from an abandoned well in the ancient Athenian Agora (2nd century BCE, Greece), suggesting that the assemblage represents a “unique window into the cultural practices associated with … [the] youngest, most vulnerable members [of their community].” In particular, they suggest that this burial environment was used by midwives to dispose of infants who died during childbirth, rather than being buried in a formal cemetery (Lison & Rotroff Reference Liston and Rotroff2013). A thousand years later, Christian burial grounds often required the deceased to have been baptized prior to death in order to be buried on consecrated land. Consequently, stillborn or young infants who had not been baptized may have been buried elsewhere, or provided a clandestine burial rather than a typical or normative burial (e.g., Cootes et al. Reference Cootes, Thomas, Jordan, Axworthy and Carlin2020; Gilchrist Reference Gilchrist2022). For example, cillíní were the designated resting places for unbaptized infants and children in Ireland (Murphy Reference Murphy2011). If buried elsewhere, the remains of infants and children may not be found during the excavation of the cemetery. If buried in a clandestine burial that is shallower than other graves, their burial contexts may be disturbed, limiting their archaeological presence (Lewis Reference Lewis2007).

In each of these contexts, cultural burial practices will limit their archaeological presence if the entire cemetery, or areas specific to infants and children, are not excavated. These alternate burial locations may not be selected for excavation, and may not be protected from future developments, meaning the skeletal remains are not recovered and can be more easily disturbed and damaged. While there may be little we can do to overcome this issue, being aware of specific cultural practices, or exploring other portions of a burial ground, may help bioarchaeologists identify and recover infant and child remains. For more examples regarding non-adult burial practices in various geographical and temporal contexts, see Children, Death, and Burial: Archaeological Discourses, edited by Eileen Murphy and Mélie Le Roy (Reference Murphy and Roy2017).

Excavation Techniques

The perceived scarcity of non-adult remains in cemetery samples may also be the result of excavation practices, which have been developed for the excavation of adult remains and may be inadequate for non-adult remains (Saunders Reference Saunders, Katzenberg and Saunders2000; Lewis Reference Lewis and Katzenberg2019). Specifically, small osteological elements may be overlooked, while unfused skeletal elements may not be identified as human remains, leading to the incomplete recovery of the individual (Baker et al. Reference Baker, Dupras and Tocheris2005; Scheuer and Black Reference Scheuer and Black2004; Saunders Reference Saunders, Katzenberg and Saunders2000; Sundick 1978; Lally Reference Lally2008). Bioturbation of soils may also obscure smaller grave cuts, leading to the misidentification of features during archaeological excavation (McFadden et al. Reference McFadden, Muir and Oxenham2022).

Techniques to help ensure the recovery of infant and perinate remains include using fine mesh for screening soils. For example, Pokines and De La Paz (2016) found that using a 6.4 mm mesh resulted in 36.8 per cent recovery, while a 1.0 mm mesh resulted in the recovery of 99.8 per cent of perinate remains. Additionally, wrapping jaws prior to transportation can help retain unmineralized dentition (used for age estimates; see Section 3) (Saunders Reference Saunders, Katzenberg and Saunders2008). Lewis (Reference Lewis and Katzenberg2019) also recommends retaining the excavated soil from around a perinate burial, to be further examined in a lab setting for better controlled conditions.

Misidentification

If the graves of infants and children are located and thoroughly excavated, there are still potential issues associated with recovery, as infant and child remains due to their immature nature are frequently confused with faunal remains by non-bioarchaeologists (Buckberry Reference Buckberry2005). For example, an amphora burial from Athens included the cremated remains (cremains) of a high-status female and additional bone that was initially identified as burned animal remains (Liston and Papadopoulos Reference Liston and Papadopoulos2004). A re-examination of the material by a trained osteologist with specialization in non-adult remains, however, demonstrated that the additional bone was not faunal, but the remains of a human fetus which were likely misinterpreted due to their highly fragmentary nature, as well as the fact that the bones had been altered by the cremation process (Liston and Papadopoulos Reference Liston and Papadopoulos2004, 18–19).

Reference texts dedicated to non-adult osteology help to limit these challenges and should be on hand when analyzing skeletal material. The foundational text for non-adult osteology was first published in 2000 as Developmental Juvenile Osteology (Scheuer and Black Reference Scheuer and Black2000). For the first time, this comprehensive text summarized almost 100 years of research into the non-adult skeleton, outlining what was known, and where the gaps remained. The second edition (Cunningham et al. Reference Cunningham, Scheuer, Black, Liversidge and Christie2016) includes updated and augmented illustrations and bibliography, incorporating an additional 15 years of research into non-adult remains, and to this day is the most comprehensive text on the topic of non-adult skeletal remains.

Following this text, Scheuer and Black (Reference Scheuer and Black2004) published The Juvenile Skeleton, which provided a condensed version, targeting students of osteology. This publication outlines significant milestone events in the maturation of the human skeleton for a less specialized audience. Shortly afterwards, Baker et al. (Reference Baker, Dupras and Tocheris2005) also produced a field manual and textbook for osteology courses entitled The Osteology of Infants and Children.

Perhaps most appropriate for in-field researchers, Schaefer and colleagues (Reference Schaefer, Black and Scheuer2009) published Juvenile Osteology: A Laboratory and Field Manual. This brief text is designed to assist researchers by providing the methodological and mathematical tools necessary to complete their own studies rather than summarizing previous works.

While numerous sources are now available for the study of infant and child remains, the choice of text will depend on the target audience, including experience and level of detail required. If possible, the use of specialized osteoarchaeologists familiar with infant and child osteology will also be beneficial in complex cases.

Broadly speaking, human and non-human skeletal material can be differentiated based on differences in architecture and maturity. Architecturally, non-human remains may have features not present in humans, such as tails, claws, or baculum. Additionally, landmarks and articular surfaces tend to be more robust in non-human remains. Architectural differences may also be the result of different locomotion patterns or diets and are important to understand. For example, as humans are bipedal, the foramen magnum in the occipital bone is positioned inferiorly, while in quadrupeds, the foramen magnum is positioned posteriorly. Differences in the pelvis, hands, and feet may also provide indications regarding locomotion practices, and can help differentiate between human and non-human remains. With regard to teeth, human dentition, with its sharp anterior teeth and blunted posterior teeth, is suited to an omnivorous diet. Comparing these teeth with those of other animals, such as rodents or ungulates, may help differentiate between human and non-human remains.

When considering the maturity of skeletal remains, infant human remains are most likely to be misidentified as “non-human” due to their small and morphologically non-descript nature, as they are missing key features and landmarks that might help differentiate human and non-human species. The presence of growth plates and unfused epiphyses should help further identify non-adult specimens, while a completely fused/developed bone of a small size should not be mistaken for non-adult remains (Figure 2).

Figure 2 Photos of (A) a non-adult human femur and (B) an adult muskrat femur, illustrating differences in architecture and maturity, despite their similar size.

Estimating Age at Death

Generally speaking, the physical human body develops from conception until shortly after puberty, at which point the physical body maintains or degrades, as an individual continues to age chronologically (Halcrow and Tayles Reference Halcrow and Tayles2008). The development of the body follows a relatively predictable pattern that has been studied using osteological collections of individuals with known ages (Perry Reference Perry2005; Lewis Reference Lewis and Katzenberg2019). To estimate age at death, we examine patterns of biological age, that is, patterns of physical changes within the human body. With the application of established methods and standards, those physical manifestations of biological age are then translated to an estimated chronological age. However, as the development, and particularly the degradation, of the human body is still slightly variable, these age estimates are precisely that: estimates (Liversidge et al. Reference Liversidge, Buckberry and Marquez-Grant2015; O’Connell Reference O’Connell, Brickley and McKinley2004). Consequently, age ranges or confidence intervals should be incorporated, demonstrating the uncertainty between biological and chronological age estimates. However, it should be noted that biological or chronological age estimates still do not equate to social age.

