Sie sind auf Seite 1von 12


Life and Death in a Civitas Capital: Metabolic Disease

and Trauma in the Children from Late Roman
Dorchester, Dorset
Mary E. Lewis*
Department of Archaeology, School of Human and Environmental Studies, University of Reading,
Reading, Berkshire RG6 6AB, UK

Poundbury Camp; rickets; scurvy; rib fractures; anemia; urbanization

The impact that Romanization" and the
development of urban centers had on the health of the
Romano-British population is little understood. A reexamination of the skeletal remains of 364 nonadults
from the civitas capital at Roman Dorchester (Durnovaria) in Dorset was carried out to measure the health of
the children living in this small urban area. The cemetery
population was divided into two groups; the rst buried
their dead organized within an eastwest alignment with
possible Christian-style graves, and the second with more
varied pagan graves, aligned northsouth. A higher
prevalence of malnutrition and trauma was evident in the
children from Dorchester than in any other published
Romano-British group, with levels similar to those seen
in postmedieval industrial communities. Cribra orbitalia
was present in 38.5% of the children, with rickets and/or

scurvy at 11.2%. Twelve children displayed fractures of

the ribs, with 50% of cases associated with rickets and/or
scurvy, suggesting that rib fractures should be considered
during the diagnosis of these conditions. The high prevalence of anemia, rickets, and scurvy in the Poundbury
children, and especially the infants, indicates that this
community may have adopted child-rearing practices that
involved fasting the newborn, a poor quality weaning
diet, and swaddling, leading to general malnutrition and
inadequate exposure to sunlight. The Pagan group
showed no evidence of scurvy or rib fractures, indicating
difference in religious and child-rearing practices but that
both burial groups were equally susceptible to rickets and
anemia suggests a shared poor standard of living in this
urban environment. Am J Phys Anthropol 000:000000,
2010. V 2009 Wiley-Liss, Inc.

This study explores the health of an urban

Romanized population living in Dorchester through the
examination of the children buried at the cemetery of
Poundbury Camp. The health of the inhabitants of
Roman Britain is little understood, and population studies that integrate skeletal evidence with archaeological
and environmental data are few. Many questions remain
about what impact the introduction of urban centers and
the gradual economic decline at the end of the Roman
Empire had on the local population. Some Roman scholars argue that Romanization" brought about improvements to the health of the people (Mattingly, 2006, p
323), with the writing of Roman architects such as Vitruvius conjuring up images of planned Roman cities, with
marbled surfaces and owing water, providing extensive
facilities for the comfort and health of the inhabitants
(Morley, 2005). Diametrically opposed is the argument
that urbanization widened the divide between the rich
and the poor, with many suffering hardships of poverty,
social unrest, and subservience to the conquering population. The city of Rome itself is seen as overcrowded
and lthy, with dogs and muggers prowling the streets,
whereas the archaeological evidence calls into question
the organization of the water supply and levels of sanitation, with latrines discovered near kitchen areas (Laurence, 1997; Morley, 2005). Garnsey (1999) has hypothesized that undernutrition was endemic in the urban
communities of the Greco-Roman world, but there is little osteological evidence to support or dispute this controversial theory. Although the importance of the skeletal remains has been acknowledged, these data are still
marginalized in discussions of Romano-British archae-

ology (Gowland, 2003, p 137). Although, for the rst

time, Roberts and Cox (2003) have drawn together all
the current osteological evidence for health in Roman
Britain, we still have little concept about what life in
Romano-British urban centers was like, and how it
affected the health of the general population.
Nonadult remains provide an effective measure of population tness, as the ability of a society to keep their
most vulnerable members alive, and in good health,
attests their ability to adapt to their environment
(Lewis, 2007). Research on nonadult skeletons from
industrial areas has demonstrated the impact a poor
environment has on the health of the children (Lewis,
2002). Previous studies of children in Roman Britain
have focused on the aspects of their burial, infanticide,
and the life course, and there is a growing interest in
examining the remains for evidence for diet and disease
(Mays, 1993; Gowland, 2001; Gowland and Chamberlain,

C 2009


Grant sponsor: AHRC (Diasporas, Migration and Identities

research programme).
*Correspondence to: Mary E. Lewis, Department of Archaeology,
School of Human and Environmental Studies, University of Reading, Reading, Berkshire RG6 6AB, UK.
Received 18 May 2009; accepted 23 October 2009
DOI 10.1002/ajpa.21239
Published online in Wiley InterScience


2002; Redfern, 2007, 2008). In understanding the features of Roman society that may have impacted on the
health of children, we face certain challenges. Documentary evidence from Roman Britain is scarce, and any information on weaning, child rearing, education, health,
marriage, and employment of children comes mainly
from the Mediterranean. Arguments surrounding the
extent to which Roman" culture was adopted by the
native population in Britain are still ongoing, with
the idea of Romanization, whether conceptualized in
terms of imperially imposed acculturation or native emulation (Millett, 1990), gradually being replaced by the
models such as creolization (Webster, 2001). There is an
increasing awareness of the diversity of the responses to
Roman" inuence and the multitude of identities
expressed in the material culture and social practice
(Mattingly, 2002, 2006).
The children excavated from the late-Romano-British
cemetery of Poundbury Camp (3rd5th century AD) represent the largest sample of nonadult skeletons dating
from the period. This work reports on the recent reexamination of the 364 nonadults from Poundbury Camp
and focuses on new evidence for high levels of rickets,
scurvy, and trauma in the children. The reasons behind
the pathology in this nonadult group are discussed
within the context of urbanization in Roman Britain.

