An Evaluation of Ancestral Diversity in 19th-Century South Shields

SUMMARY

Bioarchaeological analysis of human remains from St Hilda’s Church cemetery, South Shields was undertaken to investigate ancestral diversity in post-medieval Britain. Methodological evaluations argue that when diversity is explored through phenotypic traits, biological affiliations are best investigated through dental morphology due to its resistance to exogenous influences, whereas cranial form appears to be increasingly influenced by environmental and cultural factors in the post-medieval period. A novel quantitative model was developed for a within-group analysis of dental morphology assessed through the Arizona State University Dental Anthropology System, successfully identifying an ancestral outlier. An osteobiographical narrative for this individual, combining osteological and historical data with social theory, suggests that personhood is malleable and multifaceted. Ancestry is therefore most appropriately understood as one aspect of identity which also included sex, occupation, social status and linguistic affiliation.

INTRODUCTION

Post-medieval Britain was an ancestrally heterogenous society, with diversity increasing in magnitude due to Britain’s industrialising economy, colonialism and mercantile ventures.¹ Though a small number of primary sources relate the importance of ancestry in a post-medieval context,² historical research into ancestral diversity has been circumscribed by a general paucity of evidence and ancestry-focused osteoarchaeological investigations have been rare.³ The scarcity of osteoarchaeological research reflects the difficulties encountered when estimating ancestry through skeletal remains; past projects have been criticised for over-simplifying phenotypic diversity and perpetuating unethical, classificatory paradigms.⁴ It has also been debated to what extent biological affinities affect skeletal morphology in comparison to environmental pressures,⁵ and how well estimates of ancestry correspond to sociocultural associations.⁶

New analysis of skeletal remains excavated from post-medieval cemeteries and archived with museums and universities creates the opportunity to contribute to this understudied aspect of post-medieval life. The aim of the present study was to collect and statistically interpret dental morphological data from one such assemblage – the remains from St Hilda’s Church in South Shields – to critically evaluate methods of ancestry estimation and contextualise the nature and role of ancestral diversity in a post-medieval urban population. The aim was achieved through three research objectives: 1. to identify phenotypic outliers using dental non-metric traits and novel statistical procedures, and evaluate the further applicability of such exploratory techniques; 2. to employ a variety of methods to estimate ancestry and critically compare the suitability and efficacy of the techniques with reference to post-medieval skeletal assemblages; 3. to develop osteobiographies for any individuals of diverse ancestry (identified as statistical outliers) to explore attitudes to ancestral diversity in relation to other elements of personhood in post-medieval Britain.

BACKGROUND

THEORETICAL AND METHODOLOGICAL CONTEXT OF ANCESTRY ASSESSMENT

The frequent collocation of ‘ancestry’ with terms such as ‘race’ and ‘ethnicity’ render it difficult to delineate and commonly misunderstood.⁷ Recently, Kranioti and colleagues described ancestry as ‘an individual’s ancestral region of geographic origin.’⁸ Though this is a broad, perhaps ambiguous, definition, it does imply the time-depth and spatial separation requisite for the development of ancestral groups which are created when populations become geographically isolated and genetic admixture is limited.⁹ It is thus theorised that as geographic distance increases between human populations, so will genetic dissimilarity, though genetic heterogeneity can be accentuated or reduced by factors such as selective environmental pressures and culturally determined mating patterns.¹⁰ Such dynamic processes underpin genetic differences between population groups, yet, the subtle markers used in differentiation are ever-changing and no fixed boundaries exist.¹¹

Ancestry, therefore, represents an individual’s biological affinity; although perceptions of race can closely correspond to ancestral affinity, race has no valid scientific foundation and can be viewed as a social construct.¹² Race is a manipulable concept that has been used to oppress and segregate, and was arguably scientifically reified in 1735 when Carl Linnaeus separated humans into four subspecies distinguished by differences in their characters and physical features.¹³ Driven by the pervasive theory of biological determinism, many 19^th-century scientists perpetuated this classificatory trend, contending that an individual’s innate moral quality and intellectual capacity were defined and delimited by their physical being.¹⁴ Even famous abolitionists rarely believed that peoples of different ancestral origins were equal; Charles Darwin referred to what he perceived as the ‘civilized races’ and ‘savage races’ of man.¹⁵ Although this theory was not restricted to racial categorisations, the concept of inherited qualitative differences in humans has permeated societies and underpinned segregationist policies that represent a form of structural violence aimed at limiting the opportunities of individuals believed to be inherently inferior.¹⁶ In a conscious effort to break from the racist character of formative scientific models of race and ancestry, which ascribed membership to a specific group and emphasised differences between groups, modern critical approaches instead explore a collective diversity or a range of human variation.

ESTIMATION OF ANCESTRY: TECHNIQUES AND RATIONALE

In archaeological settings attempts have been made to explore diversity through the analysis of grave goods and mortuary rites – the logic being that people whose origin lay in a different region could express ancestral affinities through clothing, accoutrements and funerary custom. This, though, can be problematic. Several individuals in Viking Age Orcadian burials, for example, were initially identified as Norse settlers based on the cut and alignment of their graves as well as the goods contained therein. Yet, with the benefit of aDNA testing it has been revealed that in some cases these individuals did not in fact possess a Scandinavian genotype but were most likely Picts that had adopted – or at least had been buried with – the markers of a Norse cultural identity.¹⁷ These cases highlight two points. Firstly, that cultural or ethnic identity is malleable and fluid; it may reflect personal agency and social processes such as assimilation and diversification. Secondly, that the most effective means of investigating ancestry and biological diversity is through the direct exploration of genetic materials.

Due to constraints in access and funding as well as the technical difficulties associated with extracting aDNA from diagenetically degraded materials, however, most archaeological research projects cannot directly investigate genetic heterogeneity through genomic analyses and usually rely on phenotypic proxies.¹⁸ Phenotypic traits, most often in the cranium or dentition and defined either metrically or non-metrically, are formed through a varying combination of genetics, environmental influences and culturally determined behaviours.¹⁹ Non-metric traits, also known as epigenetic and quasi-continuous traits, are phenotypic idiosyncrasies in skeletal form which have been linked to ancestry due to their heritability and can be described qualitatively but are difficult or impossible to consistently measure.²⁰ Metric traits, in contrast, can be quantitatively described by measurements taken between two clearly defined landmarks – though applicable to post-skeletal and dental materials, this method is most frequently applied to cranial morphology when estimating ancestry.²¹ Phenotypic traits characterise a broad range of anatomical variation and have a strong biological underpinning; they do, however, present certain problems when investigating ancestry.

Cranial non-metric traits, for instance, have been successfully used to explore ancestry for many years. Even so, as traits are not confidently discernible until an individual reaches skeletal maturity and no trait is exclusively associated with any ancestral group (see example in Figure 1), investigations are limited to adult populations and investigator experience is fundamental in assessing subtle and subjective differences with a high degree of within-group variation.²² Furthermore, as the interpretative schemata through which cranial non-metric variation is analysed were largely developed at a time when social prejudices permeated science, they have been accused of perpetuating a racist and outdated view of ancestry in which people are coarsely grouped into ‘White’, ‘Black’ or ‘Asian’ categories.²³ In reaction to such criticisms, statistical techniques have been developed to explore cranial variation – especially metrically – with the aim of achieving greater objectivity and inferential nuance. Programs like CRANID and FORDISC, for example, employ the measurements advocated by Howell to assess the length, breadth and height of various cranial features to compare adult crania to populations of known origin.²⁴ CRANID is a particularly flexible program through which it is possible to conduct either a parametric Linear Discriminant Analysis (LDA) or non-parametric Nearest Neighbour Discriminant Analysis (NNDA). The former calculates the probability of group membership by comparing the weighted sum of an individual’s cranial measurements to mean values from sample populations, while the latter compares unstandardized scores to identify the closest matching crania.²⁵ Nevertheless, craniometric databases do not yet comprehensively account for diachronic and geographic variation which can reflect significant levels of environmental and cultural influences.²⁶

Fig. 1 Cranial non-metric traits in the nasal region are often used in ancestry estimation. These two crania illustrate variance in nasal bone contour: a) steepled, almost triangular, contour is more typical in individuals of European ancestry; b) low, rounded nasal bones are more frequently found in people of African ancestry; this individual, however, also has a pronounced nasal spine (red arrow) that is more common in European populations.

In contrast, dental non-metric traits are less affected by external influences.²⁷ Certain dental traits occur more or less frequently in ancestral groups due to assortative mating patterns and genetic isolation, and are easily recorded using the Arizona State University Dental Anthropology System (ASUDAS).²⁸ As such, it is generally believed that dental morphology represents the best alternative to direct genetic testing with some dental traits having an estimated heritability of >90%.²⁹ As with the analysis of cranial form, the frameworks through which dental non-metric traits are being interpreted are becoming increasingly complex. While in the past many researchers took a hypothetico-deductive approach to estimating ancestry, the newly developed rASUDAS program uses a Naïve Bayes classifier algorithm that assumes conditional independence between traits to calculate the probability of an individual belonging to a specific biogeographic population cluster.³⁰ Dental non-metric analyses are not without limitations though. As the observation of dental traits is typically restricted to the permanent dentition and as cultural and dietary factors – such as occlusal wear and dental caries – may still erode or obscure the dentition, many individuals are excluded from analyses.³¹ It is also debatable as to how precisely dental non-metric traits can differentiate biological affinities; some researchers have sought to identify specific familial and kinship clusters within cemeteries, whilst others more cautiously suggest that an individual or assemblage broadly confirms to a morphological pattern associated with much larger, regional population groups.³² Consequently, as no single method is free of uncertainties and problems, to achieve a substantive estimation of ancestry it is recommended that several methods are applied, though this can create contradictory results.³³

Given such methodological uncertainties, it may be wondered why ancestry is estimated at all. On a practical level, ancestral biases exist in most methodologies used to develop sex, age and stature data,³⁴ so osteoarchaeological analyses that do not consider the effects of ancestry could potentially generate skewed interpretations. Furthermore, ancestry can form an integral part of a person’s self-identity,³⁵ influence their life experiences and susceptibility to pathological conditions,³⁶ and the deep-rooted propensity for humans to recognise and respond to diversity renders any study of an individual or society incomplete if ancestral variation is not sociologically considered.³⁷ As the New York African Burial Ground Project demonstrates, the consideration of ancestry in a bioarchaeological framework can be informative and socially engaging.³⁸ In this project, researchers moved from a purely descriptive investigation of ‘ahistoric and acultural’ biological data and took an interdisciplinary approach.³⁹ Historical and ethnographic sources enhanced the interpretation of osteological remains, permitting a more meaningful and inclusive reflection on the social history of an underrepresented sector of society whilst revealing the complexity of human skeletal and cultural variation.⁴⁰

In summary, modern research into ancestry has struggled to escape its embittered past and efforts must always be made to prevent the propagation of categorical and racist paradigms on an impossibly large range of human diversity.⁴¹ Yet, reflecting holistically on the nature of human variation and the embodied experience of ancestrally diverse populations, including migrants, has the capacity to illuminate poorly explored aspects of the past and make historical research more inclusive.

