Content uploaded by L. Creighton Avery
Author content
All content in this area was uploaded by L. Creighton Avery on Apr 25, 2022
Content may be subject to copyright.
Bioarchaeology International
Volume 3, Number 3: 157–173
DOI: 10. 5744/bi.2019.1011
Copyright © 2020 University of Florida Press
Dental Health and Dietary Difference at Late
Roman Winchester
L. Creighton Avery,a* TracyL. Prowse,a and MeganB. Brickleya
aDepartment of Anthropology, McMaster University, Canada
*Correspondence to: L. Creighton Avery, McMaster University, Chester New Hall Room ,
Main St.West, Hamilton, Ontario LS L, Canada
e- mail: averylc@mcmaster . ca
ABSTRACT Gendered experiences are not static but are inuenced by other aspects of identity. is study investigates the
inuence of interrelated aspects of identity on diet through an analysis of dental health at Late Roman Win-
chester (fourth and h centuries A.D.), England. Specically, we investigate the inuence of age and social
status on gendered dietary experiences. Antemortem tooth loss (AMTL), carious lesions, and tooth wear were
recorded for adults (+ years) and compared between sex- , age- , and status- based groups, using burial type
as a proxy for status. Mann- Whitney U tests indicate that females had signicantly higher rates of total carious
lesions than males (p < .) in the combined sample, while old adult males exhibited higher rates of posterior
tooth wear than females, and middle adult males presented lower rates of AMTL than females. When burial
type is also incorporated, a more variable picture emerges. In monumental constructions the sexes exhibited no
dierences in dental health, whereas in earthen pits and cons both age and sex aected the prevalence of den-
tal disease.
ese results suggest that gendered dietary experiences at Late Roman Winchester were partly dependent on
social status and age and may be associated with dierences in occupations or daily activities. By incorporating
multiple aspects of identity, this study moves beyond broad- scale patterns of dental health, providing a more
nuanced understanding of gendered experiences in Roman Britain, and demonstrates how cross- cutting vari-
ables of identity can be used to better understand the past.
Keywords: age; gender; status
Las experiencias diferenciadas por género no son estáticas sino inuenciadas por otros aspectos de identidad.
Este estudio investiga la inuencia de los aspectos interrelacionados de la identidad sobre la dieta a través de un
análisis de la salud dental en la época romana tardía (- º siglo, dc), en Winchester, Inglaterra. En concreto, se
investiga la inuencia de la edad y el estatus social sobre las experiencias dietéticas según género. La pérdida de
dientes ante mortem (AMTL), las lesiones cariosas y el desgaste de los dientes fueron registrados en adultos
( años y más) y luego comparado entre el sexo, la edad y los grupos basados en estatus, utilizando el tipo de
entierro como representante del estatus. Las pruebas Mann- Whitney U indican que las mujeres tenían tasas sig-
nicativamente más altas de lesiones cariosas en total que los machos (p < .) en la muestra combinada, mien-
tras que los viejos adultos machos mostraron mayores tasas de desgaste de dientes posteriores que las hembras,
y los machos medios adultos presentaron menores tasas de AMTL que las hembras. Cuando el tipo de entierro
también se incor pora, una imagen más va riable aparece. En construcciones monumentales, los sexos no mostraron
diferencias en la salud dental, mientras que en los pozos de tierras y en los ataúdes, la edad y el sexo ambos afec-
taron la prevalencia de la enfermedad dental.
Estos resultados sugieren que las experiencias dietéticas de los géneros en Winchester en la época romana
tardía fueron parcialmente dependientes del estatus social y la edad, y puede estar asociadas con diferencias de
ocupaciones o actividades diarias. Mediante la incorporación de múltiples aspectos de identidad, este estudio
Received 12 February 2019
Revised 17 June 2019
Accepted 24 June 2019
Dental Health and Dietary Dierence at Late Roman Winchester
158
Much of the literature surviving from the Roman pe-
riod represents women as inferior to men (Garnsey
; Papinian, D... in Grubbs :xi). Yet it is un-
clear how appropriate and relevant ancient literary
sources are for life outside of Rome, and if they accu-
rately depict experiences in more distant territories of
the Empire, such as Roman Britain (Fuller etal. ;
Müldner ; Redfern etal. ). To move beyond
the broad assumptions about gender as portrayed in
ancient literary sources, bioarchaeologists can con-
sider embodied experiences of gender that become
visible in the human skeleton, including dierences in
dental health, to better understand gender dynamics
elsewhere in the Roman Empire.
Although literary sources present a dichotomous
relationship between men and women, experiences of
gender were further inuenced by social status and
age (Hemelrijk and Woolf ). us, investigating
gender- based dietary dierences in isolation from
other aspects of identity can lead to an oversimpli-
cation of the past (Harlow and Laurence ). To
move beyond single- scale analyses, researchers can
incorporate cross- cutting variables and consider how
multiple factors of identity inuence one another
(Joyce ). Researchers have studied dietary dif-
ferences as they relate to sex and settlement type
(e.g., Cheung etal. ), sex and status (e.g., Cucina
and Tiesler ), and even sex, status, and time
(e.g., White ).
As dental health is closely related to diet, research-
ers can learn about the type and consistency of foods
consumed through an analysis of carious lesions,
tooth wear, and antemortem tooth loss (AMTL).
Dental caries is the destructive process of dental tis-
sues, due to the acidic environment created by bacte-
ria, possibly leading to the formation of a carious
lesion (Hillson ). Although the formation of car-
ious lesions can be inuenced by a variety of factors,
diet is arguably the most important. Specically, di-
ets high in carbohydrates produce an acidic environ-
ment that promotes enamel demineralization and
may ultimately produce elevated rates of carious le-
sions (Hillson ). Conversely, diets high in pro-
tein produce an alkaline oral environment, leading
to increased rates of dental calculus, which may pro-
tect against the formation of carious lesions (Hillson
). According to Lukacs (, ), dental car-
ies can also be aected by biological processes, such
as pregnancy, where hormonal changes can alter
salivary ow rate and composition and suppress the
immune response system. Accordingly, when exam-
ining dierences in rates of carious lesions, bioarchae-
ologists must consider both social and biological
explanations (Lukacs ). Tooth wear encompasses
both attrition (tooth- on- tooth contact) and abrasion
(tooth- food interaction) and can provide insights
into the texture and consistency of foods consumed
(Hillson ). For example, a diet consisting mainly
of abrasive and hard foods will wear away dental
enamel at increased rates, while so and sticky foods
will result in minimal wear but may lead to more car-
ious lesions (Kaidonis ). AMTL is not specic to
one aspect of diet but is oen related to dental disease
(i.e., periodontal disease, gross carious lesions, and
severe wear) and/or activity and cultural factors (i.e.,
trauma, interpersonal violence, cultural dental ab-
lation). Consequently, an understanding of AMTL
rates is important for interpreting rates of other dental
health variables (Hillson ).
Rather than focusing on exactly what foods were
consumed in the past, researchers analyze dental
health to explore a wide range of questions regard-
ing social transformations and inequalities, such
as changes following the adoption of agriculture
(e.g., Lukacs ; Temple and Larsen ), the im-
pact of political or cultural transformations on diet
(e.g., Belcastro etal. ; Redfern etal. ), and
dierences in dental health in relation to age, sex,
gender, or socioeconomic status (e.g., Cucina and
Tiesler ; Keenleyside ). While some studies
uncover broad- scale patterns of variability or social
stratification between subsamples, others investi-
gate dierences among and between subsamples, in-
corporating cross- cutting variables of identity
(Joyce ).
is study explores dental health in relation to sex,
age, and burial type to better understand gender- based
dietary dierences at Late Roman Winchester, En-
gland (ca. A.D. – ). By analyzing dental health
in this manner, we elucidate the complex interrelation-
ship between dierent aspects of identity and gain a
more nuanced interpretation of gendered experiences
at Late Roman Winchester. erefore, this study has
two aims: () to determine if gender- based dietary dif-
ferences existed at Late Roman Winchester, and () to
determine if gender- based dierences were uniform at
Late Roman Winchester or if they were further inu-
enced by age or social status.
va más allá de la amplia escala de los patrones de salud dental, proporcionando una comprensión más matizada
de las experiencias según género en la Gran Bretaña Romana y muestra cómo variables transversales de identi-
dad pueden ser utilizados para mejor comprender el pasado.
Avery et al. 159
Materials and Methods
Roman Winchester
e town of Venta Belgarum (Roman Winchester) was
established ca. A.D. and was designated a civitas
capital around A.D. , functioning in part as an ad-
ministrative center and unit of government for the
surrounding conquered areas (Fig.; James ). Due
to its economic and administrative status as a civitas,
the population of Roman Winchester grew to three or
four thousand inhabitants, making Venta Belgarum
the h- largest town in Roman Britain (James ;
Stuckert and Kricun ). Archaeological excavations
have uncovered evidence for a wide range of occupa-
tions within the community, including metallurgy,
bone working, and weaving, indicating that most in-
habitants worked as traders or crasmen, with a
smaller ruling elite living in higher- status houses
within and around the community (Bonsall and Pick-
ard ; James ).