For non-adult remains, there are three main approaches to estimating biological age: dental development and eruption, long bone length, and epiphyseal fusion. When selecting a method, it is important to understand the reference collection used to establish or develop the methodology to help ensure comparability between the reference collection used to establish the method, and the sample under consideration, as age and growth can be mediated by ancestry, socio-economic status, nutritional status, and more. Other methods also exist to estimate age at death in non-adult remains, including measurements of the occipital (e.g., basilar part) and temporal bones. These are summarized in Schaefer et al. (Reference Schaefer, Black and Scheuer2009).

Dental Development and Eruption

Perhaps the most reliable method to estimate age at death is dental development and eruption, as it is least likely to be affected by external factors (e.g., under nutrition) (White and Folkens Reference White, Black and Folkens2012, 364; Brickley 2004). During their lifetime, humans have two sets of teeth: deciduous dentition (or, your baby teeth) and your permanent dentition (adult teeth). These teeth develop in conical structures from the crown of the tooth to the tip of the root. Deciduous teeth also resorb, that is, they start to disappear, starting from the roots, until the tooth falls out, making room for the permanent tooth to come up in its spot or crypt (Hillson Reference Hillson1996). The bone around the teeth is the alveolar bone, and it is constantly remodeling and adapting to the different teeth as they develop and erupt in the mouth (Figure 4).

Figure 4 Non-adult mandible with deciduous teeth in occlusion, in eruption, and in alveolus (crypt).

Photo by the author.

When discussing dentition in humans, biological anthropologists divide the mouth into quadrants, and within each quadrant, we have a set number of incisors, canines, premolars, and molars. In our deciduous dentition, each quadrant has two incisors, one canine, and two molars, for a dental formula of 2.1.0.2. In our adult dentition, each quadrant has two incisors, one canine, two premolars, and three molars, for a dental formula of 2.1.2.3. Each type of tooth has a specific use, from tearing and ripping into tough foods (like meat) to grinding and chewing fibrous foods (like plant materials).

To estimate an individuals age at death, we examine the development and eruption of each tooth or the entire dentition, observing how much of the tooth is present (e.g., just the crown, roots partly complete, roots are fully developed), and whether that tooth has erupted out of the bone (into occlusion) or it is still sitting in the alveolar bone. There are generally two approaches to consider dental development and eruption. The first is looking at a developmental chart that captures “ideal” development at particular ages (e.g., Ubelaker Reference Ubelaker1989). Then, you simply compare the dentition you are looking at to the images and find the one that most closely corresponds to it, or what two images the individual might be between, in order to get an estimated age range. The second is a little more complex. Instead of looking at the entire dentition as one item, you examine each tooth individually and come up with a very specific age range for the individual you are looking at (e.g., Gustafson and Koch Reference Gustafson and Koch1974; Moorrees et al. Reference Moorrees, Fanning and Hunt1963). It is a much more complex and time-consuming process, but the results are more specific to each individual rather than just a broad comparison. Alternatively, AlQahanti and colleagues (2010) have also developed an interactive software based on their London Tooth Atlas, whereby individuals examine dental development, compare dentition, and input dental development data and are provided an age estimate (Audsley & Khan Reference Audsley and Khan2024).

Dental development is considered most appropriate for individuals less than 16 years of age. For individuals older than this, age estimation relies heavily on the mineralization and eruption of the third molars (colloquially known as the wisdom teeth), which is subject to a great deal of variation, and consequently, may not provide a confident age estimation (White and Folkens, 2005).

Epiphyseal Fusion

Researchers can also explore patterns of epiphyseal fusion, or the fusion of the growth plates at the ends of bones, to estimate age at death. Consider the humerus (Figure 5). In an adult, it is one solid bone; however, for non-adult remains, the bone comprises multiple pieces, including the diaphysis or shaft of the bone, and epiphyses or caps. As the individual ages, the bones continue to grow along the growth plate at the end of the shaft. Ossification and bone growth stops when the cells at the growth plate stop dividing and the epiphysis fuses to the end of the long bone.

Figure 5 Right proximal humerus with the proximal epiphyseal fusion line still visible (indicated by arrow).

Photo by the author.

Fusion of epiphyses occurs in known and predictable patterns, with slight variations depending on the individual, their biological sex, and the population (White and Folkens, 2005). Most epiphyseal activity occurs between the ages of 11 and 23, making this an ideal method for conducting age estimation on adolescent remains (White and Folkens, 2012). Additionally, White and Folkens (2012, 373) note that epiphyseal fusion and dental development are complementary aging techniques, as the methods slightly overlap and are often congruent with one another.

Numerous age estimation methods using epiphyseal fusion have been established and used within bioarchaeology. Perhaps the most complete and comprehensive review of epiphyseal fusion and the appearance of secondary ossification centres is summarized by Scheuer and Black (Reference Scheuer and Black2000) (updated by Cunningham et al., Reference Cunningham, Scheuer, Black, Liversidge and Christie2016), but numerous methods have been applied based on epiphyseal fusion, consulting various reference collections (e.g., Hodges Reference Hodges1933; Greulich and Pyle Reference Greulich and Pyle.1950; Tanner et al., Reference Tanner, Whitehouse, Cameron, Marshall, Healy and Goldstien1983; Cardoso Reference Cardoso2008a, Reference Cardoso2008b). Many of these methods were developed using radiographic images and may not be entirely appropriate for macroscopic examination. Alternatively, the methods developed by Cardoso (Reference Cardoso2008a, Reference Cardoso2008b) were developed using dry-bone observations, and are perhaps more appropriate for analysis of skeletal remains. In almost all approaches, however, bones are scored as unfused, fusing, or fused and compared to charts based on the bone and biological sex (if known/estimated), to estimate age at death. However, as indicated earlier, understanding the reference collection is necessary, as epiphyseal fusion timing can be influenced by a variety of factors, most significantly, socioeconomic status of a given population (Schmeling et al., Reference Schmeling, Reisinger, Loreck, Vendura, Markus and Geserick2000, Reference Schmeling, Schulz, Danner and Rosing2006).

Social Age

As previously discussed, social age corresponds to the culturally constructed perspectives of appropriate behaviours and responsibilities within a community. As social age is not always directly related to the biology of our bodies, it can be difficult to ascertain bioarchaeologically. Why then, do we even need to consider social age? Consider a 12-year-old boy (Figure 6).

Figure 6 Approximation of a 12-year-old boy in the Anglo-Saxon, early medieval, and modern periods, as interpreted by AI.

Images generated by the author.

In the Anglo-Saxon period (England, 450–1066 CE), this individual might be married or serving in the military. In the medieval period (Europe, 1200–1700 CE) a 12-year-old boy might be an apprentice, learning a new skill and taking on new roles and responsibilities in their community. Today, however, a 12-year-old would likely be in school and starting a first job like a paper route or babysitting. Clearly, these individuals would have had different lived experiences, different expectations, and different risks. Treating these 12-year-olds as the same would be a gross oversimplification, and possibly contribute to inaccurate and misguided interpretations about their bodies and lives.

To understand social ages as they were practiced in the past, we might be able to look to historical or ancient literary sources that provide a guide to the life course models in the past. In other situations, however, we have to look at the bioarchaeological data to help us define the social age categories. Researchers are continuing to develop ways to assess social age in past populations, and it is important to note that the methods employed need to be appropriate to that community, and how social age may have been experienced or manifested at the archaeological and bioarchaeological levels.

In studying social age changes in the Romano-British period (1st–5th centuries CE), Rebecca Gowland (Reference Gowland, Gowland and Knusel2006) examines mortuary profiles by looking at grave good distributions, working to understand changes in grave goods in relation to changes in social ages. Changes in grave good assemblages for females between the ages of 18 and 24 years may suggest the period in which young girls were married and became wives, mothers, or individuals running a household (Gowland Reference Gowland, Gowland and Knusel2006).

In the Roman period Gaul (4th–6th centuries CE), Avery and colleagues (Reference Avery, Brickley, Findlay, Chapelain de Seréville-Niel and Prowse2021) explore changes in diet, using dietary stable isotopes, finding marked changes for males around 16.5 years of age. Incorporating archaeological and ancient literary evidence, they suggest that this change is related to the social age transition to adolescence for young men as they began apprenticeships or began military service (Avery et al., Reference Avery, Brickley, Findlay, Chapelain de Seréville-Niel and Prowse2021).