HISTORICAL BACKGROUND: MIGRATION, REACTIONS TO DIVERSITY, AND SOURCES

Post-medieval ancestral diversity was strongly related to economic and political disparities that developed during the period and, as Figure 2 depicts, the rapidly industrialising ports and cities of Britain became centres of migration.⁴² Previously, the effects of such large scale and varied movements of peoples have largely been investigated through official records and biographies, though both sources pose problems. Official records tend to be biased towards London and inconsistently record ancestral affinity.⁴³ Furthermore, they do not investigate the lived experience of individuals and families in question but, instead, narrowly focus upon stereotypical fears surrounding migration: indigency, spread of disease and civil unrest.⁴⁴ Biographies, in contrast, have been used to add humanising detail, describing the hardships and transformative processes experienced, but by definition are not generalisable.⁴⁵ Hence, it is difficult to capture the full impact ancestral diversity had on society or the individual.

Fig. 2 Increasing numbers of migrants, many fleeing economic hardship and political uncertainty, were seen in 19th century British port cities, increasing levels of biological and cultural diversity. Illustration from Harper's Weekly, 7th November 1874.

Despite its limitations, historical data does paint an interesting picture of diversity, which appears to have reached new levels in the 18th and 19th centuries.⁴⁶ Indeed, the magnitude of Irish and Scottish economic migrants provoked government commissions to establish how these influxes affected society and labour markets,⁴⁷ while census data reports that 30,000-60,000 Germans entered Britain during the mid-19th century to escape famine, persecution and poverty.⁴⁸ Additionally, as international trade increased in volume and frequency, especially between British colonies in the West and East Indies, there was an influx of foreign sailors of West African, Indian, Southeast Asian and Pacific origin known as lascars.⁴⁹ Contemporary reports estimated that 10,000-12,000 of these seamen manned British ships during the mid-19th century and, as the government began to enforce a restriction on outgoing British vessels that required a certain percentage of the ship’s company to be of British origin, many of these sailors were compelled to remain in Britain for long periods – of these stranded lascars, approximately 140 perished per year as a result of the poor rations and the unsanitary living conditions in which they were expected to live.⁵⁰ The expansion of international trading routes and global markets also facilitated the growth of the transatlantic slave trade and British ships – operating out of Bristol and Liverpool in particular – were involved in the transportation of millions of enslaved West Africans between the 16th and 19th centuries.⁵¹ Although most were forcibly relocated to colonies in the Caribbean and Americas, some found themselves in Britain – tombstones dedicated to enslaved Africans can still be found in Bristol’s Henbury Churchyard, for instance.⁵² Enslaved individuals occupied a variety of positions, but it became a marker of social status for wealthy families to have Africans in their households and those engaged in the domestic sphere are perhaps most visible in the historical record.⁵³ Domestic service in general was a major employer for female migrants and it was not uncommon for servants and nursemaids contracted in India and Asia to accompany their employers when they returned to Britain.⁵⁴ Though the scale of migration may seem surprising, to promote free trade the British government did not impose limitations on immigration until passing the 1905 Aliens Act, and it was only in the 1960s that legislation was enacted to substantially restrict entry on effectively racial grounds.⁵⁵

Many incomers that settled in England were economic migrants, frequently fleeing famine or hardship, and were compelled to undertake employment in low-paid, hazardous occupations as well as live in insalubrious, disease-ridden environments.⁵⁶ It is perhaps no surprise then that the 19th-century Manchester Constables Accounts reported the giving of charitable payments to poor and sickly individuals from such disparate places as Constantinople and Bengal.⁵⁷ There is also evidence of deliberate abuse against migrants, enslaved individuals and peoples of non-British ancestry for whom the protection of the law could not be guaranteed.⁵⁸ Although in some situations legal protection was forthcoming – such as a case in 1800 in which a woman was summoned to appear before a Surrey court to be charged with the cruel treatment of her Indian maid – the legal status of enslaved peoples, for example, was highly ambiguous and judicial decisions on the subject were often diametrically opposed, popularly misinterpreted, or so narrowly phrased that their impact beyond a particular case was hard to discern.⁵⁹ Nevertheless, societal attitudes and prejudices did not always prove to be an insurmountable barrier.⁶⁰ So, though many ancestrally distinct individuals were forced to occupy socially liminal positions as indentured labourers, political exiles or transients, some managed to overcome social hostilities and rise to positions of prominence, as illustrated in the biographies of Sake Deen Mahomed and Olaudah Equiano.⁶¹ Sake Deen Mahomed was born in India in 1759, but relocated to Britain where he married an Anglo-Irish wife and established a successful business in therapeutic medicine which was acknowledged by medical professionals and patronised by royalty.⁶² Olaudah Equiano’s (see Figure 3) biography, meanwhile, recounts how after having been enslaved as a child, he negotiated his own freedom and became a renowned abolitionist, politician and social celebrity.⁶³

Fig. 3 Portrait of Olaudah Equiano, a prominent Britain of non-European origin, taken from the frontispiece of his 1789 autobiography.

Despite being incompletely documented, it can be said that post-medieval British commercial centres were socially dynamic and, for some residents, ancestry played a fundamental role in shaping personhood and the life course.⁶⁴ However, attitudes towards diversity were ill-defined and there would have been great variety in both the origins and experiences of migrants and their families.⁶⁵ On the one hand, migrants could achieve great social and financial successes, but many were also employed in strenuous occupations and lived in sub-optimal conditions.⁶⁶ The lack of comprehensive documentary data can, in part, be alleviated through interdisciplinary bioarchaeological investigations which draw upon osteological analyses and archaeological theory as well as historic sources.⁶⁷ Skeletal material from port cities, such as South Shields, present viable research opportunities for such bioarchaeological explorations. Due to its prominent location and connection to maritime networks, South Shields during the 19th century was a demographically and industrially expanding city with collieries, chemical factories, shipyards and one of Britain’s largest international ports (see Figures 4 & 5).⁶⁸ Investigation of mid-19th century documents report international migrants in the Tyneside area,⁶⁹ while census data for contemporary Liverpool – a comparable port city – indicates that 1.5-2.0% of the population was not British.⁷⁰ Migration of non-Europeans to South Shields had reached such levels by the early-20th century that riots erupted protesting the associated increase in workplace competition and cultural differences.⁷¹ One local paper went so far as to state that, ‘the import of coloured seamen and firemen is a more serious menace to the health of the community… and should be prohibited.’⁷²

Fig. 4 The location of South Shields on the northeast coast made it a centre of maritime trade and industry during the 19th century. Illustration of the Tyneside area taken from Haswell.

Fig. 5 a) South Shield’s location; b) a historic map of the city.

MATERIALS

THE CORONATION STREET ASSEMBLAGE

Rescue excavations at the site of St Hilda’s Parish Church in South Shields, conducted by Oxford Archaeology North between 2006-2007, uncovered 204 individual skeletons.⁷³ The skeletal collection, known as the Coronation Street assemblage, is curated by the Department of Archaeology at the University of Sheffield.

Prior to submission to the University of Sheffield, a medium-resolution osteological analysis was conducted on the articulated skeletal remains to establish a demographic and health profile for the assemblage and highlight any individuals with notable pathologies or indicators of lifestyle – assessment of ancestry was not attempted at that time. A review of this report was undertaken to establish whether the Coronation Street assemblage was an appropriate collection through which to explore ancestral diversity. The osteological analysis revealed that the remains of most adults (n=91/117) were well-preserved with clear or only slightly eroded surface morphology, just over half the adults (n=60/117) were >75% complete, there were 100 individuals with surviving dentition, and that there were a total of 1,289 teeth within the assemblage.⁷⁴

Of the 117 adult remains, sex could be estimated for 103 individuals; this revealed a near equal split of males (n=51) and females (n=52) with a very slight female bias. The original investigators noted that the marginal female bias was not generally typical, but also not altogether unusual when compared to other assemblages from the Newcastle area and historic data from the region.⁷⁵ Age-at-death estimation in the adult population showed a peak in adult mortality in the ‘mature adult’ and ‘mature-older adult’ age categories, which were respectively defined as 36-45 and 36-45+ years of age. On the basis of these results it was suggested that the population had a life-expectancy of between 36-45 years, which is typical when compared to contemporary assemblages, though historic documents for Britain and the Tyneside area indicate that some individuals survived past their ninetieth year.⁷⁶ From long bone lengths, the mean male stature was estimated to be 171 cm with a range of 149-181 cm; female mean stature was estimated as 159 cm with a range of 149-174 cm.⁷⁷

A review of the assemblage report and historical documents revealed that, though St Hilda’s church was first recorded in the early 15th century, the excavated section was in use between the mid-18th and mid-19th centuries.⁷⁸ The burials were largely of lower socioeconomic status and most likely associated with labourers employed in local industries.⁷⁹ The existence of family vaults for more prosperous residents suggested an organised and symbolically invested burial rite.⁸⁰ Historical research has indicated that individuals from culturally diverse backgrounds may have been inclined to follow a non-Anglican or non-Christian religion,⁸¹ however, St Hilda’s was the only burial ground serving the community until the latter half of the 19th century and therefore seems to have provided for all residents.⁸² As a consequence, from the late-18th century, the cemetery was built up and expanded on several occasions in response to population growth until it was eventually officially closed in 1855, though it appears that some interments – possibly in family plots – occurred as late as the 1860s.⁸³

In summary, adult skeletal remains were well-preserved, many individuals had surviving dentition, there was a contextually typical demographic profile, and people of different backgrounds living within the environs of St Hilda’s church were unlikely to have been buried elsewhere. On the basis of this it was surmised that the Coronation Street assemblage presented a viable opportunity to explore ancestral diversity through dental non-metric traits in the South Shields population.