Cool () suggests that a key staple of the diet in
Roman Britain was grains, supplemented with sh,
meat, fruit, and vegetables, although the grain variety
and milling style was dictated by local preference
and/or status. Archaeobotanical research has demon-
strated that urban settlements in Roman Britain had
access to a greater variety of foods, while rural settle-
ments showed less diversity, particularly in relation to
foods that were introduced to Britannia during the
Roman period (van der Veen et al. ). Similarly,
sites associated with Roman military encampments
demonstrate great diversity in foods, likely due to the
trade networks established in association with the
Roman Empire.
Gendered dierences in diet are also referenced in
literary sources, with men having greater access to
meat and marine resources than women. However,
Romano- British studies investigating sex- based di-
etary dierences have produced conicting results
and do not always agree with these literary depictions.
For example, studying dierences in dental health,
Bonsall () and Grin etal. () conclude that
dierences in diet between males and females were
subtle and may be due to dierent patterns of con-
sumption or non- dietary factors rather than distinctly
gendered diets. At Poundbury Camp (England, fourth
century A.D.), anthropologists found clear sex- based
dierences in dietary stable isotope values, with males
consuming more protein or a more varied diet than
females (Richards etal. ), but these results are not
consistent throughout Roman Britain. For example,
studying dietary dierences and change in Roman
Gloucestershire, Cheung etal. () found no signi-
cant dierences in dietary isotope values between
males and females at two rural sites but did nd sig-
nicant dierences between sex- based groups at an
urban site. ey suggest that these patterns might be
the result of a greater military presence within the ur-
ban settlement and that they consequently reect di-
etary patterns of individuals who lived elsewhere in
the Roman Empire and moved to Roman Gloucester-
shire later in life (Cheung etal. ).
While some have suggested that the Romano-
British diet was largely vegetarian, zooarchaeological
analyses demonstrate that this was not the case at Late
Figure 1. Maps of (A) United Kingdom with Winchester and London identied, a nd (B) of south- central UK, with Roman cities (current name in
parentheses) and known Roman roads marked. Images created by LCA.
Dental Health and Dietary Dierence at Late Roman Winchester
160
Roman Winchester (Allason- Jones ; Booth etal.
). Rather, the quantity and variety of animal re-
mains at the site suggest that the inhabitants of Venta
Belgarum consumed higher- than- expected quantities
of meat (Booth etal. ). Stable isotope analysis con-
ducted on the Late Roman Winchester sample indi-
cates a diet based on C plants, with regular inclusion
of animal resources (Bonsall and Pickard ; Cum-
mings and Hedges ).
In this study, we examine skeletal samples from
four archaeological sites corresponding to the North-
ern Cemetery of Late Roman Winchester, all dating
between the early fourth and early h centuries
A.D. (Table). Only those with estimated sex and age
were included in this study, providing a total sample
size of individuals.
Age estimations
Age at death was estimated using tooth develop-
ment (Gustafson and Koch ), epiphyseal fusion
(Cardoso a, b), pubic symphysis morphol-
ogy (Brooks and Suchey ; Katz and Suchey ),
and transition analysis of the sacroiliac joint (Boldsen
etal. ). As this study examines sex- based dier-
ences, only those aged + years were included in the
nal study to ensure individuals were old enough to
show clear sex- related dierences in the skeleton.
Individuals were assigned to age categories of young
adult (YA, .– .years), middle adult (MA, .–
.years), and old adult (OA, .+ years) following
Buikstra and Ubelaker ().
Sex estimations and interpretation of gender
In biological anthropology, it is generally accepted that
sex and gender are not synonymous but refer to two
dierent constructs (see Sofaer and Geller
for discussion). Sex (male, female, etc.) refers to the bi-
ological construct and relates to dierences in genita-
lia and secondary sexual characteristics. In the current
study, sex- based terminology is used when discussing
osteological sex estimations and presenting dental
health data that were measured by sex- based groups.
Methodologically, sex was estimated using standard
pelvic and cranial morphology (Acsádi and Nemeskéri
; Phenice ).
Gender (man, woman, etc.) is a cultural construct
that relates to how individuals should or choose to be-
have, or are treated within society. To investigate gen-
der, biological anthropologists employ theoretical
frameworks, such as the theory of embodiment (see
Hollimon ). In this framework, gender is under-
stood to be performative, leaving marks on the skele-
ton through habitual actions (Hollimon ; Joyce
; Sofaer ). us, by considering sex- based dif-
ferences in dental health and contextualizing these
data with other lines of evidence, we can explore the
ways in which gender, or other aspects of identity, were
embodied, enacted, and experienced by people at Ro-
man Winchester. Consequently, this study uses lan-
guage associated with gender when discussing lived
experiences or contextualized data. Although non-
binary individuals certainly existed in Roman society,
the literary texts describe a binary distinction be-
tween men and women; as a result, this analysis uses
two gender- based groups (Grubbs ).
Burial construction
To consider social status dierences in the past, ar-
chaeologists and bioarchaeologists oen look at the
quantity and quality of grave goods within a burial
to estimate relative wealth and, by inference, status.
However, the use of grave goods as a simple transla-
tion of wealth can be highly problematic (see Binford
; Cannon etal. ; Pearce ). For example,
studies of Romano- British cemeteries demonstrate
that grave good distributions relate to the age and
gender of the individual, rather than being an inde-
pendent indicator of status or wealth (Cooke ;
Gowland ). At Late Roman Winchester in par-
ticular, the quantity and type of grave goods were
closely tied to social age and ethnic identity (as
interpreted from burial rites and stable isotope
analyses; see Clarke ; Eckardt et al. , ;
Evans et al. ; Gowland ). For this reason,
Gowland () cautions against oversimplifying
social identity in the mortuary record in Romano-
British contexts.
Tab le 1. Sites and Skeletal Samples from Late Roman Winchester.
Site Date Range Site Report Sample Size
Lankhills School Site A.D. – Clarke
Lankhills Expansion Early th– Early thcentury Booth eta l.
Andover Road Mid- th– Early thcentur y Ottaway etal.
Hyde Street A.D. – Early thcentury Ottaway etal.
Tot a l
Sample size represents t he number of individuals included in this study. See site reports for fu ll collection size.
Avery et al. 161
Alternatively, researchers have used burial con-
struction as a proxy for social status in Roman Brit-
ain, with the construction of an individual’s grave
providing an indication of the investment of resources
and labor by the family of the deceased (Redfern and
DeWitte ; Richards etal. ). Archaeologically,
burial constructions tend to leave clear evidence in the
mortuary record, whether in the form of wood, con
nails, con ttings, or soil stains (Booth etal. ).
However, Robb etal. () point out that, when ex-
ploring social status by grave construction, the high-
est levels of the social stratum may have been buried
elsewhere (e.g., a family mausolea on a villa estate out-
side of the community cemetery), while those of the
lowest social stratum may have been placed in shallow
unmarked graves and lost over time. Additionally,
subtle indicators of status in burial constructions may
be lost in the archaeological record. For example, ela-
borate carvings or embellishments on wooden cons
may have indicated status at the time of burial but not
leave any discernible archaeological traces (Pearce
; Redfern and DeWitte ). us, when consid-
ering burial constructions, we must acknowledge
that the highest and lowest strata of Roman society
may not be well represented in these cemeteries and
that subtle dierences between social groups may
not be distinguishable.
For the Late Roman Winchester cemeteries, ar-
chaeologists described the burial context for each in-
dividual, including burial construction (e.g., number
of con nails, con ttings, soil stains) (Booth etal.
; Clarke ; Ottaway etal. ). Based on these
descriptions, we identied three key types of burial
constructions, which are used as a proxy for social
status in this study: monumental constructions, wooden
cons, and pit burials (Fig.). Monumental construc-
tions include stepped burials and those with an enclo-
sure built around the interment, lead cons, and
stone sarcophagi. Wooden cons are those with evi-
dence of a wooden container at the time of burial,
which may include ttings, nails, or soil staining that
was interpreted by the archaeologists as the remnants
a con. Lastly, pit burials are those where no evi-
dence of a burial container exists and the individual
was placed directly in a grave cut feature. In a few
instances, it was not possible for the archaeologists to
discern the burial construction, whether due to poor
preservation or modern disturbances (Booth etal. ;
Clarke ). e burial constructions of these individ-
uals were le as unknown, and these individuals were
not incorporated into analyses based on burial con-
struction type.
Without direct evidence of how the people of Late
Roman Winchester buried their dead and may have
indicated or represented social status, it is important
to keep in mind that the analysis of the mortuary pro-
le does not determine who was high or low status.
Instead, these divisions oer broad measures of status
within the cemetery sample, identifying groups that
operate in relation to one another.
Dental health
Dentition and alveolar bone were examined macro-
scopically, and tooth presence/absence, presence/
absence of carious lesions, and degree of tooth wear
were recorded. If the alveolar bone indicated a tooth
was not present, suspected timing of tooth loss
Figure 2. Burial constr uctions used as a proxy for status at Late Roman Winchester. (A) Monumental burial
(Skeleton Number: WINCM:AY- ). (B) Wooden con burial (Skeleton Number: WINCM:AY- ).