Identifying ways in which social age may have been expressed and finding ways to explore these changes need to be considered on a site-by-site basis, incorporating other cross-cutting variables of identity including sex, gender, socio-economic status, immigration status, and more.

The Process of Aging

While age is a fundamental aspect of bioarchaeological studies, biological anthropologists’ engagement with age often takes a linear approach. First, researchers transform biological age (e.g., long bone length in non-adults) to chronological age estimates, using established osteological age estimation. However, we know that these methods are influenced by inter- and intra-population variation (Lewis Reference Lewis and Katzenberg2019). To account for this variation, researchers then employ age categories (e.g., 0–5 years). In turn, these age categories are likened to a social age category (e.g., infancy), without recognizing the social constructionism of these age categories and how they may vary depending on the sex, status, and community of the individual (Inglis & Halcrow Reference Inglis, Halcrow, Beauchesne and Agarwal2018). As Soafer (Reference Sofaer, Agarwal and Glencross2011) highlights, this linear approach to the tripartite model not only inappropriately equates the three measures of age but turns the process of aging into a series of distinct categories, removing the fluidity and variability of aging within a community.

Due to the precise nature in which biological anthropologists can estimate age at death in infants and children, we have the unique opportunity to assess the process of aging, viewing age as a continuous, rather than categorical, variable. Researchers should be critical within their studies and determine if categorical variables are necessary to their analysis, rather than replicating standard approaches commonly used when analysing adult remains. While there may be limitations in the bioarchaeological study of infants and children, there are certainly exciting prospects and opportunities that need to be seized and expanded moving forward.

Estimating Biological Sex

Although almost all parts of the human skeleton exhibit some degree of variation between biological sexes, sex estimation of adult skeletal remains traditionally focuses on features of the pelvis and cranium, where dimorphic features are most pronounced (White and Folkens 2005). Generally, these differences are described as sexual dimorphism, that is, the differences in physical characteristics or physiology between two dominantly defined sex-based groups (typically, male versus female). The challenge with this language is that it presents the expression of biological sex in a dichotomous pair rather than as within a spectrum, misrepresenting the variety of expressions within the human (and non-human) species (Wesp Reference Wesp, Agarwal and Wesp2017). Despite bioarchaeologists arguing since the 1990s that biological sex does not operate within a binary, the language of sexual dimorphism and use of male/female dichotomies persists (Hollimon Reference Hollimon, Agarwal and Wesp2017, 51).

When estimating morphological biological sex from skeletal remains, bioarchaeologists prefer to use features of the pelvis and cranium. That is, looking at the shape and form of the pelvis and cranium for features that exhibit a range of sex-linked manifestations. When pelvic and cranial methods are used in tandem, these approaches can produce the most reliable biological sex estimations from morphological assessments. However, the features examined in these methods develop during the later stage of puberty, meaning they are appropriate for post-pubertal skeletal remains, as well as peri-pubertal remains, with lower degrees of accuracy (Sanchez and Hoppa 2022). They are not, however, reliable for pre-pubertal skeletal remains.

Non-adult sex estimation methods also exist but are generally less reliable than adult specific methods. In fact, Mary Lewis (Reference Lewis2007) calls a reliable non-adult sex estimation method the “holy grail” within biological anthropology: something that has not yet been found but is highly coveted. The intention behind establishing a reliable method to estimate biological sex is not simply to further categorize humans, but rather, that biological sex may offer an important lens to understanding lived experiences in the past. For example, if boys and girls were treated very differently, incorporating biological sex estimations to help us understand the consequences of these different patterns of treatment can provide insights into long-term consequences, or health status within past populations. For example, in Colombia (1000–1400 CE), Miller and colleagues (Reference Miller, Agarwal, Langebaek, Beauchesne and Agarwal2018) found that sex-specific patterns of food consumption did not emerge later in life, but were present during childhood as well.

The common adult and non-adult methods are outlined in what follows, as well as biochemical and metric approaches that can be applied to all ages. While these methods exist, it is important to note that no morphological method for assessing biological sex in non-adult skeletal remains is considered “acceptable” by most osteologists (Buikstra and Ubelaker Reference Buikstra and Ubelaker1994, 16).

Dental Metrics

One problem with established methods is that they tend to be sample specific. However, when the method is applied to a different sample, the accuracy rate tends to be lower than in the original study (Buikstra and Ubelaker Reference Buikstra and Ubelaker1994; Cardoso Reference Cardoso2008c; Vlak et al. Reference Vlak, Roksandic and Schallaci2008). To overcome this limitation, researchers advocate for site-specific methods, where factors such as nutrition, living conditions, and even genetic variation may be better mediated (Cardoso Reference Cardoso2008c).

One approach that overcomes this limitation is site-specific dental metrics (Cardoso Reference Cardoso2008c). The use of dental metrics has been acknowledged as a viable method of sex estimation, particularly in the mandibular canines, which show the greatest degree of sexual dimorphism (Hillson Reference Hillson1996). In the approach outlined by Cardoso (Reference Cardoso2008c), permanent teeth are measured in the mesodistal (MD) and faciolingual (FL) or buccal-lingual (BL) planes, to the nearest tenth of a millimeter (Figure 7). Any tooth selected should be unobscured by dental wear or carious lesions, to ensure a full and complete measurement is taken.

Figure 7 Directions of tooth measurements indicated, including buccal-lingual (or labial lingual) (solid arrow) and mesio-distal (dotted arrow).

Photo by the author.

For each measurement (i.e., FL, MD) of each tooth, the sample mean is calculated, which serves as the sectioning point for the entire unidentified sample. Any individual with a measurement greater than the sectioning point is estimated male, while any individual with a measurement less than the sectioning point is estimated female. Where the sex-ratio is known to be imbalanced (e.g., cemetery associated with a military fort occupied predominantly by men), this approach will not be appropriate.

Beyond its site specificity, the main strength of this approach is that it does not require a comparative sample of adult remains nor a comparative subsample of known-sex individuals. Furthermore, the sectioning point procedure can be applied to poorly preserved samples and non-adult remains, so long as permanent dentition is present (Cardoso Reference Cardoso2008c). The approach is not applicable to deciduous dentition.

In using this approach, Cardoso (Reference Cardoso2008c) found that the canine was consistently the most sexually dimorphic, with an accuracy of more than 80 per cent, and was less affected by small sample sizes than other teeth in the dentition. Premolars also had high accuracy compared to other teeth (Cardoso Reference Cardoso2008c). In instances where the canine or premolar measurements disagree with one another, greater emphasis should be given to the canine, as this is the more dimorphic element of the dentition (Hillson Reference Hillson1996).

To help assess if dental metric measurements are performed reliably, intraobserver error measurements should also be taken (e.g., intraclass correlation coefficients). Measurements with a strong correlation between the first and second measurements can be considered reliable and used to estimate sex for individuals with permanent dentition present.

Exploring Gender

As discussed at the start of this section, sex and gender refer to two different constructs: the biological, and the social. While bioarchaeological methods have been developed to assess biological sex, assessment of gender is much more challenging. Some researchers work to assess gender independently of biological sex and incorporate other methods with varying degrees of success. Much like investigations into social age, the methods and approaches used to assess gender need to be site specific and relevant to the community or sample under study. For example, while grave goods may be used in some contexts to suggest gendered identities, implicating weaponry as a “masculine” item may be highly inappropriate in contexts where more than just men contributed to fighting. To date, very little research has considered gender in children, which offers an exciting area of development for future researchers: to consider how we might investigate gendered experiences in the past, and how gender may have shaped the lives of children and their communities by extension.

One avenue that offers promising results to explore gendered experiences of childhood is the sex-specific analysis of diet. By studying teeth that developed during childhood, but were archaeologically recovered from adult remains, researchers can employ more reliable sex estimation methods (e.g., pelvic and cranial dimorphism), and investigate experiences of childhood trapped in adult remains. Through the contextual analysis of the results, we can begin to understand how males and females were treated or behaved within their communities, thus offering insights into gendered patterns of behaviour.