The cemetery at Poundbury Camp in Dorset was excavated between 1966 and 1982 and is located at the foot
of an Iron Age hillfort, overlooking the River Frome
(Green, 1987). More than 1,400 graves were identied,
with 1,200 burials recovered. The cemetery rst served
the Durotrigian Iron Age hillfort (1st century BC1st
century AD), before becoming incorporated into a small
rural Romano-British settlement (13 centuries AD),
with both groups burying their dead on the slopes of the
hillfort (see Fig. 1). By the end of the 3rd century AD,
Poundbury Camp had become the main burial location
for the people living in Durnovaria (modern Dorchester)
immediately to the east, a relatively small civitas capital
(town) originally created for the Durotriges of Dorset
(Wacher, 1992). Little is understood about the living conditions within this urban settlement, and this lack of
knowledge about what occupations people carried out,
how they sourced their food and materials, and the density and identity of urban sites is a common and wellrecognized problem within Roman archaeology (Jones,
1991a; Burnham et al., 2001). However, extensive excavations at Greyhound Yard (in the center of Durnovaria)
show the development of the town from timber buildings
to extensive stone houses in the 4th century. By the end
of the Roman period, these were often subdivided into
several living spaces, with adjacent-aisled buildings suggesting accommodation for tenants, laborers, and slaves.
Bronze working and iron smithing may also have been
carried out (Woodward et al., 1993). The presence of pig,
dog, cat, and raven bones indicates a build-up of refuse
at the site that attracted scavengers to middens. It is
likely that the inhabitants of Durnovaria followed a typical diet of meat (ox, sheep, goat, pig, chicken, red, and
roe deer), marine and freshwater sh and diary products, with plums, cabbages, and pears being introduced
to the area by the Romans (King, 1991; Smith, 1993;
Cool, 2006). Evidence for imported foods such as wine,
olive oil, gs, dates, grapes, corn, and chamomile has
American Journal of Physical Anthropology

also been found in the south of Britain during this period, as the result of extensive trade networks between
Britain and Europe (Jones, 1991b; Gale, 2003).
The skeletal sample comprised 796 adults and 404
nonadults. The children, including perinates, were
evenly distributed throughout the cemetery (see Fig. 2),
suggesting that they underwent the same burial treatment as the adults. There were no clusters of particular
age groups, with the exception of a group of infants to
the east of the cemetery, buried within a smithy (R14).
In the original report, the cemetery was divided into six
separate burial areas to aid in the analysis (see Fig. 2).
The Main" cemetery area comprised the majority of burials (n 5 1,028) organized in neat rows, typical of many
Romano-British urban cemeteries (Cleary, 1992). Most of
these burials were aligned eastwest and contained single inhumations. This area of the cemetery included 11
mausolea, thought to be reserved for the burial of highstatus individuals and their families (including their
slaves), as well as wooden, lead and stone cofns, and
burials packed with gypsum or plaster (Molleson, 1989).
There was a distinct lack of gravegoods. Burials, showing these characteristics were assigned to study Group
1. The other cemetery areas comprised graves that were
aligned northsouth with a variety of grave inclusions
and a lack of cofns and hobnails (Group 2).
The meaning behind these two seemingly contemporaneous burial types is still a matter of debate. The formal
alignment of the Group 1 graves has been suggested to
be the result of managed" Christian cemeteries, where
the mode of burial and eastwest alignment was dictated
by the Romanized" elite (Philpott, 1991; Sparey-Green,
The Group 2 burials included ve prone burials, and
the only late Roman crouched burial at the site. Grave
goods included glass and jet beads, bone pins and bracelets, and copper alloy brooches, and the three cremation
burials in this section of the cemetery indicates a small
segment of the population who continued to practice traditional burial rites, whereas the general mode of burial
was undergoing change (Farwell and Molleson, 1993,
p 30). It is also possible that these burials reect a
reintroduction of ancient burial practices at the end of
Roman occupation. Food offerings and hobnails are
thought to represent the pagan tradition of furnishing
for the afterlife and the journey across the Styx (Clarke,
1979; Baldwin, 1985). The similarity in the burial orientation and gravegoods suggests that the burials in Group
2 represent individuals that shared a similar identity
that was different from those of the general population
in Group 1.
The differences in burial rite between these two
groups provide a useful tool for comparing child health
between two potentially contrasting sections of RomanoBritish society.

The sample comprised 364 nonadults (\17 years).
Age-at-death estimates were obtained using standards of
development for the deciduous and permanent dentition
published by Moorrees et al. (1963a,b) and tabulated by
Smith (1991). Where no teeth were present, diaphyseal
lengths and skeletal maturation were used to assign an
age (Ubelaker, 1989). Perinates were aged using British
standards developed by Scheuer et al. (1980) based on
diaphyseal lengths. The skeletons were divided into


Fig. 1. Map showing the location of Poundbury and Dorchester (Durnovaria). After Farwell and Molleson (1993, p 3).

Fig. 2. Distribution of child graves (shaded) within the Poundbury Camp cemetery (adapted from Farwell and Molleson, 1993).
[Color gure can be viewed in the online issue, which is available at]

seven age categories, with individuals in the last age category (14.517.0 years), estimated to be older than 17.0
years when the root of the third molar was complete
(Rc 5 16.9) but the apex open (Moorrees et al., 1963b).
These biological ages are a useful tool for examining
health within key stages of a developing childs life (i.e.,
at birth, during infancy, and adolescence), but do not

necessarily coincide with the cultural concepts of childhood within the Roman period, where girls as young as
12 could be married, and males may not have been
viewed as fully adult until they were 25 years of age
(Revell, 2005, p 50). Vitamin D deciency (rickets and
osteomalacia) was diagnosed using the criteria set out
by Ortner and colleagues (Ortner and Ericksen, 1997;
American Journal of Physical Anthropology

TABLE 1. Age at death and burial location of the Poundbury Camp nonadults

Group 1
Group 2

40 (12)
15 (38)
55 (15)









79 (27)
8 (20.5)
87 (24)

45 (98)
1 (2)
46 (13)

50 (15)
5 (13)
55 (15)

38 (12)
3 (8)
41 (11)

34 (10)
0 (0)
34 (9)

10 (3)
3 (8)
13 (3.5)

30 (9)
3 (8)
33 (9)

325 (89)
39 (11)
364 (

The number in parenthesis is the rounded % of the total for each group.
The number in parenthesis in this column is the rounded % of the total nonadults.

Ortner and Mays, 1998) and Brickley and Ives (2006).