(C) Pit burial (Skeleton Number: WI NCM:AY- ). Images adapted by LCA from Booth eta l. :
Fig.., Fig.., Fig...
Dental Health and Dietary Dierence at Late Roman Winchester
162
Figure 3. Dental health variables measured. (A) Antemortem tooth
loss of mandibular le premolars and mola rs; brackets indicate a rea of
tooth loss (Skeleton Number: LH ; old adult, female). (B) Carious
lesions of the right maxillary rst and second molar, indicated by the
arrow (Skeleton Number: LH ; middle adult, male). (C) Dental wear
of le maxillary dentition, arrow indicating toot h (rst molar) with
most marked wear (Skeleton Number: LH ; middle adult, probable
male). Photos by LCA.
(ante- or postmortem) was recorded, based on the de-
gree of resorption of the alveolar bone (Fig.a), fol-
lowing Buikstra and Ubelaker (). Antemortem
tooth loss was distinguished from congenitally absent
teeth by observing the space available for a tooth and
the presence or absence of wear facets on adjacent
teeth. Rates of AMTL were calculated as a percentage,
by dividing the number of sockets exhibiting AMTL
by the total number of observable sockets for the ante-
rior (incisors and canines), posterior (premolars and
molars), and total dentition.
Carious lesions were recorded using the system de-
veloped by Moore and Corbett (, ), record-
ing lesions where clear enamel demineralization
was visible (Fig.b). As this approach captures later
stages of carious lesions (as opposed to the recording
of “pre- cavitated” lesions), these values represent a
conservative expression of the true prevalence rate in
the sample (Hillson ). Although this approach
prevents comparison to modern populations, it does
allow for comparison to other bio archae ol o gi cal stud-
ies (Lukacs ). Caries rates were calculated as a
percentage, by dividing the number of teeth with car-
ious lesions by the total number of observable teeth,
for the anterior (incisors and canines), posterior (pre-
molars and molars), and total dentition. When con-
sidering total dentition, however, researchers must be
cautious, as each tooth type (e.g., molars, premolars,
canines, incisors) is at dierent risk for cariogenesis
(Hillson ), and individuals with inconsistent
proportions of these tooth types may present with
very dierent rates of carious lesions. In the current
study, it was decided to include these results for com-
parison to future studies; however, they should be
viewed and interpreted carefully.
Tooth wear was recorded for anterior teeth (inci-
sors, canine, premolars) following Smith () and
for posterior teeth (molars) following Scott ()
(Fig.c). e Smith () approach provides scoring
for teeth on an eight- point scale; however, the Scott
() method scores each quadrant of the molar on a
ten- point scale, allowing for scores between four and
for each molar. Consequently, the latter approach
facilitates a more detailed analysis of the molars, al-
lowing researchers to pick up on more subtle changes
in tooth wear between subsamples. Rates of tooth wear
were determined as an average, by adding all tooth
wear scores for the anterior teeth, divided by the num-
ber of anterior teeth present. e same equation was
then applied to the posterior teeth. As two dierent
methods were used to score the degree of wear, the
scores for the anterior and posterior teeth are pre-
sented and interpreted separately.
Statistical testing
Statistical analysis was completed using SPSS Statisti-
cal Package Version . e assumptions of normality
were not met. Consequently, dental health dierences
were investigated using non- parametric tests.
Because dental disease is age progressive, the age dis-
tribution of a sample can greatly inuence dental health
data and subsequent interpretations. us, dierences
in dental health were tested between age categories
(YA vs. MA, MA vs. OA, etc.) using Kruskal- Wallis,
for males (Supplemental Table S) and females (Sup-
plemental Table S), with a signicance level of %
(p < .). When a signicant dierence was identi-
ed, post hoc tests were conducted (Mann- Whitney
U) with a Bonferroni correction, to account for the
Type error (Supplemental Tables S and S). In these
instances, the signicance threshold was calculated
to be p = a/k, where a equals the original signicance
Avery et al. 163
Tab le 2 . Sample Distribution by Age, Burial Construction, and Sex.
YA MA OA Tot a l
All Burials
Male (.%) ( .%) (.%) (.%)
Female (.%) ( .%) (. %) ( . %)
Tot a l (.%) (.%) ( .%) (.%)
Pit Burials
Male (.%) (.%) (.%) (. %)
Female ( .%) (.%) (.%) ( .%)
Tot a l (.%) (.%) (.%) (.%)
Con Burials
Male (. %) (.%) (.%) (. %)
Female ( .%) (.%) (.%) (.%)
Tot a l (.%) (.%) (.%) (.%)
Monumental Burials
Male (.%) (.%) (.%) (.%)
Female ( .%) (.%) (.%) (. %)
Tot a l ( .%) (.%) (.%) (. %)
Unknown Burials
Male (. %) (.%) (. %) (.%)
Female (. %) (.%) (.%) (. %)
Tot a l (.%) (. %) ( .%) (.%)
YA = young adults (.– .years), MA = middle adu lts (.– .years),
OA = old adults (.+ yea rs).
Monumental burials include stepped burials and burials with built
enclosures.
Unknown burial construction is largely due to distu rbed burials from
modern construction projects.
level and k equals the number of hypotheses being
tested. If signicant dierences were noted in the post
hoc tests, then age- based categories were kept sepa-
rate in subsequent analyses. For example, rates of an-
terior tooth wear for males buried in cons were
signicantly dierent between young, middle, and old
adults (χ = ., df = , p = .), so these age cate-
gories are presented separately. However, within the
male and female subsamples (burial types combined,
or by independent burial types), no signicant dier-
ences were found between age categories with respect
to rates of posterior and total carious lesions. Ac-
cordingly, in the following sections these results are
not presented by age category but are grouped to-
gether to increase statistical strength. Small sample
sizes prevented age- based analysis for individuals in
monumental burials.
Dental health dierences between males and fe-
males (by burial construction and/or age) were tested
using Mann- Whitney U tests for non- parametric data,
with a signicance threshold of % (p < .).
Results
Of the individuals included in the current study,
.% (n = ) of the total sample was estimated as
male, while .% (n = ) was estimated as female
(Table ). Most individuals were buried in cons
(.%, n = ), with the least number of individuals
buried in monumental constructions (.%, n = ).
Additionally, young adults make up the largest por-
tion of the sample by age (.%, n = ), with middle
adults and old adults representing .% (n = ) and
.% (n = ), respectively.
When only sex is considered, males presented
higher rates of anterior AMTL but lower rates of
posterior and total AMTL compared to females,
with dierences for rates of posterior AMTL in mid-
dle adults being signicant (U = ., p = .;
Table). As age- based dierences in carious lesions
were not signicant for posterior dentition, these
data were collapsed together for all age categories
(males: χ = ., df = , p = .; females: χ = .,
df = , p = .). Compared to males, females exhibited
statistically higher rates of carious lesions, for poste-
rior dentition (U = ., p = .), anterior dentition
in young adults (U = ., p = .), and total dentition
in middle adults (U = ., p = .). Additionally,
males displayed higher rates of tooth wear than females,
but only the results for posterior dentition in old adults
were statistically signicant (U = ., p = . ).
Pit burials
When examining sex- based dierences for individu-
als buried in pits, rates of anterior and posterior den-
tal wear exhibited no statistical dierences (Table).
However, females buried in pits had statistically higher
rates of posterior carious lesions than males buried in
pits (U = ., p = .). In addition, in middle adult-
hood, males buried in pits demonstrated higher rates
of anterior AMTL than females buried in pits (U = .,
p = . ).
Con burials
Females buried in cons had higher rates of carious
lesions than males buried in cons; signicant dier-
ences were noted for the anterior (U = ., p = .)
and total dentitions (U = ., p = .; Table ).
Males in cons consistently exhibited higher rates of
tooth wear than females in cons, but only the re-
sults for the posterior dentition in old adulthood were
statistically signicant (U = ., p = .). Females
in cons displayed higher rates of posterior and total
AMTL than males; however, these dierences were
not statistically signicant.
Monumental burials
Considering individuals in monumental burials,
males had more posterior and total carious lesions,
greater average tooth wear for both anterior and
Dental Health and Dietary Dierence at Late Roman Winchester
164
posterior teeth, and higher rates of AMTL than fe-
males. However, Mann- Whitney U tests indicate
that there were no signicant dierences in dental
health variables between males and females buried in
monumental constructions (Table ). Small sample
sizes precluded statistical analysis by age category.