METHODS

This project consisted of two phases. Initially, a quantitative model was developed to identify significant deviations in dental non-metric trait expression within the skeletal population. Statistical outliers to this model represented individuals with traits that are non-normative, and therefore the most likely to be of non-local ancestry. Stage two was the construction of an osteobiographical narrative for outliers.

SELECTION AND STATISTICAL ANALYSIS OF DENTAL TRAITS

Due to their strong genetic component and archaeological survivability,⁸⁴ dental traits were chosen in this initial phase of analysis to detect ancestral outliers. A total of 20 observations were made per individual, on the enamel surfaces of permanent first incisors, first molars and second molars (see Table 1). Many of these traits, especially supernumerary cusps, are more commonly found in individuals with an ancestral affiliation to Sub-Saharan African, Southeast Asian and Pacific populations, whose presence in 19th-century Britain has been inferred from historical sources.⁸⁵ Traits such as Bushman canine and enamel pearls were not included (see Figure 6); although the former has a very strong association with specific ancestral group – the Khoisan – historical documents do not suggest that individuals from this population were likely to have been present in 19th-century South Shields, whilst for the latter trait specific populational frequencies have not been established.⁸⁶

Fig. 6 Small spherical masses of enamel on the tooth root, known as enamel pearls, are a dental non-metric trait whose occurrence in populations is poorly recorded and is of limited use in exploring ancestral diversity.

Table 1. Dental non-metric traits chosen.

Teeth	Trait	Teeth	Trait
UI1	Winging	LM1 & LM2	Hypoconulid
UI1	Shovelling	LM1 & LM2	Cusp 6
UM1 & UM2	Hypocone	LM1 & LM2	Cusp 7
UM1 & UM2	Carabelli’s trait	LM1 & LM2	Protostylid
UM1 & UM2	Cusp 5	LM1 & LM2	Groove pattern
UM1 & UM2	Parastyle

Traits were scored along the ASUDAS ordinal scale with reference to standardised plaques (example shown in Figure 7), pictures and written descriptions in a Microsoft Excel 2016 spreadsheet.⁸⁷ The ASU system is an efficient, easily interpretable format that promotes transparency in research methods and evaluation by external observers.⁸⁸ When available, both left and right dental quadrants of an individual’s dentition were observed; in the case of antimeric asymmetry, the greatest expression of a trait was scored, as it is believed that this represents genotype more accurately.⁸⁹ Trait scores were converted to binarized presence/absence values at standardised breakpoints.⁹⁰

Fig. 7 An ASUDAS reference plaque.

Typically, variance in dental non-metric traits is quantified by calculating either the Mean Measure of Divergence ($\mathit{MMD}$) or the Mahalanobis Distance ($D$). The latter was chosen in this instance because it was originally developed for use with biological variables, has been shown to produce comparable results to $\mathit{MMD},$ is relatively well-catered for in most statistical packages, and, by taking into account the correlations between variables, prevents observations with high collinearity receiving undue weight.⁹¹ The formula for calculating the Mahalanobis Distance for $x$ th skeleton in the sample is given as: $$D=\sqrt{{\left(x-\mu \right)}^T{\mathrm{\Sigma }}^{-1}(x-\mu )},$$ where $x={(x_1,\mathrm{ }{x}_2,\mathrm{ }{x}_3,\dots x_N)}^T$ and $\mu ={({\mu }_1,\mathrm{ }{\mu }_2,{\mu }_3,\dots {\mu }_N)}^T$ – transposed vectors of equal length that represented respectively the dental trait observations for the $x$ th skeleton and the mean observations for the sample – while $\mathrm{\Sigma }$ is the covariance matrix. The computation thus reduced the multivariate data into a univariate statistic in N-dimensional Euclidean space $\left(N=20\right)$ which measured the distance between a vector of scores for each individual and the underlying sample mean whilst incorporating the correlations between different scores so that infrequent or unusual combinations of traits generated a greater $D$ statistic.⁹²

A requirement for Mahalanobis analyses is a complete dataset. In an archaeological context, incomplete data is near ubiquitous, so prior to the calculation of $D,$ missing trait observations were imputed.⁹³ When imputing data, it is suggested that multiple data points should be estimated initially through probabilistic calculations based upon frequencies of genuine observations; the newly generated datasets represent a distribution of possible values, which can then be aggregated.⁹⁴ In this study five imputations were performed using a regression method and the average used. After the computation of the $D$ statistic, distances between individuals were visualised through Multidimensional Scaling ($\mathit{MDS}$). Kruskal’s Stress test was employed to indicate how well scaled data was visually represented in the given number of dimensions and an R² value quantified the proportion of scaled variance accounted for; visualisations were considered valid when Kruskal’s Stress was <0.1 and R² was >0.6.⁹⁵ A p value was also calculated by comparing the $D$ statistic to a χ² distribution with N degrees of freedom to ascertain the probability of each distance representing a statistically meaningful outlier.⁹⁶

Broadly speaking, an outlier is an abnormal deviation from a measure of centrality, but their statistical identification can be confused by the clustering of multiple outliers and there is no consensus regarding significance thresholds, or the alpha value (α), at which a $D$ value is deemed a significant.⁹⁷ To select an appropriate significance threshold, a power analysis was conducted. Power analyses are generally employed to determine appropriate sample sizes to achieve acceptable statistical power where traditionally α=0.05.⁹⁸ In this instance, sample size was pre-defined by archaeological survivability, permitting an inversion of custom to produce an experiment-specific significance threshold and a balancing of the statistical parameters associated with sample size and α, namely power and effect size.⁹⁹ A relatively high statistical power (≥0.9) and a strong effect size (ɳρ²=≥0.8) were desired to ensure sensitivity to pronounced variations in a dataset which was expected to display a large degree of within-group variation.¹⁰⁰ Though significant results would not directly specify the nature of an alternate hypothesis,¹⁰¹ the association of dental non-metric traits with ancestral groups inferred significant deviations were the result of ancestral differences.¹⁰² Thus, a flexible framework was established which satisfied the ethical demand for parameters to be explicitly stated a priori in ancestral research.¹⁰³ All statistical procedures were conducted in SPSS 24.

OSTEOBIOGRAPHICAL NARRATIVES

The second phase of investigation involved the construction of osteobiographies for individuals identified as significant outliers. This qualitative approach, termed the bioarchaeology of the individual, theorises that the body is an experiential site which can reflect personal choices, social mores and culturally influenced behaviours,¹⁰⁴ such as the saliency of ancestral affiliations.

A standardised suite of methodologies employed to estimate sex, age and stature are briefly described in Table 2.¹⁰⁵ A body of evidence has formed, however, which demonstrates that these techniques can at times produce inaccurate estimates for individuals not of European descent – many of them having been developed using samples with a European or Euro-American bias. For some ancestral groups, corrections have been developed and these were applied as appropriate.¹⁰⁶ Ancestry was estimated through a multifactorial process,¹⁰⁷ using methods shown in Table 3. Based upon previous research, it was expected that a ‘simplified’ Western Eurasian dental complex, characterised by trait reduction from M1-M2, would be typical in 19th-century South Shields, as would craniometric analyses inferring a similarity to European populations;¹⁰⁸ deviation from this pattern would infer a substantive ancestral difference that would justify further exploration. Cranial non-metric traits were not assessed for ancestral affiliations, as it has been argued that the associated interpretative frameworks impose a racial categorisation.¹⁰⁹

Table 2. Features used for age, sex and stature estimation.

Estimate	Description
Sex	Sexually dimorphic skull and pelvic characteristics.
Age	Developmental changes and degenerative processes with a predictable progression were assessed, such as dental eruption and ectocranial vault closure.
Stature	Long bone measurements.

Table 3. Methods to estimate ancestry.

Methods	Description	Rationale
Observation of dental traits	Inspection of dental non-metric traits in a deductive manner to infer affinity to a dental complex.	Quick and efficient; broad estimate of biological affinity.
rASUDAS	ASUDAS dental non-metric data is quantitatively assessed through a Naïve Bayes classifier algorithm.	Precise estimate of bio-geographic affiliation. Extensive, global reference samples.
Craniometrics	29 cranial measurements are compared to Howell’s database using both CRANID and FORDISC programs.	Comparative estimate to evaluate non-metric results. Historic reference samples.

Palaeopathological features were recorded through standardised terminologies that specified the lesion’s anatomical location, appearance and size, so that the patterning of lesions could be comprehensively charted throughout the skeleton; the likely pathological processes that led to lesion development were also discussed with the intention of differentiating between acquired and congenital conditions.¹¹⁰ With this information, where possible, differential diagnoses were made when the distribution and inferred pathogenesis of lesions was suggestive of a specific disease, alternative aetiologies could be reasonably excluded, and contextual data – such as historic sources – were available to provide added supported.¹¹¹ Particular interest was paid to diseases that are rare to the UK that might further suggest a non-British origin or ancestry and also lesions that reflected developmental perturbations, such as hypoplastic defects, as childhood growth can be more sensitive to cultural pressures and produce changes that persist into adulthood.¹¹² It should, nonetheless, be noted that only stressors/diseases that disrupt physiological homeostasis for long enough to cause an alteration to bone morphology can be identified osteologically; especially virulent and acute diseases may cause death before any osseous signs develop, while low-level stressors may not leave discernible traces on the skeleton. As such, it is not always possible to fully appreciate an individual’s health during life through skeletal remains, and estimates of disease presence and prevalence are likely to be conservative in nature.¹¹³

Finally, the funeral rite was explored to discuss socio-cultural attitudes. Despite the fact that the focus of an osteobiography is on the lived experience, burial rites can reflect the attitudes of society in general.¹¹⁴ Through this multidimensional approach a narrative of an individual’s life course could be inferred, the role of ancestry in relation to other elements of personhood explored, and the effects of ancestral biases in methodologies discussed.