For example, Miller and colleagues (Reference Miller, Dong, Pechenkina, Fan and Halcrow2020) use stable isotope analysis of incremental dentine segments to explore sex-specific patterns of breastfeeding, weaning, and childhood diets, in Eastern Zhou period (771–221 BCE) China. Through this gendered exploration of diet, they find that dietary differences between sex-specific groups emerged during childhood, suggesting a cultural gendering of individuals in ancient China (Miller et al., Reference Miller, Dong, Pechenkina, Fan and Halcrow2020). Using a similar methodology, Avery and colleagues (Reference Avery, Brickley, Findlay, Chapelain de Seréville-Niel and Prowse2021) explore sex-based patterns of dietary change in Roman Imperial adolescence. By exploring changes in diet along with archaeological and literary evidence, Avery et al. (Reference Avery, Brickley, Findlay, Chapelain de Seréville-Niel and Prowse2021) propose that the changing patterns correspond to changing social age roles that were different for males and females, suggesting that gender-based patterns of social age change as individuals reach the period of adolescence. In using this approach, we are once again reminded of how aspects of our identities (age, sex/gender, ethnicity, etc.) overlap and intersect with one another to produce our own individual lived experiences.

Methodological Approaches

To investigate patterns of growth and development, there are five key methodological approaches: long bone lengths, cortical thickness, body mass, enamel hypoplastic defects, and vertebral neural canal shape and size.

Cortical Thickness (Appositional Growth)

Just like bones can grow in length, bones also grow in thickness, what we call appositional growth, with the shaft of the bone increasing in width. Specifically, bone is deposited beneath the periosteum on the outer surface of the diaphysis (Mays Reference Mays, Crawford, Haldey and Shepherd2018). As new bone is forming on the outer surface, bone on the inner surface or endosteal surface is resorbed, altering the size of the internal medullary cavity. Generally, the rate of subperiosteal apposition outpaces endosteal resorption, increasing the bone width and the cortical thickness (i.e., thickness of the walls surrounding the marrow cavity) (Mays Reference Mays, Crawford, Haldey and Shepherd2018). While long bone length tracks growth relative to stature, cortical thickness is more closely related to growth in body mass (Ruff et al. Reference Ruff, Garofalo and Holmes2013). As such, the two processes (endochondral growth and appositional growth) are often investigated in tandem, to provide a more thorough understanding of growth and development in past populations (e.g., Dhavale et al. Reference Dhavale, Halcrow, Buckley, Tayles, Domett and Gray2017).

By measuring the thickness of the bone from the medullary cavity to the outside surface, we can measure the cortical thickness of the bone. In a normal, healthy individual, appositional growth occurs at a faster pace than resorption during subadult growth. This means that, in normal growth, cortical thickness should increase during non-adult development. However, when malnourished, clinical studies have demonstrated that rates of endosteal resorption increase, resulting in reduced cortical thickness for age, or a thinner bone (Mays Reference Mays, Crawford, Haldey and Shepherd2018). In past populations, cortical thickness-for-age is generally less comparable to modern populations, indicating that past populations experienced more nutritional stress than we do today. Furthermore, increased activity levels in pre-pubertal children may result in increased deposition along the subperiosteal surface; as a result, researchers may be able to differentiate between nutritional and biomechanical factors that contribute to altered cortical thickness (Mays Reference Mays, Crawford, Haldey and Shepherd2018).

To measure cortical thickness, we use radiographs or x-rays, and measure at the mid-point of the shaft of particular bones. Specifically, you measure the total subperiosteal width (T) and the medullary width (M). By subtracting the medullary width from the total subperiosteal width, you get the cortical thickness (Equation 1 and Figure 8). Cortical indices (CI) can then be calculated to help standardize the cortical volume for bone size, allowing researchers to evaluate changes in apposition growth in relation to age.

Figure 8 Illustration demonstrating difference between medullary cavity (M) and total subperiosteal thickness (T), used to calculate cortical thickness.

Illustration by the author.

$\text{C}\text{I}=\left(\text{T}–\text{M}\right)/{\text{T}}^{*}100$(1)

Equation 1 Cortical index calculation

There has been a recent push to use CT scans instead of X-rays, which are higher resolution, and allow for a more subtle measurement of differences, but getting access to this type of equipment might be hard to do in the field, so it may not be appropriate in all cases.

Generally, the acquisition of cortical bone is a more sensitive index of environmental stress than long bone length (Mays Reference Mays, Crawford, Haldey and Shepherd2018). Thus, regardless of using x-rays or CT scans, investigating cortical thicknesses is considered a more sensitive indicator of poor conditions during the growth period. The only trade-off is that you need more specialized equipment to take the measurements than measuring linear growth.

Linear Enamel Hypoplasia

Hypoplastic defects are the result of episodic growth disruptions captured in dental enamel during permanent and, to a lesser degree, deciduous tooth crown formation (Hillson Reference Hillson2014; Lewis Reference Lewis2007). Defects can manifest as furrow-type defects, pit-type defects, plane-type defects (large areas of brown stria planes), or localized hypoplasia of primary canines (LHPC) (Hillson Reference Hillson, Irish and Nelson2008, 166–167; Lukacs Reference Lukacs2009).

As your teeth are forming, the crown develops from the crown to the root. During this process, ameloblasts, the cells responsible for developing the enamel prior to mineralization are very sensitive to disruptions and do not work properly when they are under stress (Temple Reference Temple, Gowland and Halcrow2020; Lewis Reference Lewis2007). This can be nutritional stress or pathological stress, but it can even be impacted by psychosocial stress (see Kinaston et al., 2019; Marini & Flensborg Reference Marini and Flensborg2023). In these instances of prolonged increased stress, the enamel is affected. These defects can show up in the form of bands across the tooth that almost look like someone has put a really tight elastic band around it or pitting in an area where the enamel is just missing (or, very thin) (Figure 9). These marks are called hypoplastic defects, while the overall condition is called enamel hypoplasia.

Figure 9A

Figure 9B

Photos by the author.

Figure 9 Enamel hypoplastic defects including (A) linear defects and (B) linear defects and pitting. Dental samples from Apollonia, Greece.

Hypoplastic defects are generally non-specific indicators of stress during the growth period, meaning that the specific cause of growth disruption cannot be identified (Mays Reference Mays, Crawford, Haldey and Shepherd2018, 80). However, compared to linear or appositional growth, these dental defects represent periods of stress that were survived by the individual, meaning we are not entirely reliant on individuals who died during stressful periods to observe these defects.

A key benefit of exploring hypoplastic defects is that they are often visible to the naked eye or accessible by imaging techniques, meaning that x-rays, CT scans, or other potentially expensive equipment are not necessary to evaluate the condition. Due to the known development patterns of teeth, it is also possible to track where the hypoplastic defect is on the tooth and trace it back to the age at which the individual experienced developmental stress (Reid & Dean Reference Reid and Dean2006). For example, studying 18th- and 19th-century populations in Japan, Nakayama (2016) concluded that the frequency and timing of hypoplastic defects occurred between two and four years of age, which they suggest may be related to weaning stress (i.e., following the introduction of complementary foods or after the cessation of breastfeeding). Similarly, in a study of linear enamel hypoplastic defects (LEHs) in Late/Final Jamon period hunter-gatherers (Japan), Temple (Reference Temple, Gowland and Halcrow2020) found that the majority of LEHs were identified between 2.0 and 4.0 years of age, corresponding to a period of social age transition, partly mediated by the cessation of breastfeeding. However, Temple (Reference Temple, Gowland and Halcrow2020, 74) also acknowledges that the organization of enamel in the occlusal region of the tooth versus that in the cervical region of the tooth are different, which may contribute to the more weakly defined LEHs in the occlusal region of the teeth. Thus, it is important to consider both internal and external interpretations of growth disruption in the human body.