Discussion
Eating and drinking is about much more than caloric
intake and is oen related to what is considered appro-
priate for consumption, whether due to social, cul-
tural, or even nancial reasons (Cool ). In the
Roman Empire, diets were oen inuenced by social
dynamics and hierarchies of power, with men receiv-
ing a more generous share of the food in the family
unit, as they were deemed more productive and held
positions of power within the household (Garnsey
). ese perceptions of power and dietary dier-
ence permeated other aspects of society as well, includ-
ing medical treatises. For example, medical writers
advised men and women to consume dierent diets, due
to perceived gender- based nutritional requirements
and medical theories regarding the four humors. Spe-
cically, they considered women to be inherently wet
and cold and suggested that women consume hot and
dry foods to maintain balance (Garnsey ). Con-
versely, medical writers considered men to be hot and
dry and encouraged them to eat cold and wet foods
(e.g., fatty meats, marine resources) to balance their
humors (Oribasius .– , as quoted in Garnsey
:). However, these literary sources are largely
written for and about higher- status individuals in the
Tab le 3. Dental Health Data by Sex and Age Categor y.
Males Females Mann- Whitney U
Rates of AMTL % ( AMTL/ Sockets)
Anterior dentition (I, C)
YA . (/) . (/) U = ., p = .
MA . (/ ) . (/ ) U = ., p = .
OA . (/) . (/) U = ., p = .
Posterior dentition (P, M)
YA . (/) . (/) U = ., p = .
MA . ( / ) . (/) U = 218.0, p = 0.027
OA . (/ ) . (/) U = ., p = .
Total dentition
YA . (/,) . ( /) U = ., p = .
MA . (/, ) . (/ ) U = ., p = .
OA . (/) . (/ ) U = ., p = .
Rates of Carious Lesions % ( Carious Teeth/ Obser vable Teeth)
Anterior dentition (I, C)
YA . (/) . (/) U = 536.5 , p = 0.040
MA . (/) . (/) U = ., p = .
OA . ( / ) . (/ ) U = ., p = .
Posterior dentition (P, M)
All ages . (/,) . (/) U = 3031.5, p = 0.017
Total dentition
YA . (/) . (/ ) U = ., p = .
MA . (/) . (/) U = 265.0, p = 0.018
OA . (/) . (/) U = ., p = .
Avg. Dental Wear Score Avg. (Total Wear Score/ Observable Teeth)
Anterior dentition (I, C, P)
YA . (,/) . (, /) U = ., p = .
MA . (,/) . (,/) U = ., p = .
OA . ( /) . (,/) U = ., p = .
Posterior dentition (M)
YA . (,/) . (,/) U = ., p = .
MA . (,/) . (,/) U = ., p = .
OA . (,/ ) . (,/) U = 171.5, p = 0.020
Statistically signicant values are in bold. Statistically signica nt at p < .. Statistical ana lysis showed no
dierences by age categories for rates of posterior ca rious lesions, and thus, age categories were grouped together to
increase statistical strength, all others were divided by age categor y (Tables S– S).
YA = young adults (.– .years), MA = middle adu lts (.– .years), OA = old adults (.+ yea rs), All ages
(.+ years).
I = incisors, C = canine s, P = premola rs, M = molars .
Avery et al. 165
Tab le 4. Dental Health Data by Sex for Pit Burials.
Males Females Mann- Whitney U
Rates of AMTL % ( AMTL/ Sockets)
Anterior dentition (I, C)
YA . (/) . (/) U = ., p = .
MA . (/) . (/) U = 11.5, p = 0.028
OA . (/) . (/) U = ., p = .
Posterior dentition (P, M)
YA . (/) . (/) U = ., p = .
MA . (/ ) . ( / ) U = ., p = .
OA . (/ ) . (/) U = ., p = .
Total dentition
YA . (/) . (/ ) U = ., p = .
MA . (/) . (/) U = ., p = .
OA . (/) . (/) U = ., p = .
Rates of Carious Lesions % ( Carious Teeth/ Observable Teeth)
Anterior dentition (I, C)
All ages . (/ ) . (/) U = ., p = .
Posterior dentition (P, M)
All ages . (/) . (/) U = 82.0, p = 0.019
Total dentition
All ages . (/) . (/) U = ., p = .
Avg. Dental Wear Score Avg. (Total Wear Score/ Observable Teeth)
Anterior dentition (I, C, P)
All ages . (/) . (/) U = ., p = .
Posterior dentition (M)
All ages . (, / ) . (, /) U = ., p = .
Statistically signicant values are in bold. Statistically signica nt at p < .. Statistical ana lysis showed no
dierences by age categories for rates of carious lesions or dental wear, and thus, age categories were grouped
together to increase statistical strength, all others were divided by age category (Tables S–S).
YA = young adults (.– .years), MA = middle adu lts (.– .years), OA = old adults (.+ yea rs), All ages
(.+ years).
I = incisors, C = canine s, P = premola rs, M = molars .
Mediterranean region, and biological anthropologists
have questioned their applicability to life in Roman
Britain (Rohnbogner and Lewis ).
Investigating gendered dierences in diet, dental
health, and stable isotopic studies of Romano- British
samples have produced conicting results (Bonsall
; Chenery etal. ; Cheung etal. ; Fuller
etal. ; Grin etal. ; Müldner ; Redfern
etal. ). Even between archaeological sites associ-
ated with Late Roman Winchester, results are incon-
sistent. For example, studying eight sites across three
cemeteries, Bonsall and Pickard () did not nd
statistically signicant dierences in δC and δN iso-
topic values between males and females, suggesting
more equal access to resources for men and women.
Conversely, studying the Lankhills School site expan-
sion at Late Roman Winchester (used in the current
study), Cummings and Hedges () noted small but
signicant dierences in δN values, suggesting that
males consumed more marine protein or a more var-
ied diet than females.
e variability in previously conducted Romano-
British studies clearly demonstrates that factors be-
yond sex or gender contribute to dietary dierences,
whether due to dierences in social status, age com-
position of the samples, the inuence of migration, the
communities in which people were buried or had
spent previous portions of their lives, or, simply, local
diets. In this study, sample- wide results indicate that
females had higher rates of carious lesions than
males. When also considering age, females who died
in middle adulthood had higher rates of posterior
AMTL than males, while those who lived longer had
lower rates of tooth wear than males from the same
samples (Table). On the surface these results seem
to support Roman medical and social ideas that men
and women ate dierent foods, as it appears that
women living in Late Roman Winchester consumed
a soer and more cariogenic diet than men. However,
the consideration of dierent age groups complicates
this story and may indicate the long- term accumula-
tion of smaller dietary dierences rather than marked
gendered diets throughout the entire life course. Al-
ternatively, the age- related patterns could indicate
changes in diet as individuals aged, to a greater reli-
ance on soer foods for women and/or greater con-
sumption of more abrasive foods for men. Future
studies exploring the biomechanical changes in the
Dental Health and Dietary Dierence at Late Roman Winchester
166
Tab le 6. Dental Health Data by Sex for Monumental Burials.
Males Females Mann- Whitney U
Rates of AMTL % ( AMTL/ Sockets)
Anterior dentition (I, C)
All ages . (/) . (/) U = ., p = .
Posterior dentition (P, M)
All ages . ( /) . (/) U = ., p = .
Total dentition
All ages . (/) . (/) U = ., p = .
Rates of Carious Lesions % ( Carious Teeth/ Observable Teeth)
Anterior dentition (I, C)
All ages . (/) . (/) U = ., p = .
Posterior dentition (P, M)
All ages . (/) . (/) U = ., p = .
Total dentition
All ages . (/) . (/ ) U = ., p = .
Avg. Dental Wear Score Avg. (Total Wear Score/ Observable Teeth)
Anterior dentition (I, C, P)
All ages . ( /) . (/) U = ., p = .
Posterior dentition (M)
All ages . (/) . (/) U = ., p = .
Statistically signicant values are in bold. Statistically signica nt at p < .. Statistical ana lysis by age category
was not possible due to small sample sizes.
I = incisors, C = canine s, P = premola rs, M = molars. All ages = .+ yea rs.
Tab le 5. Dental Health Data by Sex for Con Burials.
Males Females Mann- Whitney U
Rates of AMTL % ( AMTL/ Sockets)
Anterior dentition (I, C)
YA . (/) . (/) U = ., p = .
MA . (/) . (/) U = ., p = .
OA . (/) . (/) U = ., p = .
Posterior dentition (P, M)
YA . (/ ) . ( /) U = ., p = .
MA . (/) . (/) U = ., p = .
OA . (/ ) . (/) U = ., p = .
Total dentition
YA . (/) . ( / ) U = ., p = .
MA . (/) . (/) U = ., p = .
OA . (/) . ( / ) U = ., p = .
Rates of Carious Lesions % ( Carious Teeth/ Observable Teeth)
Anterior dentition (I, C)
All ages . ( /) . (/) U = 1622. 5, p = 0.002
Posterior dentition (P, M)
All ages . (/,) . (/) U = ., p = .
Total dentition
All ages . (/,) . (/,) U = 1671.5 , p = 0.043
Avg. Dental Wear Score Avg. (Total Wear Score/ Observable Teeth)
Anterior dentition (I, C, P)
YA . (, / ) . (,/) U = ., p = .
MA . (,/) . (/) U = ., p = .
OA . (/) . (/ ) U = ., p = .
Posterior dentition (M)
YA . (,/) . (,/) U = ., p = .