RESULTS

QUANTITATIVE RESULTS

In total, dental non-metric data could be gathered from 85 skeletons. Trait frequencies generally conformed to the trends expected in a Western Eurasian population, with low frequencies of traits such UI1 winging (9.1%) and shovelling (8.1%), as well as a pronounced reduction in molar trait expression from M1 to M2 (see Table 4). χ² tests did not detected significant associations between traits and sex. Using a random sample of five individuals and the calculation of an intraclass correlation coefficient (ICC) using a two-way mixed model, absolute observer agreement and the reliability of the data collection procedure was estimated to be ‘good’ (ICC=0.851).¹¹⁵

Table 4. Trait frequencies.

Trait	Tooth	Presence (%)	Trait	Tooth	Presence (%)
Winging	UI1	9.1	Hypoconulid	LM1	76.1
Shovelling	UI1	8.1	Hypoconulid	LM2	12.8
Hypocone	UM1	98.0	Y groove pattern	LM1	60.5
Hypocone	UM2	87.2	Y groove pattern	LM2	5.7
Carabelli’s trait	UM1	24.5	Cusp 6	LM1	13.0
Carabelli’s trait	UM2	0.0	Cusp 6	LM2	0.0
Cusp 5	UM1	14.0	Cusp 7	LM1	10.4
Cusp 5	UM2	14.0	Cusp 7	LM2	5.0
Parastyle	UM1	7.7	Protostylid	LM1	6.5
Parastyle	UM2	4.5	Protostylid	LM2	2.6

The visual representation of Mahalanobis distances did not lead to plots that were considered valid. The Kruskal stress score of 0.302 and R² value of 0.532 for the two-dimensional plot did not meet the minimum thresholds believed acceptable; although the three-dimensional plot was associated with a reasonable R² value of 0.70, a Kruskal stress score of 0.198 suggested that data was still not accurately displayed in three-dimensions.¹¹⁶ Greater success, however, was achieved in detecting statistical outliers through the use of a tailored p value. A power analysis indicated that strong statistical power (0.98) and effects size (ɳρ²=0.88) were accomplished with a rejection of the null hypothesis at p=≤0.01.¹¹⁷ With this criterion two significant outliers were detected: SK 358 ($D$=40.531, p =0.00428) and SK 649 ($D$=46.759, p =0.00063). These skeletons were selected for the construction of osteobiographies.

QUALITATIVE RESULTS

OSTEOBIOGRAPHY 1: SK 358

For SK 358, grave cut truncation limited demographic and palaeopathological inferences to the skull, thorax and upper limb. Cranial morphology suggested a male, while ectocranial vault closure and the calcification of thyroid cartilage inferred that SK 358 was above 45 years of age at their time of death.¹¹⁸ Trotter’s regression formula for a ‘White’ individual – often used in historic European contexts¹¹⁹ – generated a stature estimate of 168.7 cm ± 4.05 which falls close to the assemblage’s male average (male mean=171 cm; range=149-181 cm).¹²⁰ Regarding ancestral affiliation, SK 358’s dentition displayed a pattern of trait loss from M1-M2 typical of the Western Eurasian dental complex,¹²¹ and an rASUDAS assessment of dental non-metric traits assigned the greatest probability to Western Eurasian ancestry (52.0%).¹²² Damage to the cranium, which had led to both fragmentation and slight distortion of the cranial vault (see Figure 8), prevented a comparative craniometric analysis. In sum, it seemed likely that SK 358 was not ancestrally distinct and their identification as an outlier resulted from a combination of normal within-group variation and the imputation of several low frequency traits (n=3).

Fig. 8 Damage to SK 358’s cranial vault.

Perhaps the most striking features noted in SK 358’s skeleton were located in the dentition. There was, for instance, pronounced erosion of the antero-labial enamel surfaces in the mandibular teeth, seen in Figure 9. This may either have been the result of using the dentition as a “third hand” or an overbite; an overbite could have been consequence of either a genetic predisposition or a response to the softer diet of the post-medieval period – it has been theorised that soft diets do not provide sufficient chewing stress to stimulate normal mandibular growth.¹²³ Though dental defects and irregularities can be the cause of pejorative remarks in both past and modern populations,¹²⁴ the extent to which such a noticeable idiosyncrasy affected SK 358 can only be speculated. Additionally, a pipe smoking notch was observed between the left mandibular canine and first premolar. For a pipe to have noticeably abraded the teeth, it can be suggested that SK 358 was a habitual smoker who used a coarse clay pipe. As socially privileged pipe smokers in the 19th century tended to use pipes made with less abrasive, meershum stems,¹²⁵ it may be therefore be inferred that SK 358’s socioeconomic status was relatively low.

Fig. 9 Wear on the anterior dentition.

The supposition that SK 358 was of lower socioeconomic status is also consistent with skeletal signs of physical injury – deviated nasal bones and a healed fracture callus on a rib – that are frequently associated with employment in lower status and dangerous professions.¹²⁶ Evidence for occupational adaptations were not, however, apparent in markers of degenerative joint disease (DJD) or alterations at sites of muscle and ligament attachment – known as entheseal change (EC). Osteophytic formation around the glenoid fossae, for instance, and EC in the pectoral girdle was largely bilateral and likely age related.¹²⁷

OSTEOBIOGRAPHY 2: SK 649

Much like SK 358, the grave of SK 649 had been truncated and only a partial skeleton was present; this included the cranium and mandible, both scapulae, claviculae and humeri, cervical and thoracic vertebrae, as well as the manubrium, sternum, and ribs. In the absence of the pelvis, sex estimation was based on cranial and mandibular morphology. Although overall it appeared that SK 649 was female – the relatively small mastoid processes and slight glabella in the cranium were typically feminine, for example – certain mandibular traits were more masculine.¹²⁸ The gonial angle, for instance, was very steep, while the mandibular ramus was wide and robust (see Figure 10). When estimating age, further curious discrepancies were found. On the basis on dental development and epiphyseal fusion, it was believed that SK 649 was in their late teens or early twenties; the maxillary third molars were not quite in occlusion and the clavicular medial epiphyses were unfused.¹²⁹ However, the sutures of the ectocranial vault were generally closed, which suggested that SK 649 was at least in their late-twenties and possibly much older (see Table 5).¹³⁰ From the length of the left humerus (34.7 cm), stature was approximated to be 174.6 cm ± 4.45, which is significantly taller than the female mean (159 cm) and is the upper extreme of the assemblage’s female stature range (149-174 cm).¹³¹

Fig. 10 SK 649's skull with notable mandibular robusticity.

Table 5. Comparison of age estimates.

Method	Age range in years
	<19	20-24	25-29	30-34	35-39	40-44	>45
Dental eruption
Epiphyseal fusion
Ectocranial vault suture closure
Ectocranial lateral-anterior suture closure
Sternal rib ends

SK 649 possessed complex dental morphology, retaining cusps in the distal molars characteristic of Asian and African groups.¹³² Analysis of dental morphology via the rASUDAS program indicated that a Western Eurasian ancestry was unlikely (13.5%), and that Southeast Asian or Polynesian ancestry was most probable (33.9%) though the likelihood of Australo-Melanesian and Micronesian affinity was also high (32.1%).¹³³ Although results from the FORDISC analysis did not reach the typicality probability criteria for classification,¹³⁴ analyses conducted in CRANID were well catered for. LDA ascribed the greatest probability (46.9%) of affiliation to female populations from Hokkaido, Japan; the weighted sum was also highest for this population (692) and SK 649’s cranium was classified as female from Hokkaido through NNDA after all weighted sums were tallied. As the recommended cut-off points for reporting results are probabilities >0.1 for LDA and NNDA weighted sums >300, these results are presented in Tables 6 and 7.¹³⁵ SK 649’s crowded dentition is also more common in Asian populations, but could similarly have been linked to a soft diet providing insufficient stress to encourage normal mandibular development.¹³⁶

Table 6. Probabilities of population affinity based on LDA in CRANID.

Order	Population	Probability
1	North Japan, Hokkaido, Female	0.46934
2	Ainu, Hokkaido, Female	0.19152
3	South Japan, Kyushu, Female	0.11317

Table 7. Results of NNDA in CRANID.

Sample	Population	Hits	Weighted Sum
45	North Japan, Hokkaido, Female	7	692
41	Moriori, Chatham Island, Female	7	434
53	Ainu, Hokkaido, Female	4	333
19	Atayal, Taiwan, Male	3	327
46	South Japan, Kyushu, Female	4	309

These ancestry estimates initially appeared disparate and contradictory. However, the Ainu population of Hokkaido, being descended from a separate migratory event from the majority of the Japanese population, are dentally classified as Sundadonts along with Southeast Asians, Polynesians and other Pacific Asian groups (see Figure 11).¹³⁷ Sundadonts typically possess a ‘conservative’ dental morphology noted for features such as supernumerary cusp retention and third molar development. Sinodont groups, on the other hand, are noted for high frequencies of traits like incisor shovelling which were not observed in the dentition of SK 649.¹³⁸ The results of dental and cranial analyses are therefore broadly consistent and it is proposed that SK 649 was of Sundadont ancestry and that their presence in South Shields was possibly connected to British colonial and mercantile ventures in Southeast or Pacific Asia.¹³⁹ This estimate of ancestry also helped to clarify the idiosyncrasies noted in sex estimation; Sundadont groups can exhibit lower levels of sexual dimorphism in skull morphology, with mandibular robusticity noted in females.¹⁴⁰ Additionally, it has been found that traditional ectocranial suture closure methods overestimate age in Southeast Asian populations by 13-22 years; adjusting for this gave an age-at-death estimate of between 18-29 years that is congruent with dental development and epiphyseal fusion.¹⁴¹ Stature was also recalculated as 160.2 cm ± 2.6 with a regression formula designed for Southeast Asian females.¹⁴²

Fig. 11 Geographic dispersal of Sundadont and Sinadont populations. Map courtesy of Austin Whittal, 2014.