Vertebral Neural Canals

The use of dimensions of non-adult vertebral neural canal (VNC) and vertebral bodies can also be used as an indicator of growth disruption (Newman & Gowland Reference Newman and Gowland2015; Amoroso & Garcia Reference Amoroso and Garcia2018). Research has demonstrated that reduced vertebral canal size has been linked with stunted growth development, and decreased health (Clark et al., Reference Clark, Hall, Amelagos, Panjabi and Wetzel1986; Watts Reference Watts2011). Particularly exciting with this approach is that, while growth disruptions captured in long bone lengths may become distorted by effects of catch-up growth, VNC captures early postnatal growth disruptions that become “locked in” and cannot be remodeled during later growth and development (Newman and Gowland Reference Newman and Gowland2015, 156; Lewis Reference Lewis and Katzenberg2019).

The vertebrae form in three ossification centres: the vertebral body, the left neural arch, and the right neural arch (Cunningham et al., Reference Cunningham, Scheuer, Black, Liversidge and Christie2016). The vertebral neural canal (VNC) then surrounds the spinal cord and is positioned within the vertebral foramen. To assess VNC measurements, researchers rely on multiple measurements. First is the vertebral body height, which increases rapidly during birth to five years of age and during the pubertal growth spurt, with a period of reduced growth rate between these periods (Newman and Gowland Reference Newman and Gowland2015). The majority of the neural arch growth, however, is completed relatively early within the postnatal growth period, with the left and right neural arch fuses along the spinous process around 1.5 years of age, and the neural arches fuses to the vertebral body by five years of age (Cunningham et al., 2017).

Measurements of vertebral body height are taken from the midline of the centrum. As the vertebrae develop and fuse at different age points, this approach should be applied to vertebral bodies C3 to L5, where possible. Measurements of the VNC assess the transverse (TR) diameter of the neural canal (i.e., the farthest distance between the medial surfaces of the pedicles) and anterior-posterior (AP) diameter (i.e., posterior surface of the vertebral body to the further opposite point of the neural canal) (Figure 10). The latter should only be applied when fusion of the neural arches and vertebral bodies has initiated. All measurements should be taken with sliding calipers to the nearest 0.01 mm.

Figure 10A

Figure 10B

Illustration by the author.

Figure 10 Vertebral neural canal measurements including (A) anterior-posterior (dotted line) and transverse (solid line), and (B) vertebral body height.

In 2023, Brzobohatá and colleagues (Reference Brzobohatá, Velimsky and Frolik2023) used VNC to assess stressful early-childhood experiences in a population associated with silver mining in medieval Czechia (13th–16th centuries CE). Incorporating TR, AP, and vertebral body height (BH) measurements, they found evidence of health stress events during early childhood, which contributed to early mortality in males. Females, however, appear to have been better buffered against these early-life stress events, raising questions about female life course experiences or biological differences in processing stress (Brzobohatá et al., Reference Brzobohatá, Velimsky and Frolik2023).

Assumptions and Methodological Considerations

As with most of bioarchaeology, there are some common caveats to keep in mind when assessing growth and development.

The first is the issue of mortality bias. When looking at skeletons, we must keep in mind that we are looking at deceased individuals, or the non-survivors of the community (Wood et al. Reference Wood, Milner, Harpending and Weiss1992). This is especially true of looking at non-adult remains. Said another way, when we examine non-adult remains and patterns of their growth and development, we are evaluating individuals that died as children. So, are their experiences really comparable or representative of non-adults that became adults? Can we say that the growth patterns of those who died are the same as the growth patterns of those who lived? Or did exposure to conditions that stunted their growth mirror conditions that contributed to their death? Would the survivors have experienced the same conditions that affected their growth and development? These questions may not have immediate answers, but can be explored in various ways, allowing us to investigate mortality bias (e.g., exploring enamel hypoplasia in non-adult remains and adult remains).

The second caveat is the issue or confounding factor of biological sex. Growth differs for males and females, especially around adolescence, due to hormones. But, as discussed in Section 4, biological sex estimation of non-adult remains is problematic, with no standard method that is generally accepted by osteologists. For many researchers, this means we must combine data for males and females into one sample, possibly blurring results and slight differences in various populations. With advancements in biomolecular approaches to biological sex estimation, there are possibilities of overcoming these issues, but not in a way that is currently accessible to all researchers and all studies.

Diet and Weaning

Weaning is the extended process in which infants transition from receiving all nutrients from breastmilk to obtaining nutrients from sources other than breastmilk (weaning cessation). During this process, infants are introduced to complementary foods; that is, foods other than breastmilk that are brought in and used in tandem with breastmilk. Typically, breastmilk has everything to support a child up until six months. After this point, the breastmilk must be supplemented with other foods to ensure children are receiving all necessary nutrition to support growth and development.

In the past, weaning was a very dangerous period of a child’s life. Weaning foods were often under-nutritious because they did not provide all the nutrients needed, or were even harmful, containing nutrients that the child is not yet able to metabolise. For example, while cow’s milk can be consumed by much of the population without any detrimental factors, the consumption of cow’s milk by infants can cause kidney failure and renal damage (Ziegler, Reference Ziegler, Agostoni and Brunser2007). As a result, consuming this type of food as a weaning food may contribute to malnutrition or even death in some instances. Conversely, human breastmilk contains a wide range of antibodies and probiotics, giving the infant immune support from the immune system of the individual providing the breastmilk. Once fully weaned, however, that infant must rely entirely on their own undeveloped immune system. As a result, a period of weaning – reducing intake of antibiotic-rich breast milk and increasing consumption of inadequate or dangerous foods – can be a dangerous, or even deadly, period in a child’s life.

To explore patterns of diet and feeding in past populations, we may look to broad indicators of diet, including dental health, or more specific indicators, like dietary stable isotopes.

Dental Health

An analysis of dental health often includes an analysis of multiple dental health conditions, including rates of carious lesions (cavities) and dental wear, which can tell us a lot about diet. Cavities can be influenced by several factors, although diet is arguably the most important, and as such, the rate of carious lesions can tell us how cariogenic food may have been, which can tell us about carbohydrate or sugar intake (Hillson Reference Hillson1996, Reference Hillson, Irish and Nelson2008). Meanwhile, dental wear can tell us how tough and abrasive food may have been. For non-adults more specifically, a wear pattern on the front baby teeth might be able to tell us if the infant was bottle fed (Scott & Halcrow Reference Scott and Halcrow2017). Or the presence and rate of cavities could tell us about the type of food used for weaning. How old an individual is when those cavities and dental wear patterns start appearing might even tell us about the timing of weaning in a population, or if there was a sudden change in child diets (e.g., Prowse et al. Reference Prowse, Saunders, Schwarcz, Garnsey, Macchiarelli and Bondioli2008).

Carious Lesions

Carious lesions should be recorded by tooth and location on the tooth, recording areas where clear demineralization had taken place, rather than areas of discolouration. Although this approach only captures later stages of cavities, and not ‘pre-cavitated’ lesions, it is the most common approach in bioarchaeology, allowing for comparisons to other published studies (Hillson Reference Hillson2001; Lukacs Reference Lukacs2008).

Following Moore and Corbett (Reference Moore and Corbett1971, Reference Moore and Corbett1973), carious lesions should be identified based on the tooth and tooth surface, including the occlusal surface, interproximal surface, lingual surface, cemento-enamel junction, or roots of the tooth. In instances where the carious lesions are too large to determine where they originated, they can defined as gross carious lesions.

There are multiple approaches to calculate caries rates, each with its own strengths and benefits. However, most studies use a caries rate by individual (Equation 2). Although this approach represents the observed caries frequency, rather than the real caries frequency, and does not take into consideration the impact of antemortem or post-mortem tooth loss, it is the most commonly employed method within bioarchaeology and allows for comparisons to other archaeological sites and samples.

$CariesRate=\frac{Numberofcariousteeth}{Numberofobservedteeth}\times 100$(2)

Equation 2 Caries rate by individual

Observation of caries rates by tooth type, location, or overall rate can help inform researchers about changes in diet, differences between samples or populations, and, with the use of theoretical frameworks, can help tell us about diet in the past.