Rickets denotes the effect of vitamin D deciency on the
chondroblasts of the growth plate in children, whereas
osteomalacia refers to the more general changes resulting from the disruption to the osteoblasts (Pettifor,
2003). Cribra orbitalia was graded using the scheme by
Stuart-Macadam (1991, p 109) and an active and remodeled lesions recorded (Mensforth et al., 1978). Porotic
hyperostosis can occur in rickets, scurvy, and anaemia
and so was recorded in relation to other pathologies on
the skeleton. Although traditionally considered to be indicative of iron-deciency anaemia (Ortner, 2003), the
exact aetiology of cribra orbitalia and porotic hyperostosis
is still in question, and, recently, it has been suggested
that the marrow expansion typical of these lesions is the
result of megalobalstic anemia due to vitamin B12 deciency and gastrointestinal infections (Walker et al.,
2009). Other pathologies were recorded according to
standard published criteria (Buikstra and Ubelaker,
1994; Ortner, 2003). All three conditions, iron deciency,
vitamin D, and vitamin C deciency, are likely to have
occurred together in a malnourished child, and we should
expect to see skeletal manifestations of each of these conditions in a single skeleton. Osteopenia (loss of bone
mass) may also be expected as, while rickets affects ossication and scurvy the production of osteoid (Brickley and
Ives, 2008, p 48), a malnourished child is also likely to be
immobile, exacerbating the skeletal changes, and leading
to increasingly fragile bones. Osteopenia was identied
macroscopically when there was evidence of a thinned
cortex, sparse cancellous bone, and reinforced vertical trabeculae, known as sclerotic atrophy (Ortner, 2003, p 411).

Of the 364 nonadults from the late Romano-British
period, the majority came from the ordered cemetery
area (Group 1, 89%), with four (1.1%) located within
Mausolea (two perinates and two children aged 10.6
14.5 years) and six (1.6%) from high-status graves and
lead-lined cofns. Group 2 comprised 39 (11%) of the
nonadult burials (Table 1, Fig. 3a). Only 331 skeletons
were complete enough to be placed within an age category. Infants (i.e., 0.01.0 years) made up the largest
group (87 or 24%) with perinates (under 41 weeks) and
the 2.66.5-year-olds making up the second largest age
categories. Perinates occurred in both Groups, however,
there were signicantly more in Group 2 (38%; X2 5
18.57; P 5 0.001, d.f.1). In addition, Group 2 had the
largest number of adolescents (i.e., 14.617.0 years; Fig.
3b), but this was not statistically signicant.

Metabolic disease
Of the 200 nonadults with orbits preserved, 77 (38.5%)
had cribra orbitalia (Table 2), with Grades 2 and 3 formAmerican Journal of Physical Anthropology

ing the majority of lesions. Fifteen (19%) children had

the most severe forms of cribra orbitalia (Grades 4 and
5). These results are only slightly higher than the 36.4%
reported by StuartMacadam (1991). The ve most
severe cases of cribra orbitalia (Grade 5) were from
Group 1 (see Fig. 4), and in all age categories (from 0.0
to 14.5 years), active lesions (n 5 42) signicantly outweighed healed lesions (n 5 25; X2 5 7.64; P 5 0.01,
d.f.1). Between the ages of 0.01.0 years, all the cases
were active (n 5 11).
Of the 248 children with postcranial bones (Table 2),
31 (12.5%) had denite evidence for vitamin D (n 5 12
or 4.8%), vitamin C (n 5 12 or 4.8%), or combined vitamin D and C deciency (n 5 7 or 2.8%). A further 20
(8%) children had changes indicative of deciency disease, but the lack of a temporal or sphenoid bone or poor
preservation of the ends of ribs or long-bones made any
specic diagnosis impossible. Ten children (4%) had surviving cranial vaults with porotic hyperostosis presenting as layers of active new bone on the ectocranium, suggestive of vitamin C or D deciency over anemic
changes, but again, the lack of any postcranial bones
meant a specic diagnosis was not possible. Hence,
although 11% of the children at Poundbury are recorded
as having a specic metabolic disease, the true prevalence in this sample is probably much higher. Infants
were most commonly affected by rickets and/or scurvy
(54%), with nearly all cases of metabolic disease (90% of
the 31 individuals affected) coming from Group 1 (Table
2, Fig. 5). It is also notable that scurvy was only evident
in this ordered part of the cemetery, but this trend
proved not to be statistically signicant (X2 5 0.83; P 5
0.01, d.f.1) and probably results from the small sample
size of Group 2. Scurvy was the only condition to persist
after 2.6 years. Lesions associated with rickets, and
osteomalacia showed a range not usually recorded in
nonadult material. For example, one skeleton demonstrated folding of the unfused iliac crest, more often seen
in adult osteomalacia (Brickley et al., 2005). Parrot bossing of the skull (see Fig. 6) and extensive new bone formation in the orbits (see Fig. 7) in scurvy was also
observed, suggesting a more advanced form of the condition than pitting alone (Ortner et al., 2001). There were
also two cases of premature suture closure (in children
aged 2 and 12 years), a reported complication of rickets
(Wang et al., 2007). A striking feature of the remains
from Poundbury Camp was the degree of osteopenia in
the affected group (see Fig. 8), with 11 individuals demonstrating thin and friable bones with thickened trabeculae macroscopically. In each case, all the surviving
bones were affected (long bone shafts, skull, pelvis, and
ribs), suggesting a systemic condition such as severe
malnutrition and possibly that these children had restricted movement and poor muscle tone. Further quantitative analysis into the bone mineral density of the
Poundbury sample in relation to nutritional stress is


Fig. 3. A: Number of nonadults in each age category for the site and for each burial group. B: Percent distribution of nonadults
in each age category, by burial group.