Several lesions suggested that SK 649 had been subjected to episodes of stress during development. For instance, slight porotic lesions were observed in SK 649’s orbits. These lesions, known as cribra orbitalia, most often develop during an individual’s early years and can be caused by a range of factors that includes nutritional insufficiency, infection and metabolic imbalances; generally, they are taken as evidence of non-specific stress experienced during infancy or childhood.¹⁴³ Bilateral linear enamel hypoplastic (LEH) defects on the mandibular canines, estimated to have formed between 3.0-3.9 years of age, further indicated that SK 649 was physiologically stressed during their formative years. LEH, which have the appearance of furrow-like depressions, form when normal enamel production is disrupted; although this can be the result of localised trauma, bilateral defects are more likely to be the consequence of a systemic perturbation.¹⁴⁴ Weaning and the cessation of maternal support has been implicated in the formation of cribra orbitalia and LEH in past populations,¹⁴⁵ however, the estimated age of LEH formation is slightly later than the typical weaning age for the period of c. 2 years.¹⁴⁶ Yet, as the lesions cannot be definitively linked to a specific cause, it is impossible to tell if this represents a weaning practice not normative for an industrial British population or an alternate explanation – such as interindividual variation in the chronology of dental development.

Entheseal and degenerative joint changes present in the left scapula – but absent in the right – appeared to be the result of physical activity. As Figure 12 shows, heterotopic ossification had occurred at the origin of the Teres major muscle on the lateral border of the scapula, whilst a depression in the glenoid fossa was differentially diagnosed as osteochondritis dissecans – a degenerative change in which a piece of cartilage and subchondral bone become detached from the joint surface.¹⁴⁷ Both of these lesions, though multifactorial, are often associated with intense or repetitive strain; the latter especially so in adolescents and younger adults.¹⁴⁸ In SK 649, this exertion appears to be focused on movements at the left glenohumeral joint, despite humeral and clavicular dimensions inferring right-handedness. As can be seen in Table 8, the right clavicle is more robust and slightly shorter, while the right humerus is longer overall and broader at sites of muscle attachment around the epicondyles; this suite of adaptations has been associated with the additional mechanical and compressive forces in the shoulder girdle and upper limb of the dominant side.¹⁴⁹ On this basis, it is speculated that SK 649 repeatedly undertook a prescriptive and possibly specialist task that dictated an imbalanced loading of the non-dominant hand resulting in trauma and subsequent osteological modifications.

Fig. 12 Ossification at the origin of Teres major on the lateral border of the left scapula.

Table 8. Clavicular and humeral measurements in millimetres.

Element	Measurement	Left	Right
Humerus	Maximum length	347.0	350.0
Humerus	Vertical diameter of head	47.2	47.2
Humerus	Epicondylar length	66.2	68.2
Clavicle	Maximum length	161.0	153.0
Clavicle	Horizontal midshaft diameter	15.7	15.8

Longitudinal cortical defects with raised margins, located bilaterally at Pectoralis major’s humeral insertions and unilaterally at the clavicular origin of the right Deltoid muscle, could have been interpreted as further evidence of activity-related trauma.¹⁵⁰ However, these lesions in younger individuals can be ascribed to an imbalance in osteoclastic and osteoblastic activity associated with the migration of fibrous muscle attachments during diaphyseal growth.¹⁵¹ The young age estimate, established through consideration of ancestral biases in aging methods, supported the latter interpretation.

SK 649's burial in an area of the cemetery thought to be consecrated and in use until c. 1816 could suggest that regardless of ancestral differences SK 649 was treated as an integrated member of the community. However, the clergy were known to exercise discretion concerning grave location; for, example, in spite of laws forbidding it, even confirmed suicides could be interred within consecrated grounds.¹⁵² Additionally, by the 19th century, St Hilda’s cemetery was beginning to experience severe overcrowding,¹⁵³ so the positioning of burials may reflect pragmatic rather than symbolic choices.

DISCUSSION

IDENTIFICATION OF OUTLIERS

As the consistency of dental morphological data collection both within and between recorders has been questioned and studies of ancestry are ethically sensitive,¹⁵⁴ it was important to consider the reliability of the assessed non-metric data. This was achieved by comparing trait frequencies to those found by other researchers and quantifying intra-observer error. After contrasting the trait frequencies found by this project with reference data for a Western Eurasian population,¹⁵⁵ a few notable differences were apparent and are presented in Table 9. These could either represent normal variation between populations, or inter-observer disagreement. Though inter-observer error can be pronounced in the recording of dental non-metric traits and make inter-project comparisons inadvisable, it does not necessarily affect the validity of the individual projects.¹⁵⁶ As intra-observer testing in this instance produced an intraclass correlation coefficient regarded as ‘good’ (ICC=0.851), data collection procedures were considered consistent and recorder error appears unlikely to have compromised results.¹⁵⁷ It should be noted though that trait selection will have biased the project’s outcomes. The selection of a large number of supernumerary molar cusps favoured the detection of individuals with a Sundadont or Sub-Saharan African dental complex, both of which are noted for supernumerary cusp retention.¹⁵⁸ Potentially, outliers from different ancestral groups went undetected.

Table 9. Comparison of anomalous trait frequencies in the Coronation Street sample to the Western Eurasian mean.

Trait	Coronation Street (%)	Western Eurasian mean (%)
Hypoconulid (LM1)	76.1	92.2
Y Groove pattern (LM2)	5.7	27.2
Cusp 7 (LM1)	10.4	4.5

A novel feature of this study was a customised significance threshold. Although this revealed two significant outliers (SK 358 and SK 649), SK 358’s estimated Western Eurasian ancestry was considered typical for 19th-century South Shields.¹⁵⁹ A simple solution for this instance of Type 1 error could be to reduce this project’s high statistical power (0.986), leading to a more conservative alpha threshold and a lower likelihood of false positives.¹⁶⁰ However, decreasing power may lead to an unacceptably high chance of Type 2 error in an archaeological context.¹⁶¹ At Hazel Grove’s Wesleyan chapel, for example, a group of skeletons thought to have African affinities were recovered; a grouping like this would cause statistical clustering, rendering such individuals imperceptible with a conservative alpha.¹⁶² Furthermore, it is questionable whether individuals of mixed ancestry, for whom non-metric traits are generally only expressed to an intermediary extent,¹⁶³ would be sufficiently different dentally to register as outliers – even though they may have been perceived or self-identified as being distinctive in life.¹⁶⁴ As the biographies of Olaudah Equiano and Sake Deen Mahomed indicate, both men had children with British women; an inability to accommodate admixture is a real limitation of this statistical framework.¹⁶⁵

Overall, it is estimated that 1.2% of the sample are ancestrally distinct (n=1/85), which is congruous with documentary sources from Liverpool approximating that 1.5-2.0% of the population was not British.¹⁶⁶ However, these figures are likely underestimates; for many migrants Britain was a waystation on a journey and cross-sectional cemetery samples and census data cannot fully represent such a dynamic situation.¹⁶⁷ Despite these limitations, a level of success was achieved as an ancestral outlier (SK 649) was identified through an objective procedure which was validated through a multifactorial estimation of ancestry. The further applicability of this method would likely be limited to historic niches where the majority of the population was phenotypically homogenous, outliers were few, and there was little admixture. The identification of individuals whose phenotypic traits lie beyond what could be considered normal within-group variation as ‘outliers’, however, arguably imposes a statistical typology and implies exaggerated differences between groups.¹⁶⁸ Thus, before any attempt is made to further apply this or a similar method, it would be essential to consider the ethical implications.

ANCESTRY ESTIMATION IN THE POST-MEDIEVAL PERIOD

This project highlights the practical benefits of estimating ancestry in several ways. As the case of SK 649 illustrates, estimating ancestry – as opposed to assuming a European affiliation in a historic British context – led to a more precise osteological profile. The more specific age estimate assisted in the interpretation of cortical defects, which could either have been viewed as normal developmental changes or activity-related adaptation dependent upon age.¹⁶⁹ Furthermore, for SK 358 and SK 649, ancestry estimates were made with reference to dental complexes. To achieve this a suite of biological features were assessed, preventing an oversimplified reliance on a small number of traits.¹⁷⁰ These estimates of biological affiliation represented an improvement on the traditional three-group model of African, Asian and European ancestry which has been termed a ‘folk taxonomy’ due to its close correspondence to social categorisations.¹⁷¹

It should be noted, however, that in the case of SK 358, a stature estimate was made using Trotter’s regression formulae which prompted a categorisation of SK 358 as ‘White.’¹⁷² Though this is common practice,¹⁷³ the use of a categorisation imbued with racial overtones did undermine the attempt to produce a more ethically considerate biological profile.¹⁷⁴ The issue was less pronounced in the analysis of SK 649’s remains, as in Southeast Asia attempts have been made to counter-balance the Euro-American focus of many sex, age and stature estimation methodologies.¹⁷⁵ Yet, even here, the final stature estimate was generated for someone of Southeast Asian rather than Sundadont ancestry;¹⁷⁶ though the two designations closely correspond, they are not synonymous.