Stable Isotopes

Beyond macroscopic analyses of dental structures, dietary stable isotope analysis offers a biochemical approach to better understand dietary intake and consumption. This is a destructive approach (i.e., a sample must be destroyed to obtain the associated data) but has been widely used within biological anthropology with applications to bones and teeth. A detailed review of dietary stable isotopes is beyond the scope of this Element, but a thorough introduction can be found in Katzenberg and Waters-Rist (Reference Katzenberg, Waters-Rist, Katzenberg and Grauer2019).

Broadly, stable isotope analysis is based on the phrase “you are what you eat.” As you eat food, your body uses the nutrients and minerals to form the tissues of your body. This means that, by examining the body (e.g., bones, teeth), we can work backwards to learn a bit about the foods you consumed. Now this approach does not produce a menu, listing every item an individual consumed as a child, but can tell us about the stable carbon and nitrogen values, from which bioarchaeologists can make inferences about the types of foods that individual consumed in life.

Particularly beneficial for studies of non-adults is the analysis of nitrogen isotopes, which is used to differentiate between trophic level of foods consumed. As we move up the food chain (e.g., plants to herbivores, to carnivores), nitrogen values should increase by approximately three to five per mil (Katzenberg & Waters-Rist Reference Katzenberg, Waters-Rist, Katzenberg and Grauer2019). In the case of infants, breastfeeding children should be approximately three per mil higher than the individual from which they are consuming breastmilk, as the infants are consuming their tissues and products (i.e., breastmilk). As it is often impossible to associate a deceased breastfeeding infant with the deceased individual from which they were breastfeeding, researchers use the average nitrogen values of adult females as the baseline (e.g., Richards et al. Reference Richards, Fuller and Molleson2006). Infants that are then approximately three per mil higher than this proxy are assumed to be breastfed. By incorporating incremental sections of dentine (i.e., the portion of the tooth that incorporates dietary protein), or sampling across various ages within a sample, researchers can identify when nitrogen values are above threshold, indicating breastfeeding, and when the nitrogen values start to drop to match adult average values, from which they infer when weaning occurred (e.g., Smith et al. Reference Smith, Reitsema, Fornaciari and Sineo2023; Salahuddin & Prowse Reference Salahuddin and Prowse2023).

Timing of Trauma

The timing of trauma is often differentiated in relation to the death event, including antemortem (before death), perimortem (around the time of death), and post-mortem (after death) (Christensen et al. Reference Christensen, Passalacqua and Bartelink2014) (Table 1). These are broad categories that include significant variation. For example, antemortem could include the month before death or twenty years before death. Perimortem could include aspects that contributed to the cause of death, as well as other injuries that occurred shortly before or shortly after death but are unrelated to the death event. Lastly, post-mortem could occur after death while community members are still living, or could include damage inflicted during excavation (e.g., trowel trauma). In the latter example, the damage may not be considered “trauma” but needs to be differentiated from antemortem and perimortem trauma, to help understand trauma and injury, as well as post-mortem environments.

To differentiate between the three timings, key features should be observed. Antemortem trauma should include some evidence of healing whether in the form of new bone growth, a bony callus, or complete unification of the fracture margins. As the bone was alive during this traumatic evidence, the colour should be fairly uniform. Perimortem trauma should also exhibit uniform colouration between the fracture and surrounding bone, but without any evidence of healing. Hinge fractures may be present (Christensen et al. Reference Christensen, Passalacqua and Bartelink2014). Post-mortem trauma often includes colour differences between the fracture site and the surrounding bone, as the two areas have been exposed to the soils around them for different periods of time.

A Note on Child Abuse

In 2022, almost 400,000 incidents of physical child abuse were reported to the National Children’s Alliance in the United States (National Children’s Alliance 2023). In the same year, an estimated 1,990 children died from abuse and/or neglect in the United States alone (National Children’s Alliance 2024). The almost complete absence of any reported cases of child abuse in the archaeological record has led many anthropologists to suggest that child abuse, in its current form, is a recent phenomenon. However, the documentary evidence from past societies would suggest otherwise, with literary evidence of corporal punishment and child abuse having been experienced or delivered by many individuals in the past. Thus, the lack of bioarchaeological data regarding child abuse may not stem from a lack of child abuse, but rather, due to our inability to recognize evidence of abuse in non-adult skeletal remains. Unique cases exist (see Wheeler et al. Reference Wheeler, Williams, Beauchesne and Dupras2013 for a well-defined and supported example), but not at the rates we might expect.

Evidence of child abuse is often interpreted from subtle and non-specific lesions that are left behind, contributing to issues associated with differential diagnosis, and the inability to confidently attribute the osteological markers of experiences of trauma (Martin and Harrod Reference Martin and Harrod2015). Additionally, it would be inappropriate to define malnutrition as child abuse in contexts where malnutrition was rampant throughout a community. Consequently, much of our knowledge regarding lesions commonly associated with child abuse comes from forensic medical literature.

In a systematic review of published reports concerning skeletal traumatic injuries in individuals under the age of eighteen, Kemp et al. (Reference Kemp, Dunstan, Harrison, Morris, Mann, Rolfe, Datta, Thomas, Sibert and Maguire2008) found that no fracture in isolation is indicative of physical abuse, but that there are some consistent patterns that are more common in cases associated with child abuse compared to non-abusive traumatic injuries. These include:

• Presence of multiple fractures

• Presence of rib fractures

• Femoral fracture in those not yet walking

• Humeral fracture in children under the age of three

• Mid-shaft fractures

• Skull fractures, particularly on the parietal or linear skull fractures

Additionally, some features may be of particular interest to bioarchaeologists. For example, multiple fractures exhibiting differential stages of healing may suggest a pattern of repetitive traumatic experiences rather than an isolated incident, corresponding with patterns of continued child abuse rather than an isolated fall or accident (Martin and Harrod Reference Martin and Harrod2015). Additionally, new bone formation on the diaphysis of the humerus or radius and ulna may be consistent with the stripping of the periosteum from the bone, which can happen when an individual is forcefully grasped, pulled, or shaken (Caffey Reference Caffey1974).

When looking to the past, it is possible that many of the issues that affect the bioarchaeological studies of childhood more broadly also limit our ability to identify and assess evidence of child abuse. These can include, but are not limited to the following:

1. 1. Burial location: Children who died from abuse may have been buried in clandestine burials, rather than within the general cemetery. As a result, the remains may not be recovered during archaeological excavations.

2. 2. Sample sizes: Based on modern data, an osteological collection of 2,000 infants and children may include one individual who suffered abuse during their life. While large cemetery samples exist in the bioarchaeological record, few include datasets of that size.

3. 3. Preservation: Based on medical data, diagnosis of child abuse relies on the preservation of osteological elements, including the ribs and cranial bones. With fragmentary remains or poorly preserved remains (or improper excavation of complete skeletons), researchers may not have an entire skeleton to examine and assess patterns of trauma across the entire skeleton.

4. 4. Osteological paradox: As evident in statistics from the United States in 2022 (National Children’s Alliance 2023), most victims of abuse may survive, and their injuries will heal. Ultimately, this means that evidence of child abuse will remain invisible in the bioarchaeological record.

With this in mind, bioarchaeologists looking to explore experiences of child abuse in the past may need to consider burial locations and associated contexts in which child abuse may have been more prevalent. Examples include examining the remains associated with specific social institutions (e.g., workhouses) or those associated with systems of oppression (e.g., subject to slavery).

Biological Changes

A major component of the period of adolescence is the biological changes associated with puberty. Puberty is initiated at the hormonal level, generating physical changes in almost every aspect of the body (Hagg and Taranger Reference Hagg and Taranger1982). Rather than discussing puberty as a single event, researchers often subdivide this period of development into five stages: prepubertal, acceleration, peak height velocity (PHV), deceleration, and post-puberty. While the timing and pace through these stages is dependent on many factors (e.g., genetics, nutrition), the general pattern follows a predictable manner (Tanner Reference Tanner1962; Figure 11):

1. 1. Prepuberty encompasses a cascade of hormonal changes, occurring before any morphological changes take place.

2. 2. Acceleration includes the first outward signs of puberty, including breast budding in females and growth of the testes and scrotum in males.