Single and multiple rib fractures were recorded in 12
children (5.4% with ribs), with a total of 20 fractured
ribs (Table 3). All were located in Group 1. The ages
ranged from perinate to c.7 years of age. Where the ribs
could be identied, the majority were right mid ribs,
with between two and ve ribs affected in some individuals, and a mix of healed and healing lesions (see Fig. 9).
One child, aged around 1.5 years, had a possible healed
fracture of the rst rib, while another in the same age

group had a fractured left second rib. Six of these children also had evidence for other pathology on their skeleton (i.e., rickets, porotic hyperostosis, or periostitis). In
addition to fractures of the ribs, a child aged around 7
months from Group 1 presented a bucket-handle fracture of the distal metaphyses of the right tibia (see Fig.
10) conforming to the SalterHarris Scheme B (Salter
and Harris, 1963). Only one other case of child trauma,
a possible bowing deformity at Cannington (Brothwell
and Powers, 2000) has ever been reported in nonadult
remains from a Romano-British site.
American Journal of Physical Anthropology

TABLE 2. Number and percent metabolic disease in the nonadults, by cemetery location


Cribra orbitalia

Porotic hyperostosis

Vitamin D deciency

Vitamin C deciency

Vitamin C and D

Group 1
Total G1


2 (6)
14 (32)
18 (35)
19 (38)
17 (39)
1 (12.5)
71 (40)


5 (7)
6 (15)
3 (6)
2 (6)
2 (13)
18 (6.5)


0 (0)a
7 (16)
3 (8)
10 (4.5)


3 (7)
4 (11)
2 (5)
1 (3)
1 (3)
1 (75)
12 (5)


1 (3)
4 (9)
2 (5.5)
7 (3)


1 (25)
2 (100)
2 (75)
1 (100)
6 (25)


4 (50)
4 (14)
22 (7)


2 (40)
2 (7)
12 (4.8)


12 (4.8)


7 (2.8)

Group 2
Total G2

The number in parenthesis is the rounded % of the total for each group.

Fig. 4. True prevalence rate (TPR) of cribra orbitalia in the nonadults by burial group and severity of lesion.

The aim of this study was to re-examine the skeletal
remains of the children from Poundbury Camp to
explore the health of a community living in an urban
settlement during the Romano-British period. The division of the cemetery into two groups, the rst which buried its dead organized within an eastwest alignment
(Group 1, Christian), and the second with more varied
graves aligned northsouth (Group 2 pagan"), was carried out in an attempt identify any possible differences
in the life styles of the two groups. Statistically signicant differences between the two groups were not
American Journal of Physical Anthropology

evident with both showing a high prevalence of rickets

and anemia. Although the pagan" group did not have
any cases of rib fractures, this may be a result of the
smaller sample size. The lack of scurvy in the second
group may hint at differences in childhood diet, but this
study has not provided any conclusive evidence that the
different burial practices at Poundbury Camp reect a
different way of life for these two groups.
That Poundbury may represent a rather unusual population in Roman Britain has been suggested by Cool
(2006), who argues that the adoption of Christianity may
have resulted in the women of Poundbury Camp practicing an extreme form of fasting. Although it is hard to


Fig. 5. True prevalence rate (TPR) of children displaying metabolic disease in each burial group.

Fig. 7. Active new bone formation on the orbits of PC1308.

These lesions are thought to be the result of bleeding into the
orbits in scurvy.

Fig. 6. Classic parrot bossing" of the skull of PC1096. Indicative of scurvy.

substantiate such theories, the level of malnutrition in

the Poundbury Camp children is notable, and some cases
may be related to maternal health during pregnancy.
Most studies that report cribra orbitalia in the
Romano-British samples provide crude prevalence rates
(percent of total individuals affected) that are fairly low,
ranging from probably less than 5% in the Eastern
Cemetery in London (Conheeney, 2000, p 285) to 22.5%
at Ancaster (Cox, 1989). The true prevalence rate (based
on the number of children with orbits present) of 38.5%
at Poundbury Camp is high in comparison and is more
similar to those seen on the Continent where prevalence
rates range from 85% to 28% (Salvadei et al., 2001; Facchini et al., 2004; Cucina et al., 2006). Results from else-

where in the Roman Empire suggest that the high level

of cribra orbitalia at Poundbury Camp is not unusual for
children in the Roman period, but the rates from Dorchester, to date, are the highest reported in Britain.
In the original Poundbury Camp report, only three
possible cases of nonadult rickets were identied based
on the appearance of bowed bones and, apart from the
evidence for anemia, no other nutritional diseases were
noted (Farwell and Molleson, 1993, p 179). Advances in
palaeopathological diagnosis now mean that we are more
aware of the subtle changes related to rickets and/or
scurvy in the nonadult skeleton, such as pitting of the
sternal ends of ribs, metaphyseal fraying and friable,
and undermineralized diaphyses (Lewis, 2007). In the
past, such signs were probably interpreted as poor preservation, compounding the commonly held assumptions
that nonadult skeletons do not preserve well in the burial environment. Many of the bones of the children, suffering from rickets and scurvy in the Poundbury Camp
sample, were extremely fragile, and without knowledge
of the appearance of poorly ossied bone, they would
appear to be damaged postmortem. In this study, 31
American Journal of Physical Anthropology


Fig. 8. Osteopenia of the right ilium, ischium, and pubis of PC558. Note the porosity and thinning of the cortex (A) and thickened trabeculae (B) on the ilium that is separate to the postmortem damages at the edges of these bones (C).

TABLE 3. Rib fractures divided by age and cemetery location

Age category

Group 1


Group 2



The number in parenthesis is the rounded % of the total for
each group.

Fig. 9. Left rib of PC210 demonstrating a transverse rib

fracture and callus formation.

(11%) of the Poundbury children were diagnosed as having a specic metabolic disease. A literature review of
other Romano-British samples revealed only three cases
of rickets, and one possible case of infantile scurvy, from
sites in London (Conheeney, 2000; Mackinder, 2000).
This may be the result of differences in the methods
used to identify the conditions in these earlier reports,
rather than a true difference in the health patterns of
these samples. Once again, the rates at Poundbury
Camp are more similar to those reported from sites in
postmedieval England, where Brickley and colleagues
American Journal of Physical Anthropology

Fig. 10. Bucket-handle fracture of the right distal tibia of

PC1180. Note the extension of the left aspect of distal metaphysis beyond the level of the metaphyseal surface.


(2006) provide a true prevalence rate of 13% rickets and
10.5% scurvy in the infants from industrial Birmingham.