Perhaps more important is the lack of specificity in estimates based on dental morphology. 19th-century European migrants established communities, businesses and eventually churches in South Shields,¹⁷⁷ thereby actively contributing to social diversity despite being indistinguishable through dental traits. Additionally, upon migration to Britain, German migrants underwent a transformative process; on the Continent they would have belonged to a politically distinct Germanic state, but in Britain they were ‘Germans’ collectively identified through language.¹⁷⁸ It is doubtful that SK 358 and SK 649 ever self-identified or were socially recognised as being of ‘Western Eurasian’ or ‘Sundadont’ ancestry; it is more likely that they distinguished themselves and constructed a sense of self-identity through language and culture. Although an expansion of craniometric databases could offset this lack of synergism – cranial morphology represents a relatively specific bio-geographic and temporal niche¹⁷⁹ – it has been hypothesised that over the last three centuries humans have been subject to a ‘technophysio evolution’.¹⁸⁰ Potentially, human physiology is now less subject to genetic constraints and biological changes are increasingly dictated by the unprecedented control humans exert on their environment.¹⁸¹ Secular changes in European skull morphology support this theory,¹⁸² and potentially SK 358’s possible overbite and SK 649’s mandibular dental crowding may both have resulted from the increasingly soft 19^th-century diet.¹⁸³ As such, post-medieval skull morphology would appear to increasingly reflect human influences rather than biological affinities.¹⁸⁴

In summary, it appears as though teeth, due to their strong genetic control and resistance to external leverages, still offer significant advantages when exploring biological affiliation.¹⁸⁵ However, perceived ancestry was likely influenced by socio-political context and language. As such fluid, contextual aspects of identity would be unidentifiable through unresponsive dental traits, the anthropological value of estimating ancestry is most evident when it is combined with historical and sociological data.¹⁸⁶

ANCESTRY AS AN ELEMENT OF PERSONHOOD

It still remains to be discussed to what extent ancestry may have shaped the life courses of SK 358 and SK 649 and contributed to the formation of the osteological features noted in their remains. Whilst doing this, ancestry will be contrasted with other facets of identity – such as sex, social status, and age – and, with refence to documentary data, the importance of these aspects of personhood will be contextualised in a 19th-century setting.

The traces of traumatic events consistent with intense or prolonged strain were seen in remains of both SK 358 and SK 649,¹⁸⁷ and it was speculated that these were related to employment. As many 19th-century migrants were employed in onerous and dangerous trades, for SK 649 these osteological markers may have been linked to ancestry.¹⁸⁸ However, other aspects of personhood would also have influenced occupation. For instance, social conventions and a lack of educational opportunities in poorer communities forced many to choose a profession that was deemed appropriate for their class – perhaps following peers and close family members.¹⁸⁹ Employment options for most women would have been even more restricted; women were perceived by many to be inherently less able and it was claimed that a man was capable of achieving ‘higher eminence, in whatever he takes up, than woman can attain’.¹⁹⁰ After marriage, in particular, many women would have been expected to remain in the domestic sphere, their personal identity potentially overshadowed by their role as a wife or mother.¹⁹¹ As it was suggested that SK 358 was of a lower socioeconomic status and SK 649 was female, the development of traumatic lesions associated with activity and occupation may therefore have been influenced by status and sex. In contrast, the hypoplastic defects and cribra orbitalia observed in SK 649 are strong evidence of infant or childhood adversity and suggest that age also played an influential role in a person’s experiences.¹⁹² Again, though age may be relevant in part due to its relation to occupation and work capacity; the diet of dependent offspring in 19th-century British families was generally of very low quality in poorer urban communities, but did improve when the child began to work and contribute to the household income.¹⁹³

It is worth considering that age may not have only affected an individual’s life experiences, but also the extent to which those experiences can be embodied. Adult skeletons are far less responsive than the bones of immature individuals,¹⁹⁴ which could mean that more subtle social stressors are less likely to be reflected in the adult skeleton. Similarly, although there is no direct skeletal evidence that unambiguously suggests that SK 649’s status as an ancestral outlier shaped their life, it does not mean that this is the case. Historical texts make plain that the opposite was likely true; in the stratified 19th-century society there was a documented hostility directed towards individuals whose ancestral origins led them to be perceived as the ‘other’.¹⁹⁵ Sources do however show that attacks often collocated characterisations of sex, social class and ancestry; certain ancestral groups were regarded as ‘effeminate,’ while ‘sickly physiques’ became associated with lower-class criminals and non-Europeans.¹⁹⁶ Such language lacks the precisely defined and institutionalised racism seen in other contexts, but arguably demonstrates the pervasive nature of biological determinism – the scientific paradigm that propagated qualitative categorisations regarding ancestry and race as well as sex and social class.¹⁹⁷ The impact such attitudes had are to an extent described by historic sources – Olaudah Equiano’s poignant personal accounts detail the deep psychological wounds suffered that could cause a ‘collapse’ of self-identity¹⁹⁸ – and modern research which has shown that marginalised and ancestrally distinct subgroups are the victims of entrenched health inequalities and barriers to attainment.¹⁹⁹

As such experiences are impossible to generalise, however, it cannot be said to what extent SK 649 was affected by racism. The discussion does highlight though that the study of ancestral diversity in the past cannot be broached without reference to racism and structural inequalities perpetuated by scientific trends. Yet, it is also important to acknowledge that deterministic theories did not always reflect everyday practice. One 19th-century writer travelling in the Tyneside region commented on ‘the singular practice of engaging women as labourers’ which, though he found it to be ‘unfavourable’ and ‘disgraceful’, was an economic necessity in the rapidly expanding industrial region.²⁰⁰ As such, it may be implied that economic pressures often defined the nature of an individual’s experiences and the importance of a particular aspect of identity; the race-riots in South Shields, for instance, occurred decades after the development of migrant communities, but were coincident with a sharp economic downturn which led to increasingly intense work-place competition.²⁰¹

On reflection, confidently drawing out interpretations pertaining to the subtleties of personhood is complex and best achieved by integrating multiple evidentiary strands. This assessment of osteological and historical data exemplifies how social identities are multifaceted, relationally bound and contextually dependent; even though aspects of personhood are theoretically partible, considering them separately and without reference to cultural environment is unrepresentative of social reality.²⁰² An individual’s experiences would have been shaped by, inter alia, their sex, age, social status and occupation as well as ancestry – or the racialised perceptions of ancestral differences for those whose biological affinities made them distinct.²⁰³ The facet of personhood which was most significant likely shifted over a life course and across social and historical contexts, making generalisations impossible. The interpretation of osteoarchaeological and historical data specific to a 19th-century British industrial context, suggests that the treatment of marginalised groups and individuals in this environment was influenced by a mixture of sometimes competing pressures rooted in deterministic prejudices and socioeconomic necessities.

CONCLUSION

The purpose of this project was to identify and contextualise ancestral diversity whilst critically evaluating associated methodologies. It appeared that the demographic impact of ancestral outliers was relatively small in early to mid-19th century South Shields (1.2%), and the presence of people from non-European backgrounds was likely the result of globally expanding commercial networks.²⁰⁴ It was further demonstrated that estimating ancestry can improve the accuracy of an individual’s biological profile and assist in the interpretation of palaeopathological lesions.²⁰⁵ Another salient feature discussed was the malleability of cranial morphology; recent research suggests that osseous adaptations in response to environmental pressures may be more pronounced that previously recognised.²⁰⁶ Dental morphology was deemed the most effective means of assessing biological affinities due to its resistance to environmental pressures, but with current techniques only broad ancestral affiliations could be estimated – the continued development of quantitative methods does, however, suggest that greater accuracy and precision will be available in the future.²⁰⁷ It is paramount though that progress is made ethically and does not propagate categorical paradigms or reduce human biological and cultural variation into statistical typologies.²⁰⁸

The impact ancestral diversity had upon an individual’s life course was much harder to define, partly reflecting the difficulty in discerning cultural nuance from a limited range of osteological reactions.²⁰⁹ This limitation, however, was mitigated through the inclusion of historical data. By integrating multiple strands of evidence it was possible to explore deterministic perceptions of personhood, investigate reactions to diversity in the Tyneside region, and argue that ancestral affiliation is best considered as one facet of identity, most appropriately understood and contextualised through its intersection with sex, social status, occupation and language.²¹⁰ As such, it is proposed that bioarchaeology, which is by its nature interdisciplinary, is in an ideal position to contribute to discussions relating to diversity and the persistent legacy of structural inequalities, especially in historic populations where textual data can inform and supplement research conducted on skeletal remains.²¹¹ Regarding future research, a greater appreciation of diversity in historic Britain could be facilitated by bioarchaeological investigations that compare assemblages recovered from port cities – isotopic analyses that would enable an exploration of geographic origin as well as population mobility may be especially informative.

ABBREVIATIONS

ASUDAS	=	Arizona State University Anthropology System
DJD	=	degenerative joint disease
EC	=	entheseal change
I	=	incisor
LEH	=	linear enamel hypoplasia
L	=	lower (e.g. LM1= lower first molar)
LDA	=	Linear Discriminant Analysis
M	=	permanent molar
NNDA	=	Nearest Neighbour Discriminant Analysis
SK	=	skeleton
U	=	upper (e.g. UM1= upper first molar)

SUMMARY IN FRENCH, GERMAN, ITALIAN AND SPANISH

RESUME

Une évaluation de la diversité génétique dans les South Shields du XIXe siècle

L’analyse bioarchéologique de restes humains du cimetière de l’église de Sainte- Hilda, South Shields, a été menée afin d’investiguer la diversité génétique dans la Grande-Bretagne postmédiévale. Des évaluations méthodologiques montrent que, lorsque la diversité est étudiée à travers des traits phénotypiques, les affiliations biologiques sont mieux investiguées à travers la morphologie dentaire grâce à sa résistance aux influences exogènes, qu’à travers la forme crânienne. En effet, cette dernière semble être progressivement influencée par des facteurs culturels et environnementaux dans la période postmédiévale. Un modèle quantitatif novateur a été développé dans le cadre d’une analyse au sein du groupe de morphologie dentaire, évaluée par le Système anthropologique dentaire de l’Université de Arizona State. Ce modèle a permis d’identifier avec succès un outlier génétique. Un récit ostéobiographique pour cet individu, combinant ostéologie et données historiques avec une théorie sociale, suggère que le statut de personne est malléable et multifacette. L’ascendance est ainsi mieux appréhendée comme aspect de l’identité, identité qui inclut également le sexe, l’occupation, le statut social et l’affiliation linguistique.