3. 3. Peak Height Velocity (PHV) is the stage in which we see the most amount of outwardly physical changes, including the adolescent growth spurt, with an average height increase of 9.0 cm/year for girls and 10.3 cm/year for boys (Rogol et al., Reference Rogol, Roemmich and Clark2002). Menarche (a female’s first menstruation) also typically occurs about one year after PHV.

4. 4. Deceleration is when the rate of physical changes begins to slow and corresponds with spermatogenesis and the dropping of the male voice.

5. 5. Post-puberty corresponds to the time when all vertical growth is achieved, and breast and genital development is complete. Some growth in width and breadth (e.g., pelvis), as well as increased musculature may still occur after this point.

Figure 11 Sequence of pubertal events in boys and girls (adapted from Tanner Reference Tanner1986, 433). PHV – Peak Height Velocity.

While we often associate puberty with changes in soft tissue or hormones, there are several associated osteological changes because skeletal and somatic maturity are influenced by the same biological systems (Dreizen et al. Reference Dreizen, Spirakis and Stone1967; Ford et al. Reference Ford, Khoury and Biro2009). Specifically, the pituitary and gonadal secretions responsible for the onset of puberty are also responsible for the ossification of the epiphyseal cartilage and subsequent growth of bones (Demirjian et al. Reference Demirjian, Buschang, Tanguay and Kingnorth1985). Saggese et al. (Reference Saggese, Baroncelli and Bertelloni2002) state that, beyond developing soft tissue, puberty plays a key role for bone development in adolescents.

Pubertal Timing Methods

The use of skeletal maturity to estimate pubertal stage has been most strongly advocated for by orthodontists, as orthodontic treatments (e.g., dental braces, headgear) are most successful when applied during a period of rapid growth, like the pubertal growth spurt (Uysal et al. Reference Uysal, Ramoglu, Basciftci and Sari2006). By looking at key features on x-rays, orthodontists attempt to predict the timing of the pubertal growth spurt, determine the growth velocity during PHV, and estimate the proportion of growth remaining, in order to know when to best apply orthodontic treatments (Santiago et al. Reference Santiago, Coasta, Vitral, Fraga, Bolognese and Maia2012). Over the years, several skeletal features have been successfully linked with puberty, including hand and wrist maturation, canine mineralization, iliac crest fusion, and cervical vertebrae maturation. Initially applied to x-rays, the methods were subsequently adapted for use on dry bone and summarized by Shapland and Lewis (Reference Shapland and Lewis2013, Reference Shapland and Lewis2014).

1. 1. Canine

Mandibular canine mineralization is assessed following Demirjian et al. (Reference Demirjian, Buschang, Tanguay and Kingnorth1985), identifying teeth from stages D (crown complete) through H (roots complete) (Figure 12). Using x-rays can help better identify root development in cases where the tooth is in occlusion, otherwise, the root cannot be fully observed and should be recorded as unobservable.

Figure 12 Mineralization of the canine tooth according to Demirjian stages (Demirjian et al. Reference Demirjian, Buschang, Tanguay and Kingnorth1985). Stage D: Crown complete. Stage E: Crown height is greater than the length of the root. Stage F: Apex ends in a funnel shape; root is greater in length than crown height. Stage G: Apical end is still partly open. Stage H: Apical end of the root canal is completely closed.

Clinical studies have demonstrated that Demirjian stage F for the mandibular canine correlates with the initiation of puberty, while Demirjian stage G corresponds to immediately prior to PHV (Demirjian et al. Reference Demirjian, Goldstein and Tanner1973; Coutinho et al. Reference Coutinho, Buschang and Miranda1993; Chertkow and Fatti Reference Chertkow and Fatti1979). Shapland and Lewis (Reference Shapland and Lewis2013) suggest that stage H coincides with PHV or immediately after.

1. 2. Hamate

The hook of the hamate is assessed following Tanner et al. (Reference Tanner, Healy, Goldstein and Cameron2001) with the additional stage recommended by Shapland and Lewis (Reference Shapland and Lewis2013), as more subtle changes are observable on dry bone than in clinical x-rays. Stages of hamate development include undeveloped (stage G), appearing (stage H), developing (stage H.5), or complete (stage I) (Figure 13).

Figure 13 Development of the hook of the hamate.

According to Chertkow (Reference Chertkow1980) and Grave and Brown (Reference Grave and Brown1976), stage G suggests an individual is prepubertal, while stage H (or H.5) indicates the acceleration stages of pubertal growth spurt, and stage I indicates that PHV has been achieved.

1. 3. Phalanges

With regards to pubertal status, fusion of the phalanges is assessed as unfused, partially fused, or fused. According to Grave and Brown (Reference Grave and Brown1976), the point where the phalangeal epiphyses are equal in width to the metaphysis and begin to cap the metaphysis is correlated with PHV, as is the initial fusion. However, unfused epiphyses are infrequently recovered in archaeological contexts and may not be properly associated to the correct digits. Thus, rather than considering the capping or the width of the unfused epiphyses, fusion of the phalangeal epiphyses is adopted as an indicator that the deceleration phase of puberty had already begun (Hagg and Taranger Reference Hagg and Taranger1982).

Furthermore, the fusion of the distal phalanx of the second finger has been found to closely correlate with menarche (Buehl and Pyle 1942). However, as this particular digit may be difficult to identify, Shapland and Lewis (Reference Shapland and Lewis2013, 303) suggest that early fusion of the distal phalangeal epiphyses (where a fusion line is still present) indicates that menarche had recently passed, while complete fusion indicates the individual is of a post-menarcheal age.

1. 4. Arm Bones

The distal radius, distal humerus, and proximal ulna are also assessed as unfused, partially fused, or fused. According to Hagg and Taranger (Reference Hagg and Taranger1980, Reference Hagg and Taranger1982), a fusing distal radius indicates the individual is in the deceleration stage, while complete fusion indicates the individual is post-pubertal. Initial fusion of the proximal ulna and distal humerus corresponds to PHV in both males and females (Roche Reference Roch and Fuchs1976), suggesting that complete fusion of these elements indicates the deceleration and post-pubertal stages.

1. 5. Iliac Crest

In clinical studies, the amount of ossified but unfused iliac crest is used to differentiate stages of puberty (Risser Reference Risser1958). However, this is not easily applied in bioarchaeology, as ossified but unfused iliac crests are infrequently recovered due to their fragile nature (Figure 14). Thus, for use in bioarchaeology, a simplified approach is taken, with the iliac crest assessed as unfused and not present, ossified but unfused, fusing, or fused. In this division, an ossified but unfused iliac crest indicates the individual is within the deceleration stage, while complete ossification and fusion indicates the individual is post-pubertal (Biondi et al. 1985). However, an unfused and absent iliac crest is not indicative of a particular pubertal stage, as it could have been ossified and simply not recovered archaeologically or may not have begun to ossify at all.

Figure 14 Ossified but unfused iliac crests are rare in the archaeological record, in part due to the fragile nature of the iliac crest that results in significant post-depositional damage. Osteological element from Lisieux-Michelet, France.

Photo by author.

1. 6. Cervical Vertebrae

Cervical vertebrae maturation (CVM) is assessed, following Hassel and Farman (Reference Hassel and Farman1995), by assessing the general shape and size of the third cervical vertebra (C3). In this method, changes of the cervical body correspond to the entire pubertal growth spurt. However, in clinical and bioarchaeological studies, this method has produced inconsistent results with other pubertal stage indicators (see Santiago et al. Reference Santiago, Coasta, Vitral, Fraga, Bolognese and Maia2012; Arthur et al. Reference Arthur, Gowland and Redfern2016 for clinical and bioarchaeological studies that discuss this further). One possible explanation for these discrepancies is that CVM divides the period of puberty into six stages rather than five (Ozer et al. Reference Ozer, Jama and Ozer2006). Although most clinical researchers recommend combining stages 4 and 5 with the deceleration stage, this inconsistency may contribute to the variable results (Shapland and Lewis Reference Shapland and Lewis2014). Within bioarchaeology, it is generally acceptable for the CVM stage to be slightly out-of-line with other pubertal stage indictors, acknowledging the variability and inconsistency associated with this osteological feature. The stages of puberty and observed CVM characteristics are outlined in Table 2.