Child-rearing practices
We have no documentary evidence from Britain to
indicate whether Roman feeding practices were adopted
by the Iron Age population, but these methods may go
some way in explaining the level of malnutrition in the
youngest children from Poundbury Camp. Writing in the
early 2nd century AD, Soranus of Ephesus provides a
disturbing account of infant-feeding practices where children were starved for the rst 2 days of their lives,
denied colostrum, and often fed honey and goats milk as
a breastmilk substitute (Jackson, 1988; Temkin, 1991).
This diet would have left the child susceptible to dysentery and, as nonhuman protein causes irritation to
immature intestines, anemia (Jelliffe and Blackman,
1962). The gradual introduction of solid foods was recommended after 6 months of age (Temkin, 1991). Prowse
and co-workers (2008) caution that, as medical texts
were addressed to the wealthier members of society, or
to wet nurses charged with feeding slaves or foundlings,
they should not be seen to reect the wide variety of
practices that might be expected across the Empire. Isotopic evidence from Roman Egypt (Dupras et al., 2001)
and Italy (Prowse et al., 2004, 2007, 2008) indicates that
weaning was complete by 3 years. This is reected in
the carbon and nitrogen isotope evidence from Queensford Farm in England (Fuller et al., 2006) where a complete cessation of breastfeeding seemed to occur between
3 and 4 years. However, solid foods appear to have
entered the diet later than Roman medical texts suggest
at 1.5 years of age.
That rickets was a health problem in Roman Italy
may be inferred by the comments Soranus makes about
child rearing. Once the child is released from their swaddling clothes, mothers are warned not to let the children
walk so quickly as their legs . . . may become distorted
in the region of the thighs . . . the ground being too hard,
causing the bones to give a little, since the bones have
not yet become strong" (Temkin, 1991, p 116). It is possible that the practice of swaddling, limiting the skins
access to ultraviolet light, sporadic access to breastmilk,
and early weaning resulted in the high prevalence of
rickets in the infants from Dorchester. The widespread
distribution of vitamin D deciency throughout the cemetery site, and in both Groups, indicates that this was a
systemic problem in the community. Such high levels of
rickets in later populations are usually attributed to air
pollution in an urban or industrial environment (Lewis,
2002; Brickley et al., 2006).
Ascorbic acid is accessible to infants through mothers
breastmilk and in their fetal stores. If children are not
breastfed (Shorbe, 1953) or breastfeeding mothers are
decient in vitamin C (Fain, 2005), then infantile scurvy
may occur. Scurvy normally occurs in infants between 8
and 11 months (Brickley and Ives, 2006), and clinical
manifestations can appear after the child have lacked
the vitamin for as little as 24 months (Tamura et al.,
2000). Because of rapid growth in infancy, there is a
greater chance that they will develop defective blood vessels and show more obvious skeletal manifestations of
the condition than adults (Brickley and Ives, 2006). At
Poundbury Camp, seven infants had evidence of scurvy,
and this persisted in the children aged between 2.6 and

10.5 years, with 11 cases. It is difcult to understand

why the older children in urban Dorchester should have
been denied access to vitamin C, as fruit and vegetables
in the summer months should have been plentiful.
Although scurvy is most common during the winter
months of January and February (Brickley and Ives,
2008), other sources such as cabbage and root vegetables
should have been accessible. Perhaps, the condition
reects problems with a limited childhood diet, religious
dietary constraints, or food production, as vitamin C
would be destroyed in heated milk or stewed fruit
(Brickley and Ives, 2008, p 4244). Scurvy has been
linked to other acute illnesses or infectious disease
(Follis et al., 1950), and it is possible that these children
were already chronically ill, and scurvy developed during
their convalescence. Whatever the cause, cases of scurvy
occur throughout the organized cemetery, in both the
Mausolea and cofn burials of Group 1, but no cases
were noted in Group 2. It is possible that the diet of the
children from the pagan" group differed to the rest of
the population when it came to access to fresh fruit and
vegetables, despite their shared susceptibility to rickets
and anemia.
In 1882, Cheadle noted the coexistence of scurvy and
rickets in cases where children lacked access to fresh
fruit and were given boiled milk (Cheadle, 1882). Anemia, caused by the presence of scurvy, has also been discussed in the literature, either as the result of bleeding
into the tissues or the direct effect of ascorbic acids on
the production of red blood cells (Goldberg, 1963). Diagnosis of the coexistence of rickets and scurvy is challenging if the skull and long bones are not both present, as
the cranial bones are needed to identify the early pitted
changes in vitamin C deciency (Mays, 2008). It is also
likely that any sign of ossied haematomas in children
with both conditions would be absent due to the nature
of rickets, which hinders ossication. The presence of
both rickets and scurvy in the children from Group 1
(seven cases) indicates a lack of fresh fruit and vegetables, in addition to poor quantities of sunlight, and perhaps a vitamin-decient mother.
The recent suggestion by Walker et al. (2009) that the
marrow hypertrophy demonstrated in cribra orbitalia is
more likely to be the result of megloblastic anemia due
to deciencies in vitamins B12 and B9 (folic acid) is also
relevant. If mothers were denied access to animal products rich in vitamin B (i.e., sh, shellsh, beef, lamb,
eggs, and cheese) either through famine or religious
practices, they would pass this deciency on to their
newborn children, whose weaning diet would provide little of this nutrient. Poor-weaning foods, if coupled with
the unsanitary conditions of an urban environment,
could lead to diarrheal diseases, exacerbating the anemia and ultimately causing the death of the child. Such
a scenario would explain the high levels of cribra orbitalia in the young children at Poundbury Camp.

The prevalence of rib fractures in the children from
Poundbury Camp is remarkable. Although rib fractures
are the most commonly recorded type of fracture in British archaeological material (Brickley, 2006), they are
rarely considered in nonadults. In modern clinical medicine, rib fractures are commonly associated with physical
child abuse. Also, diagnostic is spiral and metaphyseal
fractures of the humerus, femur, and tibia (Brogdon,
American Journal of Physical Anthropology