ZUSAMMENFASSUNG

Eine Untersuchung zur „Ancestral Diversity“ in South Shields, U.K., während des 19. Jahrhunderts

Der Artikel behandelt die bioarchäologische Untersuchung nachmittelalterlicher, menschlicher Überreste vom Friedhof der St. Hilda’s Church in South Shields, U.K. In bisherigen methodischen Untersuchungen wird oftmals argumentiert, dass bei der Erforschung der Diversität durch phänotypische Merkmale biologische Zugehörigkeiten aufgrund ihrer Resistenz gegenüber exogenen Einflüssen am besten durch die Zahnmorphologie untersucht werden, während die Schädelform in der nachmittelalterlichen Zeit zunehmend von Umwelt- und Kulturfaktoren beeinflusst zu werden scheint. Für eine Analyse der Zahnmorphologie, die mit dem Dental Anthropology System der Arizona State University durchgeführt wurde, wurde ein neuartiges quantitatives Modell entwickelt, mit dem ein Individuum herausgestellt werden konnte, dass eine andere Herkunft aufweist. Eine osteobiographische Erzählung für dieses Individuum, die osteologische und historische Daten mit Sozialtheorie kombiniert, legt nahe, dass Abstammung formbar und vielfältig ist. Abstammung wird daher am besten als ein Aspekt der Identität verstanden, der auch Geschlecht, Beruf, soziale Stellung und sprachliche Zugehörigkeit umfasst.

RIASSUNTO

Una valutazione della diversità ancestrale nella South Shields del XIX secolo

L’analisi bioarcheologica dei resti umani provenienti dal cimitero di St Hilda’s Church a South Shields, fu condotta con lo scopo di analizzare la diversità ancestrale nella Gran Bretagna post-medievale. Le considerazioni metodologiche suggeriscono che quando la diversità è indagata attraverso i tratti fenotipici, la discendenza biologica si accerta meglio attraverso la morfologia dentale in ragione della sua resistenza alle influenze esogene; la forma del cranio, invece, sembra essere progressivamente influenzata da fattori ambientali e culturali durante il periodo post-medievale. Un nuovo modello quantitativo è stato sviluppato per l’analisi della morfologia dentale all’interno di un gruppo: tramite il Dental Anthropology System della Arizona State University è stato identificato con successo un valore anomalo nel tracciamento ancestrale. Una narrazione basata sulla storia osteologica di questo individuo, che unisca dati osteologici e storici con elementi sociologici, suggerisce che l’individualità è plasmabile e proteiforme. La discendenza è quindi più opportunamente comprensibile come un aspetto dell’identità, che comprende anche il sesso, l’occupazione, lo status sociale e l’identità linguistica.

RESUMEN

Evaluación de la diversidad ancestral en el South Shields del siglo XIX

Se ha realizado un análisis bioarqueológico de los restos humanos del cementerio de la iglesia de Santa Hilda, South Shields, para investigar la diversidad ancestral en la Gran Bretaña temprano-moderna. Las evaluaciones metodológicas argumentan que cuando la diversidad se explora a través de rasgos fenotípicos, las afiliaciones biológicas se investigan mejor a través de la morfología dental por su resistencia a influencias exógenas, mientras que la forma craneal parece estar cada vez más influenciada por factores medioambientales y culturales en el período temprano-moderno. Se ha creado un modelo cuantitativo nuevo para analizar la morfología dental intergrupal, evaluada a través del Sistema de Antropología Dental de la Universidad Estatal de Arizona, gracias al cual se ha identificado con éxito un valor ancestral atípico. La narrativa osteobiográfica de este individuo, combinando datos históricos y osteológicos con la teoría social, sugiere que la personalidad es maleable y multifacética, por lo que es más apropiado entender la ascendencia como un aspecto de la identidad que también incluye sexo, ocupación, estatus social y afiliación lingüística.

The skeletal collection is curated at the Archaeology Department, The University of Sheffield. Department of Archaeology, Minalloy House, Regent Street, Sheffield, S10 2TN. [[email protected]]

ACKNOWLEDGMENTS

This paper is based upon research undertaken in the summer of 2018 as part of a Master’s dissertation project at the Department of Archaeology, University of Sheffield. I would like to acknowledge and thank Dr E. Craig-Atkins for her continued help and guidance throughout the research and writing process. Also, many thanks to Dr S. Newman for advice, organising access to materials and information on unpublished reports. I am further grateful for the constructive feedback offered by two anonymous reviewers and the help of the journal’s editorial team.

Notes

1 Fuller 2016; Qureshi 2011; Panayi 1995; Visram 1986.

2 Mahomet 1794; Equiano 1789.

3 Newman & Holst 2016; Myers 1996; Visram 1986.

4 Spradley & Hefner 2016; Edgar 2009; Smay and Armelagos 2000.

5 Jellinghaus et al. 2018; Fogel 2004.

6 Blakey 2001.

7 McLean 2020; Burchard et al. 2003, 1170-1; Mukhopadhyay & Henze 2003, 670.

8 Kranioti et al. 2018, 265.

9 Cunningham 2019; Zack 2002, 69; Appiah 1996, 73; Konigsberg 1990; Lewontin 1972.

10 Cunningham 2019; Relethford 2004, 499.

11 Spradley & Hefner 2016; Edgar 2009; Ousley et al. 2009; Cavalli-Sforza et al. 1994.

12 McLean 2020; McLean 2019, 481; Msimang 2019, 4; Hochman 2017, 62; Ousley et al. 2009, 68; Mallon 2007; Appiah 2006, 381; Mukhopadhyay & Henze 2003, 670; Zack 2002, 87-8; Appiah 1996.

13 Müller-Wille 2014; Gould 1981, 66.

14 Watkins 2007, 187-188; Gould 1981; Broca 1877, 191.

15 Gould 1981, 417; Darwin 1871, 201.

16 Msimang 2019, 8; Erasmus 2010, 388; Watkins 2007, 187-188; Gould 1981, 418; Galtung 1969; Darwin 1871.

17 Margaryan et al. 2020, 392; Ritchie 1976, 190.

18 Willmott et al. 2020, 189; Stone and Ozga 2019, 183-5; Jellinghaus et al. 2018; Cross & Wright 2015, 58; Chatters 2017, 24.

19 Jellinghaus et al. 2018; Spradley & Hefner 2016.

20 White & Folkens 2005, 404.

21 Buikstra & Ubelaker 1994, 69.

22 Hefner 2009; Edgar 2009.

23 Watkins 2007, 187-188; Mukhopadhyay & Henze 2003, 670; Blakey 2001, 411; Gould 1981.

24 Howell 1973.

25 Cunningham 2019; Kallenberger & Pilbrow 2012; Wright 2012; Buikstra & Ubelaker 1994, 69.

26 Jellinghaus et al. 2018; Sierp & Henneberg 2015.

27 Scott & Turner 1997; Hillson 1996.

28 Scott et al. 2018; Scott & Turner 1997; Turner et al. 1991.

29 Hasegawa et al. 2010, 7; Arte et al. 2001, 1448.

30 Scott et al. 2018; Scott & Turner 1997, 238-42.

31 Hillson 1996.

32 Scott et al. 2018; Scott & Turner 1997; Alt & Vach 1995.

33 Scott et al. 2018; Chatters 2017; Cross & Wright 2015.

34 Ruengdita et al. 2018, 78; Ubelaker & DeGaglia 2017; Kondo et al. 2017, 86; Traithepchanapai et al. 2016, 167e4; Gocha 2015, 318e3; Green & Curnoe 2009, 529.

35 McLean 2019, 481; Blakey 2001, 397.

36 Thoreson et al. 2015; Walker & Segal 1997.

37 Mallon 2007, 103; Appiah 2006, 365; Gelman 2003, 105; Tyrrell 1999.

38 Goodman & Leatherman 1998.

39 Pierre 2020, 127; Blakey 2001, 409.

40 Blakey 2001.

41 McLean 2020; Edgar 2009; Smay & Armelagos 2000.

42 Tyrrell 1999; Friedlander 1992; Ravenstein 1885, 199.

43 Myers 1996, 18-37.

44 Myers 1996, 19.

45 Jones-Katz 2017, 214; Paul 2009; Visram 1986; Mahomet 1794; Equiano 1789.

46 Ravenstein 1885, 196-9.

47 Redford & Chaloner 1976, 132-7.

48 Panayi 1995, 35.

49 Fuller 2016; Myers 1996, 110; Visram 1986; Doyle 1978, 64; Shyllon 1977, 122; George 1926, 122; Hughes 1855, 5.

50 Myers 1996, 110; Visram 1986; Shyllon 1977, 122; George 1926, 122; Hughes 1855, 5.

51 Otele 2012, 155-6; Richardson 2005, 41; Dresser 2001, 9.

52 Otele 2012; Richardson 2005; Dresser 2001, 9.

53 Dresser 2001, 81; Visram 1986; Hecht 1954.

54 Visram 1986; Hecht 1954.

55 Bashford & McAdam 2014, 311.

56 Newman 2016; Myers 1996; Panayi 1995.

57 Redford & Chaloner 1976, 133.

58 Jackman 2018, 141-3; Dresser 2001; Visram 1986, 17.

59 Jackman 2018, 141-2; Dresser 2001, 62; Visram 1986, 17; Hecht 1954, 52.

60 Dresser 2001; Visram 1986, 30-2.

61 Qureshi 2011.

62 Teltscher 2000; Mahomet 1794.

63 Equiano 1789.

64 Jones-Katz 2017; Myers 1996.

65 Qureshi 2011; Myers 1996; Visram 1986.

66 Hecht 1954, 52; Mahomet 1794; Equiano 1789.

67 Blakey 2001; Rankin-Hill 1997; Owsley et al. 1987.

68 Newman 2016, 8-10; Eastlake 2013, 6; Bland 2006, 68.

69 Nolan 1998, 33-6.

70 Myers 1996, 22.

71 Bland 2006.

72 Byrne 1977, 268.

73 Raynor et al. 2011, 6.

74 Raynor et al. 2011; Brickley 2018, 7.

75 Raynor et al. 2011, 53.

76 Raynor et al. 2011, 44-5; Buckberry & Chamberlain 2002; Molleson et al. 1993 Mackenzie 1827, 730-5.

77 Raynor et al. 2011, 47.

78 Breeze 2006, 115; Surtees 1820, 94-104.

79 Newman 2016, 69.

80 Raynor et al. 2011.

81 Myers 1996.

82 Raynor et al. 2011, 11; Mackenzie 1834, 31.

83 Raynor et al. 2011, 15; Salmon 1856, 64.

84 Hasegawa et al. 2010; Arte et al. 2001; Scott & Turner 1997; Hillson 1996.

85 Scott & Turner 1997; Turner 1990a.

86 Fuller 2016; Myers 1996; Visram 1986.

87 Scott & Turner 1997; Turner et al. 1991.

88 Scott & Turner 1997, 69-73.

89 Fuller 2016; Scott & Turner 1997, 96; Myers 1996; Turner et al. 1991; Visram 1986.

90 Scott & Turner 1997; Turner et al. 1991.

91 Ghorbani 2019, 588; Elfadaly et al. 2016, 116; Nikita 2015; Irish 2010; Field 2009, 215-7; Leese & Main 1994, 309-10; Mahalanobis 1936.