MA . (,/ ) . ( / ) U = . , p = .
OA . (,/) . (,/) U = 60.0, p = 0.002
Statistically signicant values are in bold. Statistically signica nt at p < .. Statistical ana lysis showed no
dierences by age categories for rates of carious lesions, and thus, age categor ies were grouped together to increase
statistical strength, all others were divided by age categor y (Tables S-S).
YA = young adults (.– .years), MA = middle adu lts (.– .years), OA = old adults (.+ yea rs), All ages
(.+ years).
I = incisors, C = canine s, P = premola rs, M = molars .
Avery et al. 167
mandible in response to severe AMTL and age may
help elucidate this relationship (e.g., Hart etal. ).
We must also consider that the current sample is not
a representation of the living individuals at Late Ro-
man Winchester, but the dead. us, the age- related
dental health dierences could speak to dierent di-
ets consumed by, or conditions experienced by, those
who survived to old adulthood versus those who died
at a younger age. is area would certainly benet
from more focused research to more thoroughly
understand dietary changes over the life course in
Roman Britain.
Although Roman texts present women as inferior
to men, the lived experiences of men and women were
likely diverse and depended in part on an individual’s
status (Garnsey ). For example, legal texts demon-
strate that elite women in Rome oen had privileges
and responsibilities that were denied to freedwomen
or women who were slaves, such as the right to di-
vorce and remarry (Grubbs ). Meanwhile,
men with greater wealth and “cultural capital” ex-
erted greater inuence and control than freedmen
or slaves (Hemelrijk and Woolf :). erefore, by
exploring sex- based dental health differences
within status- based groups (as inferred through
burial construction), we can move beyond broad-
scale patterns and work to uncover some of the nu-
ances in gender- , age- , and status- based experiences
in past communities.
Pit burials
For those buried in earthen pits, the dental health data
indicate that females (ages combined) had higher rates
of posterior carious lesions than males (U = .,
p = .; Table ). Modern clinical literature has
demonstrated that higher rates of carious lesions in
females can sometimes be attributed to biological
factors. For instance, during pregnancy females ex-
perience an increase in estrogen, which can change
the composition of saliva and reduce the rate of sali-
vary ow (Laine ). Saliva plays an important role
in dental health by washing away debris and sugars
from the teeth; it also contains antimicrobial proteins
that buer against the formation of carious lesions
(Laine ). us, with altered salivary composition
and rate of salivary ow, pregnant females are oen at
risk for increased caries rates.
In studying changes associated with the adoption
of agriculture, Lukacs () applied these concepts
to an analysis of dental health and concluded that
higher caries rates in females aer the adoption of ag-
riculture were not due to an increased consumption
of cariogenic foods but rather to a higher frequency
of pregnancy, and was ultimately an indication of in-
creased fertility within the sample. Accordingly,
the higher rates of carious lesions in females buried
in pits at Late Roman Winchester may be due to rates
of fertility and pregnancy rather than strictly related
to dietary dierences between men and women.
While Roman fertility rates are dicult to assess
(Scheidel ), Molleson (:) comments that
fertility was low at the Romano- British cemetery of
Poundbury Camp. However, these authors do not de-
scribe how they determined fertility rates or if fertility
rates diered depending on burial construction. Simi-
lar data were not found for Roman Winchester, al-
though this could certainly help researchers better
understand the role of pregnancy and fertility on the
rates of carious lesions at this site.
Rates of carious lesions can also be aected by the
consumption of dierent foods, particularly carbohy-
drates and proteins (Hillson ). With this in mind,
the higher rates of carious lesions for females buried in
pits could be the result of women consuming greater
quantities of carbohydrates and sugars than men.
However, the lack of dierences in rates of tooth wear
(Table) suggests the consistency of the foods re-
mained similar.
When an individual consumes food, the pH level of
the oral environment falls within minutes, at which
point the enamel is more susceptible to demineraliza-
tion and the formation of a cavity (Hillson ).
Consuming additional foods before pH neutrality has
been met keeps the oral environment acidic for longer
periods, facilitating the formation of carious lesions
(Hillson ). us, considering rates of carious le-
sions and tooth wear concurrently, these data suggest
that women buried in pits consumed foods more fre-
quently than men buried in pits.
When focusing on middle adults, the dental health
results for those buried in pits also indicate that
males exhibited higher rates of anterior AMTL than
females. As AMTL can be caused by severe dental
wear and/or severe carious lesions, as well as cultural
and traumatic experiences, this does not provide
concrete information regarding dietary dierences
but reminds us to be cautious when interpreting den-
tal health data. Teeth are used for a variety of func-
tions, and diet is not static throug hout the life course.
Rather, individuals of varying ages may have had
dierent relationships with food over their life span.
Additionally, we must keep in mind that individuals
may have experienced a change in social position
through their life course. While an analysis of den-
tal health explores the entire life span (at least, aer
the formation of the permanent dentition), using
burial style as a proxy for status only captures the
Dental Health and Dietary Dierence at Late Roman Winchester
168
social position of an individual at their time of
death. us, discrepancies may exist when consid-
ering the ways in which people lived and the ways
in which they are buried.
Con burials
Examining dental health dierences for individuals
buried in cons, a dierent picture of gendered expe-
riences at Late Roman Winchester emerges. Rather
than sex- based dierences in rates of AMTL, as seen
in the pit burials, old adult males buried in cons ex-
hibited higher rates of posterior tooth wear (U = .,
p = .). Additionally, females exhibited higher rates
of carious lesions for the anterior and total dentitions
than males buried in cons (Table).
Exploring tooth wear patterns at Isola Sacra (rst to
third centuries A.D.; Italy), Prowse () hypothe-
sized that higher rates of tooth wear in males could
be due in part to the ingestion of sand and grit found
in seafood. is ts well with Roman medical texts,
which suggest that marine resources were appropriate
foodstus for men but not for women (Rufus .– ,
, , as quoted in Garnsey :).
Inhabitants of Roman Britain consumed a wide va-
riety of marine resources, although regional varia-
tions in these dietary practices existed (Locker ).
At Roman Winchester, shellsh are routinely recov-
ered in archaeological excavations, suggesting that
they formed a substantial part of the marine diet for
this community (Booth etal. ). As a civitas capi-
tal, this is not particularly surprising, as Roman Win-
chester would have had access to a wide variety of
foods during the Roman period (van der Veen etal.
). Stable isotope analyses of a Late Roman Win-
chester sample (included in this study) further demon-
strate that men may have consumed more marine
foods than women, as males had slightly higher δC
(by .‰) and δN (by .‰) values than females
(Cummings and Hedges ). Although the dier-
ences between the sex- based groups were small (and
the sample was not divided by burial construction),
the dierences in δN values between the sexes were
statistically signicant (t = ., df = , p = .)
(Cummings and Hedges :). us, at Late Ro-
man Winchester, the dierences in tooth wear for old
adults buried in cons may be the long- term result of
men consuming greater quantities of marine resources
than women, including sand, grit, bones, and shells.
However, as dierences in tooth wear were not pres-
ent between males and females in monumental buri-
als or those buried in pits, explanations based solely
on gender cannot adequately explain the dierences
found between males and females buried in cons,
and the role of status must also be considered.
Roman literary texts also suggest that middle- or
lower- status men oen worked outside the house, in
elds or workshops, while lower- or middle- status
women oen worked closer to a domestic space (Har-
low and Laurence ). While men and women likely
consumed some meals together, others were con-
sumed apart, and it may be dierences in these foods
or their patterns of consumption that resulted in the
dierent rates of tooth wear and carious lesions be-
tween males and females buried in cons at Late Ro-
man Winchester. At the Romano- British site of
Poundbury Camp, skeletal modications indicative of
squatting are more prominent in females than males
(Farwell and Molleson ). Although caution is
needed when inferring occupation from entheseal
changes (see Henderson and Cardoso ), these
modications may be the result of milling and grinding
and thus indicate that food preparation was “ woman’s
work” (Cool :). Consequently, women partici-
pating in these activities may have had greater access
to soened and prepared foods than men (Cool ).
is picture ts well with the data presented here,
as females have lower rates of tooth wear than males
in old adulthood, and higher rates of carious lesions,
suggesting that women had greater access to pro-
cessed foods than men. Conversely, these men who
were away from the house for hours a day may have
consumed food that was grittier and less processed.
Monumental burials
In contrast to individuals buried in cons, those in
monumental burials at Late Roman Winchester ex-
hibited very similar dental health (Table). ese re-
sults suggest that the diets consumed by men and
women in the higher strata of society in Late Roman
Winchester were more similar than ancient literary
sources suggest, both in terms of consistency and tex-
ture. While individuals buried in pits and cons
may have processed their own foods, Cool () sug-
gests that some people living in Roman Britain would
have had the means to buy processed meal and our
instead of milling it themselves. us the lack of dif-
ferences in dental health between males and females
in monumental burials may be related to their eco-
nomic position and their ability to aord processed
foods, which both men and women consumed.
However, the small sample size of this group means
that these interpretations must be made with caution.