1998; Resnick and Goergen, 2002). Fractures of the rst

rib and pelvis, in particular, are considered to represent
the most severe injuries in a child (Wilber and Thompson, 1998; Starling et al., 2002) brought about by compression forces, shaking or slamming the child down
onto a hard surface (Strouse and Owings, 1995). In a
survey of 165 postcranial fractures from suspected child
abuse cases, Kleinman and colleagues (1995) identied
rib fractures in 31 (51%) cases. However, Glass et al.
(2002) caution that rib fractures can also result from
generalized bone dysplasia, metabolic diseases, tumors,
and in infants with rickets, rib fractures due to general
osteopenia may be the rst clinical sign of this disease
(Pettifor and Daniels, 1997). Kopcsanyi and colleagues
(1969) discuss rib fractures associated with whooping
cough or any other chest infection associated with malnutrition and were seen in 50% children with rickets in
their sample. All the children with rib fractures at
Poundbury came from Group 1, and 50% also had evidence for rickets and/or scurvy and, in some cases, osteopenia. It is possible that the simple act of lifting these
children, perhaps while they were swaddled, caused fractures to the chest. That rib fractures in individuals with
rickets have not been previously reported in the archaeological literature may illustrate the severity of the condition at Poundbury Camp. In one case, rib lesions were
associated with extensive porotic hyperostosis and
changes to the facial bones, which suggest thalassaemia
intermedia as a cause for the rib changes, rather than
trauma (Lewis, Forthcoming).
Secondary rib fractures as the result of metabolic disease may explain some of the trauma at the site, but it
does not explain the presence of fractured ribs in the
other six children. Three of the children were too poorly
preserved to identify any other pathology, but three
(aged c. 44 weeks, 2 years, and 7 years) had rib fractures
unrelated to metabolic disease. Combined with the
bucket handle fracture on the tibia in a 7-month-old
child, this indicates a certain level of acute trauma in
this group, possibly as the result of physical abuse or, in
the older children, through corporal punishment, rough
games, or accidents. Modern cases of physical abuse are
seen to reect a high level of stress within the community as the result of overcrowding or parental consumption of alcohol, and the cases of infant trauma may demonstrate this tension. However, evidence for abuse in the
archaeological record is rare and needs to be put into
context (Lewis, 2007, p 181182). Roman society relished
violent sports; and attitudes to child rearing reportedly
involved the use of physical punishment that members
of Roman society would not have recognized as abuse"
(Rawson, 2003, p 176).

The impact that urban centers had on the health of
the Romano-British population has been a matter of
debate for many years. The lack of comparative skeletal
studies and a paucity of documentary evidence for this
part of the Roman Empire means that we are unsure
which aspects of Roman culture were adopted by the
general population after the conquest. Evidence from the
children buried at Poundbury Camp shows that there
was a high level of malnutrition in the sample, with
many children suffering from anemia, rickets, scurvy,
and general osteopenia. Suggestions that Poundbury
Camp is in some way unusual for Roman Britain are yet
American Journal of Physical Anthropology

to be substantiated and stem from a lack of data on the

health of individuals, and especially children, from contrasting sites from this period. However, levels of cribra
orbitalia are similar to those seen from urban centers
from the rest of the Roman Empire, and the prevalence
of rickets and scurvy is comparable to children living in
postmedieval England.
Several suggestions have been made for the high level
of anemia and vitamin deciencies in this group. It
seems likely that child-rearing practices that involved
starving the newborn, denying them colostrum, and
swaddling led to general malnutrition, infections, and
inadequate exposure to sunlight. Rib fractures may have
resulted from minor pressure being placed on fragile
bones as the swaddled child was picked up. If cribra
orbitalia and porotic hyperostosis are indeed the result
of vitamin B deciency in the population, then a maternal and infant diet of limited animal protein may be the
cause. Scurvy may also reect this overall pattern of
malnutrition, with a lack of access to fresh fruit and vegetables in the childhood diet. That scurvy and rib fractures were only evident in Group 1 may suggest that
these individuals adopted a different lifestyle to those in
the pagan" burial area, perhaps including religious
practices such as fasting and corporal punishment. But
both groups were equally susceptible to rickets and anemia, suggesting a shared polluted and unsanitary urban
living environment.
In conclusion, it seems that life in the civitas capital of
Durnovaria was detrimental to the health of the children
living there, with high levels of malnutrition and possible evidence for physical abuse. However, more work on
the health of nonadults from contrasting Roman environments (i.e., rural areas, villa sites, and military communities) is needed before we can establish whether an
urban environment was the cause.

I thank Drs. Robert Krusynski and Margaret Clegg
(Natural History Museum, London) for access to the
Poundbury Camp material, and permission to photograph the specimens, Edeltraud Asopeck for help with
Figure 2, and the Editor and anonymous reviewers for
their comments on the original manuscript. Dr. Hella
Eckardt and Prof. Mike Fulford provided thoughtful and
constructive comments on earlier drafts of this work for
which I am grateful.

Baldwin R. 1985. Intrusive burial groups in the later Roman
cemetery at Lankhills. Winchestera reassessment of the evidence. Oxf J Arch 4:93103.
Brickley M. 2006. Rib fractures in the archaeological record: a
useful source of sociocultural information? Int J Osteoarch
Brickley M, Berry H, Western G. 2006. The people: physical anthropology. In: Brickley M, Buteux S, Adams J, Cherrington
R, editors. St Martins uncovered investigations in the churchyard of St Martins-in-the-Bull Ring, Birmingham, 2001.
Oxford: Oxbow Books. p 90151.
Brickley M, Ives R. 2006. Skeletal manifestations of infantile
scurvy. Am J Phys Anthropol 129:163172.
Brickley M, Ives R. 2008. The bioarchaeology of metabolic bone
disease. Oxford: Academic Press.