92 Ghorbani 2019, 588; Elfadaly et al. 2016, 116; Filzmoser et al. 2014; Field 2009, 215-7; Leese & Main 1994, 309-10; Mahalanobis 1936.

93 Hayati Rezvan et al. 2015.

94 Pedersen et al. 2017; Hayati Rezvan et al. 2015; Newman 2014.

95 Irish 2010; Irish 2006, 534; Kruskal & Wish 1978.

96 Ghorbani 2019, 590; Elfadaly et al. 2016, 116; Brereton 2015, 10; Field 2009, 51; Leese & Main 1994, 309.

97 Elfadaly et al. 2016; Field 2009: 215-7.

98 Smith 2018; Cohen 1988.

99 Cohen 1988, 14.

100 Field 2009, 55-8; Scott & Turner 1997, 161; Cohen 1988, 14.

101 Smith 2018.

102 Scott et al. 2018.

103 Blakey 2001; Gould 1981.

104 Roffey et al. 2018; Hamilakis et al. 2002.

105 Ubelaker 1989a; Meindl & Lovejoy 1985; Ferembach et al. 1980; Trotter 1970.

106 Ubelaker & DeGaglia 2017.

107 Scott et al. 2018; Scott & Turner 1997, 238-42; Wright 2012; Jantz & Ousley 2005; Howell 1973.

108 Jellinghaus et al. 2018; Scott & Turner 1997, 236; Bland 2006, 68.

109 Blakey 2001, 411.

110 Klaus 2017; Villotte et al. 2016; Buikstra & Ubelaker 1994, 107-159.

111 Mays 2020, 91-3; Klaus & Lynnerup 2019, 81; Klaus 2017; Villotte et al. 2016; Ortner 2003, 45-63; Buikstra & Ubelaker 1994, 107-159.

112 DeWitte & Stojanowski 2015, 416-8; Primeau et al. 2015.

113 DeWitte & Stojanowski 2015; Wood et al. 1992.

114 Nystrom 2014; Zvelebil & Weber 2013, 275; Parker Pearson 1999, 42-3.

115 Koo & Li 2016, 155.

116 Irish 2010; Irish 2006, 534; Kruskal & Wish 1978.

117 Field 2009, 55-8; Cohen 1988, 14.

118 Mupparapu & Vuppalapati, 2005; Loth & Henneberg 1996; Meindl & Lovejoy 1985.

119 Mays 2016; Trotter 1970.

120 Raynor et al. 2011, 47.

121 Scott & Turner 1997, 236.

122 Scott et al. 2018.

123 Pilloud 2018, 14; Mew 2014, 555; Ortner 2003 598.

124 Scheffel et al. 2014; Lombroso 1911, 21; Broca 1877, 195.

125 Eastlake 2013, 49; Henderson & Walker 2012, 796.

126 Van Staa et al. 2001, 517-22.

127 Niinimäki & Baiges Sotos 2013; Knusel 2000.

128 Loth and Henneberg 1996; Ferembach et al. 1980.

129 Scheuer & Black: 252; Ubelaker 1989a; Iscan et al. 1984.

130 Meindl & Lovejoy 1985.

131 Raynor et al. 2011, 49; Trotter 1970.

132 Scott & Turner 1997, 236-52

133 Scott et al. 2018, 23-5.

134 Jantz & Ousley 2005.

135 Wright 2012.

136 Pilloud 2018, 14; Mew 2014 555.

137 Howell 1996: 172-3.

138 Turner 1990a.

139 Fuller 2016; Scott & Turner 1997; Visram 1986.

140 Kondo et al. 2017, 86; Green & Curnoe 2009, 529; Kean & Houghton 1990, 361.

141 Traithepchanapai et al. 2016, 167e4.

142 Ruengdita et al. 2018, 78; Gocha 2015, 318e3; Ubelaker 1989a; Meindl & Lovejoy 1985; Iscan et al. 1984.

143 Gamboa et al. 2011; Mahoney-Swales & Nystrom 2009, 34; Ortner 2003, 387-8.

144 Primeau et al. 2015.

145 Tsutaya 2017; Newman 2016; Pooley 2013; Beaumont et al. 2013; Nitsch et al. 2011.

146 Tsutaya 2017; Newman 2016; Pooley 2013; Beaumont et al. 2013; Nitsch et al. 2011.

147 Villotte & Knusel 2013, 138; Blondiaux et al. 2012, 226; Waldron 2009, 153-4.

148 Villotte & Knusel 2013, 138; Blondiaux et al. 2012, 226; Waldron 2009, 153-4; Rossi & Dragoni 2006.

149 Henderson et al. 2013b, 197; Niinimaki & Baiges Sotos 2013; Rhodes & Knusel 2005; Knusel 2000; Mays et al. 1999.

150 Villotte et al. 2016, 53; Gosling et al. 2002.

151 Villotte & Knusel 2013: 140-1.

152 Cox et al. 1990, 578.

153 Salmon 1856.

154 Smay & Armelagos 2000; Tyrrell 1999, 161.

155 Scott & Turner 1997, 323.

156 Tyrrell 1999, 161.

157 Koo and Li 2016, 155.

158 Scott et al. 2018; Scott & Turner 1997; Turner 1990a.

159 Scott et al. 2018; Bland 2006.

160 Cohen 1988, 14.

161 Field 2009, 55-8; Cohen 1988, 14.

162 Elfadaly et al. 2016; Newman & Holst 2016.

163 L’Abbé et al. 2011; Edgar 2009.

164 Aspinall 2018; Mahomet 1794; Equiano 1789.

165 Myers 1996, 22.

166 Field 2009, 784; Myers 1996; Visram 1986.

167 Turk 2005; Myers 1996; Panayi 1995; Visram 1986; Doyle 1978, 222.

168 McLean 2020, 43; Gannet 2001.

169 Villotte et al. 2016, 53.

170 L’Abbé et al. 2011.

171 Spradley & Hefner 2016; Edgar 2009; Mallon 2007, 103; Watkins 2007, 187-188; Appiah 2006, 367-70; Mukhopadhyay & Henze 2003, 670; Smay & Armelagos 2000.

172 Trotter 1970.

173 Mays 2016.

174 Smay & Armelagos 2000.

175 Ruengdita et al. 2018; Gocha 2015; Green & Curnoe 2009.

176 Traithepchanapai et al. 2016.

177 Panayi 1995, 163.

178 Turk 2005; Panayi 1995; Doyle 1978, 222.

179 Larsen 1997, 237-8; Buikstra & Ubelaker 1994, 69.

180 Fogel 2004.

181 Fogel 2004, 20-1.

182 Jellinghaus et al. 2018; Jantz & Jantz 2016.

183 Pilloud 2018.

184 Jellinghaus et al. 2018; Fogel 2004; Relethford 2004.

185 Hasegawa et al. 2010, 7; Arte et al. 2001, 1448; Hillson 1996, 181.

186 Blakey 2001.

187 Villotte et al. 2016; Waldron 2009.

188 Myers 1996; Visram 1986.

189 Karlsson & Stanfors 2011; Laband & Lentz 1983, 478.

190 Gould 1981, 418; Darwin 1871.

191 Henderson et al. 2013b, 205-6; Pooley 2013, 87; Raynor et al. 2011, 108; Hiley 1979; Mackenzie 1827, 731-2.

192 DeWitte & Stojanowski 2015, 423-6; Primeau et al. 2015; Ortner 2003, 387-8.

193 Tsutaya 2017; Newman 2016; Pooley 2013; Beaumont et al. 2013; Nitsch et al. 2011.

194 DeWitte & Stojanowski 2015; Wood et al. 1992.

195 Erasmus 2010, 388; Paul 2009; Gould 1981; Galtung 1969.

196 Gould 1981; Lombroso 1911.

197 McLean 2020, 43; Aspinall 2018; Blakey 2001; Owsley et al. 1987; Galtung 1969.

198 Paul 2009, 855; Fanon 1986, 154; Equiano 1789, 137.

199 McLean 2020; Mendoza et al. 2017; Fullwiley 2015, 42-3; Galtung 1969.

200 Raynor et al. 2011, 108; Hiley 1979; Mackenzie 1827, 731-2.

201 Bland 2006; Byrne 1977, 268.

202 Appuhamilage 2017; Appiah 2006, 366; Parker Pearson 1999, 72-6.

203 Msimang 2019; Hochman 2017; Mallon 2007, 94; Appiah 2006, 367.

204 Fuller 2016; Doyle 1978.

205 Ruengdita et al. 2018; Villotte et al. 2016; Mays 2016; Gocha 2015.

206 Jellinghaus et al. 2018; Jantz & Jantz 2016; Fogel 2004; Larsen 1997, 237-8.

207 Scott et al. 2018; Hasegawa et al. 2010; Scott & Turner 1997.

208 McLean 2020, 43; Spradley & Hefner 2016; Edgar 2009; Mukhopadhyay & Henze 2003, 670; Gannet 2001; Smay & Armelagos 2000.

209 Klaus 2017.

210 Pierre 2020, 127; Appuhamilage 2017; Parker Pearson 1999, 72-6; Gould 1981.

211 Fricke 2018; Blakey 2001; Rankin-Hill 1997; Owsley et al. 1987.