Impact of status and age on dental health
and dietary dierences
Although sex- based dierences exist between males
and females at the sample level, further analysis
Avery et al. 169
demonstrates that the story is much more complex
when also considering burial construction and age.
Yet, the conclusions of the current study are not en-
tirely unique. Rather, similar results have emerged
from Poundbury Camp (Fuller etal. ; Richards
etal. ). Specically, Richards etal. () noted
that those buried in mausolea and lead- lined cons
at Poundbury Camp exhibited similar dietary isotope
values, suggesting that these men and women con-
sumed similar diets (Richards et al. ). Mean-
while, individuals buried in cons exhibited a great
deal of variation in dietary stable isotope values, sug-
gesting that some individuals received all their protein
from plant sources, while others received a large por-
tion of their protein from animal sources; sex- based
dierences were also present (Fuller etal. ). How-
ever, as Redfern etal. () demonstrate, Poundbury
Camp does not represent a “typical” Romano- British
cemetery and is anomalous in terms of demographic
proles, health patterns, and funerary practices.
Even so, the results of the current study and those at
Poundbury Camp demonstrate that dietary dierences
between men and women were not universal in Roman
Britain but were further inuenced by other factors of
identity. Additional studies of sex- and status- based
dierences are needed to better understand if this pat-
tern is true throughout Britannia or if Late Roman
Winchester and Poundbury Camp are exceptional ex-
amples of gendered experiences.
In Rome, literary sources suggest that food distri-
bution within the family unit was based on necessity
(Garnsey :). In a family where food resources
were readily available, dietary dierences may have
been minimized between men and women; alterna-
tively, in families where resources were strained, the
most productive members of the household may have
had preferential access to particular foods (Garnsey
). More recently, however, Garnsey etal. (:)
have suggested that this quintessential idea of a patri-
archal Roman family may not be an accurate represen-
tation of family life and that this notion persists due
to the lack of research directly challenging these con-
cepts. Romano- British researchers have also ques-
tioned this assumption, as diverse groups of peoples
populated the province of Britannia and likely held a
variety of ideas regarding the status of women and
“productive members” of the household (Allason-
Jones ; Sherratt and Moore ). In this vein, the
inuence of migration also needs to be considered.
Stable isotope analyses indicate that up to % of those
buried at Late Roman Winchester were born else-
where (Eckardt etal. ; Evans etal. ). ere is
also a group of individuals with distinct mortuary
practices whom Clarke () perceived to be “Panno-
nian” (i.e., from present- day Hungary, east of the
Danube). Examining strontium and oxygen isotopes,
Evans etal. () concluded that some of the indi-
viduals whom Clarke () identied were born out-
side of Roman Winchester but, for others, their
mortuary practices indicate a cultural or ethnic con-
nection, whether through kinship or marriage. us,
it is likely that these individuals (migrants and those
with a cultural/ethnic connection) and other migrant
populations at Late Roman Winchester may have had
dierent dietary preferences and practices, and possi-
bly dierent views regarding the roles and status of
men and women that did not conform to Roman
ideals as represented in ancient literary sources.
To move beyond these patriarchal ideals repre-
sented in Roman literature, researchers may need to
consider gendered divisions of labor. Groen- Vallinga
() suggests that, while women worked within the
domestic service in the Roman Empire, industries
such as building and transportation were exclusive to
men. In Roman Britain, epigraphic and archaeologi-
cal evidence for gendered divisions of labor is limited;
however, bioarchaeologists can use osteological data
and mortuary proles to investigate economic activi-
ties (Sherratt and Moore ). By using these lines of
evidence along with theoretical frameworks, such as
embodiment, we can begin to understand how dier-
ent biological outcomes for males and females (e.g.,
rates of dental health variables) may be the result of
daily activities and participation in dierent economic
industries.
As a civitas capital, Roman Winchester was home
to a ruling elite serving in political and administra-
tive capacities (James ). Participating in these
types of occupations may have allowed men and
women the opportunity to consume foods with simi-
lar consistencies and properties (i.e., carbohydrates
and proteins), eliminating dierences in dental health
between males and females buried in monumental
constructions. Although small sample sizes may ob-
scure subtle dierences for those buried in monumen-
tal constructions, these results support Garnsey
etal.’s () assertion that Roman society was not in-
herently patriarchal.
Beyond functioning as an administrative center,
Late Roman Winchester also had a variety of commer-
cial and cra industries typical of an urban commu-
nity, including a market, metal production, and a
possible textile factory or gynaecium (Bonsall ;
see Booth et al. :– for discussion). Analyz-
ing grave good distributions at Late Roman Win-
chester, archaeologists indicate that artifacts such as
spindle whorls and loom weights were found in as-
sociation with adolescent and adult female remains,
likely indicating an association with textile produc-
tion (Booth etal. :).
Dental Health and Dietary Dierence at Late Roman Winchester
170
Bridging the archaeological and osteological evi-
dence, our study suggests that dental health outcomes
for men and women at Late Roman Winchester were
not the result of an inherently patriarchal society and
a gendered division of diet, but due to participation in
dierent occupations and daily activities. In particu-
lar, women working within domestic spheres may
have eaten more frequently (as suggested by females
buried in earthen pits) and/or consumed soer and
more processed foods (as suggested by females buried
in cons) than men within the same social groups.
e incorporation of age categories, however, demon-
strates that these were not drastic dierences between
men and women, but subtler dierences that accumu-
lated over the life course.
Researchers have used osteological indicators to in-
vestigate gendered divisions of labor in other areas of
the Roman Empire as well. For example, studying
fracture patterns at the site of Aquincum in Roman
Hungary, Gilmour etal. () concluded that fracture
types and distributions suggest that men participated
more frequently in manual labor with greater hazards,
while women did not participate in the same risky
activities. Additionally, at Poundbury Camp, Farwell
and Molleson’s () study of musculoskeletal mark-
ers may suggest that women were largely responsible
for corn- grinding at this Romano- British site.
Conclusions
When considering only sex- based dierences in den-
tal health, the sample- wide results of this study sug-
gest that gender- based dietary dierences did exist at
Late Roman Winchester, with women consuming a
soer and more cariogenic diet than men. ese re-
sults seem to support Roman literary sources and
medical treatises regarding gendered dierences in
diet. However, these broad- scale patterns have the po-
tential to homogenize the past and obscure the range
and variation of lived experiences. us we must move
beyond dichotomous groupings and more consciously
consider cross- cutting variables of identity in the past.
When also incorporating age and burial construc-
tion, it is clear that gender dierences were not uni-
form but were inuenced by other aspects of identity
at Late Roman Winchester. e results indicate that,
for individuals buried in monumental constructions,
sex- based dierences in dental health were minimal.
However, we must be cautious in interpreting these
results, as small sample sizes prevented analysis by
age categories. For those buried in cons or in
earthen pits, dental health dierences were present
but may not be indicative of Roman patriarchal per-
ceptions of power. Rather, dental health dierences
at this site may be the result of dierent occupations
held by men and women, with women having
greater access to processed foods or more frequent
consumption than men. Furthermore, these dier-
ences took time to aect the dental health of both
males and females, as dierences in tooth wear and
AMTL were only present in middle and old adult-
hood. Although it is not surprising that there was
no singular experience for women or men living in
Late Roman Winchester, by incorporating two addi-
tional aspects of identity this study highlights that
moving beyond broad- scale patterns of dietary dif-
ference is possible and can result in a more nuanced
understanding of the past. e analysis of dental
health by sex, age, and status (as inferred through
burial construction) demonstrates that gendered ex-
periences at Late Roman Winchester were more
complex than previously considered and were inu-
enced by other aspects of identity. e current study
also emphasizes that dental health variables must be
considered in tandem with one another in order to
fully understand the dierences in rates ofAMTL,
carious lesions, and tooth wear. Future bio archae ol-
o gi cal analyses will benet from incorporating mul-
tiple aspects of identity to better understand the
complexity of life and the multitude of social groups
that existed in the past. To accomplish this, research-
ers may need to incorporate greater statistical meth-
ods (e.g., log- linear analysis or principal component
analysis) and work to explore dierences between and
within subsamples. By incorporating additional as-
pects of identity, such as gender, ability/disability,
ethnicity, population ancestry/aliation, religion/
beliefs, migrant status, and occupation, we may be
better able to move beyond categories represented
in binaries and see people in the past as complex and
diverse human beings.
Acknowledgments
e authors thank Helen Rees of the Hampshire Cul-
tural Trust, Katherine Barclay of Winchester Excava-
tions Committee, and Louise Loe at Oxford
Archaeology for access to the skeletal collections in-
cluded in this study. Additional thanks to the Win-
chester Excavations Committee for permission to
use the photographs included in this article, and to
Xuan Wei for her support and assistance during data
collection in the UK. Lastly, we thank the three
anonymous reviewers and the associate editor,
whose comments and suggestions helped improve
and clarify this manuscript.
is research was undertaken, in part, thanks to
funding from the Canada Research Chairs program.
Avery et al. 171
Fieldwork in the UK was supported by the Social
Sciences and Humanities Research Council of
Canada Insight Grant, File Number - -
(ID).