Table 2 Stages of cervical vertebra maturation

Stage	Inferior border	Body shape
Initiation (1)	Flat	Wedge-shaped
Acceleration (2)	Concavity appears	Nearly rectangular
Transition (3)	Developing	Rectangular
Deceleration (4)	Distinct concavity	Nearly square
Maturation (5)	Accentuated concavity	Square
Completion (6)	Deep concavity	Taller than it is wide

1. 7. Pubertal Assessment

Once all features are assessed, an individual is placed into a pubertal stage-at-death. Divisions based on six or seven categories are available elsewhere (Falys and Lewis Reference Falys and Lewis2020). Table 3, however, provides a division according to five stages. Compared to other approaches, dividing data across five categories allows us to maximize sample sizes in each category and more closely aligns with medical descriptions of puberty according to the Tanner Stages (Marshall and Tanner Reference Marshall and Tanner1969, Reference Marshall and Tanner1970) or Sexual Maturity Ratings. To be assigned a pubertal stage, an individual should have at least three features present and in agreement with one another. As CVM is known to produce slightly inconsistent results, it is not considered problematic if CVM is one stage ahead or behind other features. In these instances, priority is given to other features, provided there were enough features to properly assess the pubertal stage.

Table 3 Pubertal assessment based on osteological features

	Pre-puberty	Acceleration	PHV	Deceleration	Post-puberty
Canine	Stage D–F	Stage G	Stage H	Stage H	Stage H
Hamate	Stage G	Stage H/H.5	Stage I	Stage I	Stage I
Phalanges	Unfused	Unfused	Unfused or partly fused	Partly fused or fused	Fused
Radius	Unfused	Unfused	Unfused	Partly fused	Fused
Ulna	Unfused	Unfused	Partly fused	Fused	Fused
Humerus	Unfused	Unfused	Partly fused	Fused	Fused
Iliac crest	Unfused	Unfused	Unfused	Ossified and unfused	Partly fused or fused
CVM	Stage 1	Stage 2	Stage 3	Stage 4/5	Stage 6

In some instances, whether due to small sample sizes, or fragmentary remains, researchers may need to consider exploring samples based on pre-PHV and post-PHV, rather than by five discrete stages. While this approach hides some of the variability observed within puberty, it can help maximize sample sizes to permit more robust statistical analyses.

Social Changes

As the study of adolescence has only recently developed, the methods currently employed to investigate social age changes are limited. No adolescent-specific bioarchaeological approaches have been developed to date, but rather existing methods are being applied to this particular age group, including mobility isotopes (e.g., Lewis and Montgomery Reference Lewis and Montgomery2023), dietary stable isotopes (e.g., Avery et al. Reference Avery, Brickley, Findlay, Chapelain de Seréville-Niel and Prowse2021, Reference Avery, Brickley, Findlay, Bondioli, Sperduti and Prowse2023b), analysis of mortuary patterns (e.g., Scott et al. Reference Scott, MacInnes, Hughes, Munkittrick, Harris and Grimes2023), and investigation of pathological conditions (e.g., Pererva Reference Pererva2017). As the discipline continues to develop and grow, so too will the approaches used to understand this period of the human life course, offering exciting avenues of future research into the lives of adolescents and young people in the past.

Perinatal Bioarchaeology

As we continue to consider the lives of non-adults, bioarchaeologists are exploring those at the periphery with greater interest. To date, this has included a wave of research regarding adolescents (e.g. see special issue Bioarchaeology International: Emerging Adolescence, edited by Creighton Avery, Megan Brickley, and Mary Lewis); however, the youngest members of society remain elusive, existing “on the margins of discussion” (Hodson Reference Hodson2021). Fetal, perinatal, and infants are often viewed as those without agency or identity, as they have not necessarily lived within their communities, societies, or households (Halcrow Reference Halcrow, Gowland and Halcrow2019). However, as Hodson (Reference Hodson2021) highlights, these individuals uniquely capture both social and biological data. Biologically, their bodies capture in utero conditions, often reflecting maternal health conditions. In a study of perinatal remains from the Spring Street Presbyterian Church burial vaults (New York City; 19th century CE), Ellis (Reference Ellis2020) identified perinatal remains, including a mother–infant pair (based on osteological remains and corresponding records). The remains of the mother exhibit evidence of a cariogenic diet and tobacco staining along their teeth; the infant in the same burial was likely exposed to the same harmful substances (Ellis Reference Ellis2020, 197).

Socially, the burial of a perinate may speak to the treatment of the youngest and most vulnerable members of a community (Scott & Betsinger Reference Scott, Betsinger, Han and Tomori2021). In a study of a burial associated with the Middle Holocene in Brazil, Solari and colleagues (Reference Solari, Pessis, Martin and Guidon2020) discuss the remains of perinatal remains that were afforded the same burial treatment as other non-adults and adults within the burial complex, suggesting a level of social identity within its group.

By capturing biological and social conditions, analysis of perinatal remains can provide unparalleled insights into pre- and post-natal experiences within different sociocultural, temporal, and economic environments. Methodological developments in terms of age estimation, growth disruptions, and pathology will contribute to our ability to understand these youngest members, and allow us new avenues of exploration, such as exploring birth experiences, including stillbirths (Booth et al. Reference Booth, Redfern and Gowland2016). During the birthing process and breastfeeding, bacteria enter the gastrointestinal tract. However, stillborn infants, who are not exposed to these bacteria through the gastrointestinal tract, do not have evidence of microstructural changes caused by bacterial bioerosion. A study by Ullinger and colleagues (Reference Ullinger, Gregoricka, Bernardos, Reich, Langston, Ferreri and Ingram2022) used micro-CT scans to examine evidence of bioerosion and ancient DNA to assess a biological relationship, concluding that a double burial was, in fact, that of fraternal stillborn twins.

Exploring Childhood in Adult Remains

As we continue to develop biochemical and morphological methods, researchers are identifying new ways to explore experiences of childhood that are trapped or captured in adult remains. The most salient example is that of stable isotopes. Considering that teeth develop during childhood and remain relatively static through adulthood, researchers can extract teeth from individuals who died as adults, but observe isotopic values captured during childhood. Such approaches allow researchers to consider the Osteological Paradox (i.e., by exploring dietary consumption patterns of those that died in childhood, versus those who survived the period of childhood), incorporate more reliable sex estimation methods, and expand sample sizes, if necessary. Beyond biochemical approaches, some pathological conditions or manifestations (e.g., vitamin D deficiency and rickets) may be observed in adult remains, allowing us to consider conditions in childhood that affected later adult life. According to Mays et al. (2017), integrated studies of non-adult and adult skeletal remains will only contribute to a more holistic view of life in the past. Continued approaches to studying the last effects of trauma, occupational changes, and growth disruptions will continue to expand our understanding of childhood for those that survived this often-tumultuous period of life.

Age Estimation

To ensure the greatest sample size can be included, multiple age methods can be used, acknowledging that not all features are consistently present in bioarchaeological collections. In instances where the various methods produce conflicting results, dental development and eruption should be given priority, followed by epiphyseal fusion, and lastly, linear growth (White & Folkens 2005).

Biological Sex Estimation

Dental Health

Examine dentition (teeth and alveolus, if available), and record the dental presence and presence of carious lesions. You can use standard designations according to Buikstra and Ubelaker (1994, 48–55) or develop your own simplified coding system that captures all required information. For example:

Presence/Absence: P – present, AP – absent, post-mortem loss (tooth absent, alveolus present and open), AA – absent, antemortem loss (tooth absent, alveolus present but complete or partial crypt closure indicates antemortem tooth loss) U – unobservable (tooth and alveolus not present).

Carious lesions according to surface affected: I – interproximal, O – occlusal, R – below the cemento-enamel junction, B – buccal/labial, L – lingual, G – gross (when more than one surface is affected).

Maxillary Dentition

Mandibular Dentition