Brickley M, Mays S, Ives R. 2005. Skeletal manifestations of
vitamin D deciency osteomalacia in documented historical
collections. Int J Osteoarch 15:389403.
Brogdon B. 1998. Child abuse. In: Brogdon B, editor. Forensic
radiology. New York: CRC Press. p 281314.
Brothwell D, Powers R. 2000. The human biology. In: Rathz P,
Hirst S, Wright S, editors. Cannington cemetery. London:
English Heritage.
Buikstra JE, Ubelaker D, editors. 1994. Standards for data collection from human skeletal remains. Fayetteville: Arkansas
Archeological Survey.
Burnham B, Collis J, Dobinson C, Haselgrove C, Jones M. 2001.
Themes for urban research c. 100 BC to AD 200. In: James S,
Millett M, editors. Britons and Romans: advancing an
archaeological agenda. Oxford: CBA. p 6776.
Cheadle W. 1882. Osteal or periosteal cachexia and scurvy. Lancet 2:4849.
Clarke G. 1979. Winchester studies 3: pre-Roman and Roman
Winchester Part II: the roman cemetery at Lankhills. Oxford:
Clarendon Press.
Cleary SE. 1992. Town and country in Roman Britain. In: Bassett S, editor. Death in towns: urban responses to the dying
and the dead, 100-1600. Leicester: Leicester University Press.
p 2842.
Conheeney J. 2000. Inhumation burials. In: Barber B, Bowsher
D, editors. The Eastern cemetery of Roman London excavations 1983-1990. London: Museum of London Archaeological
Services. p 277297.
Cool HEM. 2006. Eating and drinking in Roman Britain. Cambridge: Cambridge University Press.
Cox M. 1989. The human bones from Ancaster. London: English
Heritage Ancient Monuments Laboratory Report 98/89.
Cucina A, Vargiu R, Mancinelli D, Ricci R, Santandrea E, Catalano P, Coppa A. 2006. The Necropolis of Vallerano (Rome,
2nd-3rd century AD): an anthropological perspective on the
ancient Romans in Suburbium. Int J Osteoarch 16:104117.
Dupras T, Schwartz H, Fairgrieve SI. 2001. Infant feeding and
weaning practices in Roman Egypt. Am J Phys Anthropol
Facchini F, Rastelli E, Brasili P. 2004. Cribra orbitalia and cribra cranii in Roman skeletal remains from the Ravenna area
and Rimini (I-IV century AD). Int J Osteoarch 14:126136.
Fain O. 2005. Musculoskeletal manifestations of scurvy. Joint
Bone Spine 72:124128.
Farwell DE, Molleson TI. 1993. Excavations at Poundbury
1966-80 Volume II: the cemeteries. In: Draper J, editor. Dorset: Dorset Natural History and Archaeological Society.
Follis B, Park E, Jackson D. 1950. The prevalence of scurvy at
autopsy during the rst two years of age. Bull John Hopkins
Hosp 87:569591.
Fuller B, Molleson TI, Harris D, Gilmore L, Hedges R. 2006.
Isotopic evidence for breastfeeding and possible adult dietary
differences from late/sub Roman Britain. Am J Phys Anthropol 129:4554.
Gale J. 2003. Prehistoric Dorset. Stroud: Tempus.
Garnsey P. 1999. Food and society in classical antiquity. Cambridge: Cambridge University Press.
Glass R, Norton K, Mitre S, Kang E. 2002. Pediatric ribs: a
spectrum of abnormalities. Radiographics 22:87104.
Goldberg A. 1963. The anaemia of scurvy. Quart J Med 32:5164.
Gowland R. 2001. Playing dead: implications of mortuary evidence for the social construction of childhood in Roman Britain. In: Davies G, Gardner A, Lockyear K, editors. TRAC
2000 proceedings of the tenth annual theoretical Roman
archaeology conference, University College London April
2000. Oxford: Oxbow. p 152168.
Gowland R. 2003. The social identity of health in Roman Britain. In: Croxford B, Eckhardt H, Meade J, editors. Leicester:
Oxbow. p 135146.
Gowland R, Chamberlain A. 2002. A Bayesian approach to ageing perinatal skeletal material from archaeological sites:
implications for the evidence for infanticide in Roman-Britain.
J Arch Sci 29:677685.


Green C. 1987. Excavations at Poundbury, Dorchester, Dorset

1966-1982, Vol. 1: the settlements. London: Proceedings of the
Dorset natural history archaeological society, Monograph 7.
Jackson R. 1988. Doctors and diseases in the Roman Empire.
London: British Museum Press.
Jelliffe DB, Blackman V. 1962. Bahima disease: possible milk
anemia in late childhood. Tropical Pediatr 61:774779.
Jones R. 1991a. Cultural change in Roman Britain. In: Jones R,
editor. Roman Britain: recent trends. Shefeld: JR Collis. p
Jones R. 1991b. The urbanisation of Roman Britain. In: Jones
R, editor. Roman Britain: recent trends. Shefeld: JR Collis.
p 5366.
King A. 1991. Food production and consumptionmeat. In:
Jones R, editor. Roman Britain: recent trends. Shefeld: JR
Collis. p 1520.
Kleinman P, Marks S, Richmond J, Blackbourne B. 1995.
Inicted skeletal injury: a postmortem radiologic-histopathologic study in 31 infants. Am J Radiol 165:647650.
Kopcsanyi I, Laczay A, Nagy L. 1969. Cough fractures of ribs in
infants with dyspnoea. Acta Paediatr Acad Sci Hung 10:9398.
Laurence R. 1997. Writing the Roman metropolis. In: Parkins
H, editor. Roman urbanism: beyond the consumer society.
New York: Routledge. p 120.
Lewis M. 2002. The impact of industrialisation: comparative
study of child health in four sites from medieval and post-medieval England (850-1859 AD). Am J Phys Anthropol 119:211
Lewis M. 2007. The bioarchaeology of children. Perspectives
from biological and forensic anthropology. Cambridge: Cambridge University Press.
Lewis. Forthcoming. Thalassaemia intermedia at Poundbury
Mackinder A. 2000. A Romano-British Cemetery on Watling
Street: excavations at 165 Great Dover Street, Southwalk in
London. London: Museum of London Archaeological Services.
Mattingly D. 2002. Vulgar and weak: Romanisation, or time for
a paradigm shift? J Rom Arch 15:53640.
Mattingly D. 2006. An imperial possession. Britain in the
Roman Empire, 54 BC-AD 409. London: Penguin Books.
Mays S. 1993. Infanticide in Roman Britain. Antiquity 67:883
Mays S. 2008. A likely case of scurvy from early Bronze Age
Britain. Int J Osteoarch 18:178187.
Mensforth RP, Lovejoy OC, Lallo JW, Armelagos GJ. 1978. The
role of constitutional factors, diet and infectious disease in the
etiology of porotic hyperostosis and periosteal reactions in
prehistoric infants and children. Med Anthropol 2:159.
Millett M. 1990. The Romanisation of Britain. Cambridge: Cambridge University Press.
Molleson T, Cox M. 1988. A neonate with cut bones from Poundbury Camp, 4th century AD, England. Bull Soc Belge Anthropol Prehist 99:5359.
Molleson TI. 1989. Social implications of mortality patterns of
Juveniles from Poundbury Camp, Romano-British Cemetery.
Anthrop Anz 47:2738.
Moorrees CFA, Fanning EA, Hunt EE. 1963a. Formation and
resorption of three deciduous teeth in children. Am J Phys
Anthropol 21:205213.
Moorrees CFA, Fanning EA, Hunt EE. 1963b. Age variation of
formation stages for ten permanent teeth. J Dent Res
Morley N. 2005. The salubriousness of the Roman city. In: King
H, editor. Health in antiquity. New York: Routledge. p 192204.
Ortner D. 2003. Identication of pathological conditions in
human skeletal remains. New York: Academic Press.
Ortner DJ, Ericksen MF. 1997. Bone changes in the human
skull probably resulting from scurvy in infancy and childhood.
Int J Osteoarch 7:212220.
Ortner DJ, Butler W, Cafarella J, Milligan L. 2001. Evidence of
probable scurvy in subadults from archaeological sites in
North America. Am J Phys Anthropol 114:343351.
Ortner DJ, Mays S. 1998. Dry-bone manifestations of rickets in
infancy and early childhood. Int J Osteoarch 8:4555.