References Cited
Acsádi, György, and János Nemeskéri. . History of Human
Life Span and Mortality. Akademiai Kiado, Budapest.
Allason- Jones, Lindsay. . Women in Roman Britain. nded.
Council for British Archaeology, York.
Allason- Jones, Lindsay. . Women in Roman Britain. In
ACompanion to Women in the Ancient World, edited by Sha-
ronL. James and Sheila Dillon. Wiley- Blackwell, Chichester,
UK, pp.– .
Belcastro, Giovanna, Elisa Rastelli, Valentina Mariotti, Chiara
Consiglio, Fiorenzo Facchini, and Benedetta Bonglioli. .
Continuity or discontinuity of the life- style in central Italy
during the Roman Imperial Age– Early Middle Ages transition:
Diet, health, and behavior. American Journal of Physical An-
thropology ():– . DOI: . / ajpa . .
Binford, LewisR. . Mortuary practices: eir study and their
potential. American Antiquity :– .
Boldsen, JesperL., George R. Milner, LyleW. Konigsberg, and
James W. Wood. . Transition analysis: A new method
for estimating age from skeletons. In Paleodemography: Age
Distributions from Skeletal Samples, edited by Robert D.
Hoppa and James W. Vaupel. Cambridge University Press,
Cambridge, pp.– .
Bonsal l, Laura. . A comparison of female and male oral health
in skeleta l populations from Late Roman Britain: Implications
for diet. Archives of Oral Biology ():– . DOI: .
/ j . archoralbio . . . .
Bonsall, LauraA., and Catriona Pickard. . Stable isotope and
dental pathology evidence for diet in Late Roman Winchester,
England. Journal of Archaeological Science: Reports :– .
DOI: . / j . jasrep . . . .
Booth, Paul, Andrew Simmonds, Angela Boyle, Sharon Clough,
HillaryE. M. Cool, and Daniel Poore. . e Late Roman
Cemetery at Lankhills, Winchester: Excavations 2000– 2005.
Oxford Archaeology Monograph No.. Oxford Archaeology,
Oxford.
Brooks, Sheilagh, and Judy M. Suchey. . Skeletal age deter-
mination based on the os pubis: A comparison of the Acsádi-
Nemeskéri and Suchey- Brooks methods. Human Evolution
():– . DOI: . / BF.
Buikstra, Jane, and Douglas H. Ubelaker. . Standards for
Data Collection from Human Skeletal Remains. Arkansas Ar-
cheological Survey, Fayetteville.
Cannon, Aubrey, Brad Bartel, Richard Bradley, R. W. Chapman,
Mary Lou Curran, DavidW. J. Gill, S. C. Humphreys, C. Mas-
set, Ian Morris, Jerey Quilter, NanA. Rothschild, and Curtis
Runnels. . e historical dimension in mortuary expres-
sions of status and sentiment [and comments and reply]. Cur-
rent Anthropology ():– . DOI: . / .
Cardoso, HugoF. V. a. Epiphyseal union at the innominate
and lower limb in a modern Portuguese skeletal sample, and
age estimation in adolescent and young adult male and female
skeletons. American Journal of Physical Anthropology ():
– . DOI: . / ajpa . .
Cardoso, Hugo F. V. b. Age estimation of adolescent and
young adult male and female skeletons II, epiphyseal union at
the upper limb and scapular girdle in a modern Portuguese
skeletal sample. American Journal of Physical Anthropology
():– . DOI: . / ajpa . .
Chenery, Carolyn, Hella Eckardt, and Gundula Müldner. .
Cosmopolitan Catterick? Isotopic evidence for population mo-
bility on Rome’s northern frontier. Journal of Archaeological
Science :– . DOI: . / j . jas . . . .
Cheung, Christina, Hannes Schroeder, and RobertE. M. Hedges.
. Diet, social dierentiation and cultural change in Roman
Britain: New isotopic evidence from Gloucestershire. Archae-
ological and Anthropological Sciences ():– . DOI: .
/ s - - - y.
Clarke, Giles. . Pre- Roman and Roman Winchester: e
Roman Cemetery at Lankhills, Vol.. Clarendon Press, New
Yor k .
Cooke, Nicholas. . e Denition and Interpretation of Late
Roman Burial Rites in the Western Empire. Ph.D. dissertation,
University College London, London, UK.
Cool, HillaryE. M. . Eating and Drinking in Roman Britain.
Cambridge University Press, Cambridge.
Cucina, Andrea, and Vera Tiesler. . Dental caries and ante-
mortem tooth loss in the northern Peten area, Mexico: A bio-
cultural perspective on social status dierences among the
Classic Maya. American Journal of Physical Anthropology
():– . DOI: . / ajpa . .
Cummings, Colleen, and Robert Hedges. . Carbon and ni-
trogen stable isotopes analyses. In e Late Roman Cemetery at
Lankhills, Winchester: Excavations 2000– 2005, edited by Paul
Booth, Andrew Simmonds, Angela Boyle, Sharon Clough,
Hillary E. M. Cool, and Daniel Poore. Oxford Archaeolog y
Monograph No., Oxford Archaeology, Oxford, pp.– .
Eckardt, Hella, Carolyn Chenery, Paul Booth, Jane Evans,
A.Lamb, and Grundula Müldner. . Oxygen and strontium
isotope evidence for mobility in Roman Winchester. Journal
of Archaeological Science ():– . DOI: . / j . jas
. . . .
Eckardt, Hella, Gundu la Müldner, and Mar y Lewis. . People on
the move in Roman Britain. World Archaeology (): – .
DOI: . / . . .
Evans, Jane, Nick Stoodley, and Carolyn Chenery. . A stron-
tium and oxygen isotope assessment of a possible fourth cen-
tury immigrant population in a Hampshire cemetery, southern
England. Journal of Archaeological Science ():– . DOI:
. / j . jas . . . .
Farwell, Dennis E., and eyaL. Molleson. . Excavations
at Poundbury 1966– 80, Volume II: e Cemeteries. Dorset
Natural History and Archaeological Society Monograph Se-
ries , Dorset Natural History and Archaeological Society,
Dorset.
Fuller, BenjaminT., eyaI. Molleson, David A. Harris, Lau-
renT. Gilmour, and RobertE. M. Hedges. . Isotopic evi-
dence for breastfeeding and possible adult dietary dierences
from Late/Sub- Roman Britain. American Journal of Physical
Anthropology ():– . DOI: . / ajpa . .
Garnsey, Peter. . Food and Society in Classical Antiquity.
Cambridge University Press, Cambridge.
Garnsey, Peter, Richard Saller, and Martin Goodman. . e
Roman Empire: Economy, Society, and Culture.nded. Univer-
sity of California Press, Oakland.
Geller, PamelaL. . e Bioarchaeology of Socio- Sexual Lives:
Queering Common Sense about Sex, Gender, and Sexuality.
Springer, Switzerland.
Gilmour, RebeccaJ., Rebecca Gowland, Charlotte Roberts, Zsolt
Bernert, Katalin Klára Kiss, and Gabor Lassányi. . Gen-
dered dierences in accidental trauma to upper and lower limb
bones at Aquincum, Roman Hungary. International Journal of
Paleopathology :– . DOI: . / j . ijpp . . . .
Dental Health and Dietary Dierence at Late Roman Winchester
172
Gowland, Rebecca. . Playing dead: Implications of mortuary
evidence for the social construction of childhood in Roman
Britain. In TRAC 2000: Proceedings of the Tenth Annual eo-
retical Roman Archaeology Conference, edited by Gwyn
Davies, Andrew Gardner, and Kris Lockyear. Oxbow Books,
Oxford, pp.– .
Gowland, Rebecca. . e social identity of health in Late
Roman Britain. In TRAC 2003: Proceedings of the irteenth
Annual eoretical Roman Archaeology Conference, edited by
Ben Croxford, Hella Eckardt, Judy Meade, and Jake Weekes.
Oxbow Books, Oxford, pp.– .
Grin, Rebecca, Martin Pitts, Richard Smith, and Alan Brook.
. Inequality at Late Roman Baldock, UK: e impact of so-
cial factors on health and diet. Journal of Anthropological Re-
search ():– . DOI: . / jar . . . .
Groen- Vallinga, Mariam J. . Desperate housewives? e
adaptive family economy and female participation in the
Roman urban labour market. In Women and the Roman City
in the Latin West, edited by Emily Hemelrijk and Greg Woolf.
Brill, Leiden, e Netherlands, pp.– .
Grubbs, Judith Evans. . Women and the Law in the Roman
Empire. Routledge, New York.
Gustafson, Gösta, and Göran Koch. . Age estimation up to
years of age based on dental development. Odontologisk Revy
():– .
Harlow, Mary, and Laurence Ray. . Growing Up and Grow-
ing Old in Ancient Rome: A Life Course Approach. Routledge,
Ne w Yor k .
Hart, RichardT., VincentV. Hennebel, Nisra ongpreda, Wil-
liamC. Van Buskirk, and RonaldC. Anderson. . Model-
ing the biomechanics of the mandible: A three- dimensional
nite element study. Journal of Biomechanics ():– .
Hemelrijk, Emily, and Greg Woolf. . Introduction. In Women
and the Roman City in the Latin West, edited by Emily Hemel-
rijk and Greg Woolf. Brill, Leiden, e Netherlands, pp.– .
Henderson CharlotteY., and F. Alves Cardoso. . Special is-
sue entheseal changes and occupation: Technical and theore-
tical advances and their applications. International Journal of
Osteoarchaeology :– . DOI: . / oa . .
Hillson, Simon. . Dental Anthropology. Cambridge Univer-
sity Press, Cambridge.
Hillson, Simon. . Recording dental caries in archaeological
human remains. International Journal of Osteoarchaeology
():– . DOI: . / oa . .
Hillson, Simon. . Teeth . nded. Cambridge University Press,
Cambridge.
Hillson, Simon. . e current state of dental decay. In Tec h-
nique and Application in Dental Anthropology, edited by JoelD.
Irish and GregC. Nelson. Cambridge University Press, Cam-
bridge, pp.– .
Hollimon, Sandra E. . Sex and gender in bioarchaeological
research: eory, method, and interpretation. In Social Bioar-
chaeology, edited by Sabrina C. Agarwal and Bonnie A. Glen-
cross. Wiley-Blackwell, Oxford, pp. –.
Hollimon, SandraE. . Bio archae ol o gi cal approaches to non-
binary genders. In Exploring Sex and Gender in Bioarchaeol-
ogy, edited by SabrinaC. Agarwa l and JulieK. Wesp. University
of New Mexico Press, Albuquerque, pp.– .
James, Tom Beaumont. . English Heritage: Winchester. e
Bath Press, Bath, UK.
Joyce, RosemaryA. . Feminist theories of embodiment and
anthropological imagination: Making bodies matter. In Femi-
nist Anthropology: Past, Pre sent, and Future, edited by Pa melaL.
Geller and Miranda K. Stockett. University of Pennsylvania
Press, Philadelphia, pp.– .
Joyce, RosemaryA. . Sex, gender, and anthropology: Moving
bioarchaeology outside the subdiscipline. In Exploring Sex and
Gender in Bioarchaeology, edited by SabrinaC. Agarwal and
JulieK. Wesp. University of New Mexico Press, Albuquerque,
pp.– .
Kaidonis, JohnA. . Tooth wear: e view of the anthropol-
ogist. Clinical Oral Investigations (S):– . DOI: .
/ s - - - .
Katz, Darryl, and Judy Myers Suchey. . Age determination of
the male os pubis. American Journal of Physical Anthropology
():– . DOI: . / ajpa . .
Keenleyside, Anne. . Dental pathology and diet at Apollo-
nia, a Greek colony on the Black Sea. International Journal of
Osteoarchaeology ():– . DOI: . / oa . .
Laine, MerjaA. . Eect of pregnancy on periodontal and
dental health. Acta Odontologica Scandinavica ():– .
DOI: . / .
Locker, Alison. . In piscibus diversis: e bone evidence for
sh consumption in Roman Britain. Britannia :– . DOI:
. / .
Lukacs, John R. . Fertility and agriculture accentuate sex
dierences in dental caries rates. Current Anthropology ():
– . DOI: . / .
Lukacs, JohnR. . Sex dierences in dental caries experience:
Clinical evidence, complex etiology. Clinical Oral Investiga-
tions ():– . DOI: . / s - - - .
Lukacs, John R. . Bioarchaeology of oral health: Sex and
gender dierences in dental disease. In Exploring Sex
and Gender in Bioarchaeology, edited by Sabrina C Agrawal,
and Julie K Wesp. University of New Mexico Press, Albuquer-
que, pp.– .
Molleson, eyaI. . Social implications of mortality patterns
of juveniles from Poundbury Camp, Romano- British cemetery.
Anthropologischer Anzeiger ():– .
Moore, William, and Elisabeth Corbett. . e distribution of
dental caries in ancient British populations : Anglo- Saxon pe-
riod. Caries Research ():– . DOI: . / .
Moore, William, and Elisabeth Corbett. . e distribution of
dental caries in ancient British populations : Iron Age,
Romano- British and Mediaeval periods. Caries Research
():– . DOI: . / .
Müldner, Gundula. . Stable isotopes and diet: eir contri-
bution to Romano- British research. Antiquity ():– .
DOI: . / SX.
Ottaway, Patrick, K. Qualmann, Helen Rees, and G. Scobie. .
e Roman Cemeteries and Suburbs of Winchester: Excavations
1971– 86. Winchester Museums, Winchester, UK.
Pearce, John. . Status and burial. In e Oxford Handbook of
Roman Britain, edited by Martin Millett, Louise Revel, and
Alison Moore. Oxford University Press, Oxford, pp.– .
Phenice, Terrell Wayne. . A newly developed visual method
of sexing in the os pubis. American Journal of Physical Anthro-
pology ():– . DOI: . / ajpa . .
Prowse, TracyL. . Diet and dental health through the life
course in Roman Italy. In Social Bioarchaeology, edited by
SabrinaC. Agarwal and BonnieA. Glencross. Wiley- Blackwell
Press, Oxford, pp.– .
Redfern, RebeccaC., and SharonN. DeWitte. . Status and
health in Roman Dorset: e eect of status on risk or mor-
tality in post- conquest populations. American Journal of
Physical Anthropology ():– . DOI: . / ajpa
. .
Redfern, Rebecca, SharonN. DeWitte, John Pearce, Christine
Hamlin, and Kirsten Egging Dinwiddy. . Urban- rural dif-
ferences in Roman Dorset, England: A bio archae ol o gi cal per-
spective on Roman settlements. American Journal of Physical
Anthropology ():– . DOI: . / ajpa . .
Redfern, RebeccaC., Christine Hamlin, and Nancy Beavan Ath-
eld. . Temporal changes in diet: A stable isotope analysis
Avery et al. 173
of late Iron Age and Roman Dorset, Britain. Journal of Ar-
chaeological Science ():– . DOI: . / j . jas . .
. .
Redfern, RebeccaC., Andrew Millard, and Christine Hamlin.
. A regional investigation of subadult dietary patterns in
health in late Iron Age and Roman Dorset, England. Journal
of Archaeological Science ():– . DOI: . / j . jas
. . . .
Richards, M. P., R. E. M. Hedges, T. I. Molleson, and J. C. Vogel.
. Stable isotope analysis reveals variations in human diet
at Poundbury Camp cemetery site. Journal of Archaeological
Science ():– . DOI: . / jasc . . .
Robb, John, Renzo Bigazzi, Luca Lazzarini, Caterina Scarsini, and
Fiorenza Sonego. . Social “status” and biological “status”.
A comparison of grave goods and skeletal indicators from Pon-
tecagnano. American Journal of Physical Anthropology ():
– . DOI: . / ajpa . .
Rohnbogner, Anna, and Mary Lewis. . Dental caries as a
measure of diet, health, and dierence in non- adults from ur-
ban and rural Roman Britain. Dental Anthropology ():–
. DOI: . / daj . vi . .
Scheidel, Walter. . Progress and problems in Roman de-
mography. In Debating Roman Demography, edited by Walter
Scheidel. Brill, Leiden, e Netherlands, pp.– .
Scott, Eugenie Carol. . Dental wear scoring technique. Amer-
ican Journal of Physical Anthropology ():– . DOI:
. / ajpa . .
Sherratt, Melanie, and Alison Moore. . Gender in Roman
Britain. In e Oxford Handbook of Roman Britain, edited by
Martin Millett, Louise Revell, and Alison Moore. Oxford Uni-
versity Press, Oxford, pp.– .
Smith, B. Holly. . Patterns of molar wear in hunter- gatherers
and agriculturalists. American Journal of Physical Anthro-
pology ():– . DOI: . / ajpa . .
Sofaer, JoannaR. . Sex and gender. In e Body as Material
Culture: A eoretical Osteoarchaeology, edited by JoannaR.
Sofaer. Cambridge University Press, Cambridge, pp.– .
Stuckert, Caroline, and MorrieE. Kricun. . A case of bilat-
eral forefoot amputation from the Romano- British cemetery of
Lankhills, Winchester, UK. International Journal of Paleopa-
thology ():– . DOI: . / j . ijpp . . . .
Temple, DanielH., and Clark Spencer Larsen. . Dental caries
prevalence as evidence for agriculture and subsistence varia-
tion during the Yayoi period in prehistoric Japan: Biocultural
interpretations of an economy in transition. American Jour-
nal of Physical Anthropology ():– . DOI: . / ajpa
. .
van der Veen, Marijke, Alexandra Livarda, and Alistair Hill.
. New plant foods in Roman Britain— Dispersal and so-
cial access. Environmental Archaeology ():– . DOI: .
/ X.
White, ChristineD. . Gendered food behaviour among the
Maya: Time, place, status, and ritual. Journal of Social Archae-
ology ():– . DOI: . / .