American Journal of Physical Anthropology



Pettifor J. 2003. Nutritional rickets. In: Glorieux F, Pettifor J,

Juppner M, editors. Pediatric bone: biology and diseases. New
York: Academic Press. p 541565.
Pettifor J, Daniels E. 1997. Vitamin D deciency and nutritional
rickets in children. In: Feldman D, Glorieux F, Pike J, editors.
Vitamin D. New York: Academic Press. p 663678.
Philpott R. 1991. Burial practices in Roman Britain: a survey of
grave treatment and furnishing AD43-410. Oxford: Archaeo Press.
Prowse T, Schwarcz HP, Garnsey P, Knyf M, Macchiarelli R,
Bondioli L. 2007. Isotopic evidence for age related immigration to Imperial Rome. Am J Phys Anthropol 132:510519.
Prowse T, Saunders S, Schwarcz H, Garnsey P, Macchiarelli R,
Bondioli L. 2008. Isotopic and dental evidence for infant and
young child feeding practices in an Imperial Roman skeletal
sample. Am J Phys Anthropol 137:294309.
Prowse T, Schwarcz H, Saunders S, Macchiarelli R, Bondioli L.
2004. Isotopic palaeodiet studies of skeletons from the Imperial Roman-age cemetery of Isola Sacra, Italy. J Arch Sci
Rawson B. 2003. Children and childhood in Roman Italy.
Oxford: Oxford University Press.
Redfern R. 2007. The inuence of culture upon childhood: an
osteological study of Iron Age and Romano-British Dorset, England. In: Harlow M, Laurence R, editors. Age and ageing in the
Roman Empire: approaches to the Roman life course; Journal
of Roman Archaeology Supplementary Series. p 171190.
Redfern R. 2008. A bioarchaeological investigation of diaspora
in late Iron Age and Roman Dorset, England (mid-to-late 4th
century BC to the end of the 4th century AD). Britannia
Resnick D, Goergen T. 2002. Physical injury: concepts and terminology. In: Resnick D, editor. Diagnosis of bone and joint disorders. Philadelphia: WB Saunders and Company. p 26272789.
Revell L. 2005. The Roman life course: a view from the inscriptions. Eur J Arch 8:4363.
Roberts C, Cox M. 2003. Health and disease in Britain. Gloucestershire: Sutton.
Salter R, Harris W. 1963. Injuries involving the epiphyseal
plate. JBJS 45:587622.
Salvadei L, Ricci F, Manzi G. 2001. Porotic hyperostosis as a
marker of health and nutritional conditions during childhood:
studies at the transition between Imperial Rome and the
early Middle Ages. Am J Hum Biol 13:709717.
Scheuer JL, Musgrave JH, Evans SP. 1980. The estimation of
late fetal and perinatal age from limb bone length by linear
and logarithmic regression. Ann Hum Biol 7:257265.

American Journal of Physical Anthropology

Shorbe H. 1953. Infantile scurvy. Clin Orthoped 1:4955.

Smith H. 1991. Standards of human tooth formation and dental
age assessment. In: Kelley MA, Larsen CS, editors. Advances
in dental anthropology. New York: Wiley-Liss. p 143168.
Smith R, editor. 1993. Excavations at County Hall, Dorchester,
Dorset, 1988. In the north-west quarter of Durnovaria. Nottingham: Wessex Archaeology.
Sparey-Green C. 2004. Living amongst the dead: from Roman
cemetery to post-Roman monastic settlement at Poundbury.
In: Collins R, Gerrard J, editors. Debating late antiquity in
Britain AD300-700. Oxford: Archaeo Press. p 103112.
Starling S, Heller R, Jenny C. 2002. Pelvic fractures in infants
as a sign of physical abuse. Child Abuse Neglect 26:475480.
Strouse P, Owings C. 1995. Fractures of the rst rib in child
abuse. Radiology 197:763765.
Stuart-Macadam PL. 1991. Anemia in Roman Britain: Poundbury Camp. In: Bush H, Zvelebil M, editors. Health in past
societies: biocultural interpretations of human skeletal
remains in archaeological contexts. Oxford: British Archaeological Research International Series. p 101113.
Tamura Y, Welch D, Zic A, Coopere W, Stein S, Hummell D.
2000. Scurvy presenting as a painful gait with bruising in a
young boy. Arch Pediatric Adolesc Med 154:732735.
Temkin O. 1991. Soranus gynecology. Baltimore: The John Hopkins University Press.
Ubelaker DH. 1989. Human skeletal remains. Excavation, analysis, interpretation. Washington: Taraxacum.
Wacher J. 1992. The towns of Roman Britain. London: BT Batsford.
Walker P, Bathurst R, Richman R, Gjerdrum T, Andrushko V.
2009. The causes of porotic hyperostosis and cribra orbitalia:
a reappraisal of the iron-deciency-anaemia hypothesis. Am J
Phys Anthropol 139:109125.
Wang P, Marcus J, Fuchs H, Mukundan S. 2007. Craniosynostosis secondary to rickets: manifestations on computed tomography. Radiol Case Rep 2:14.
Webster J. 2001. Creolizing the Roman Provinces. Am J Arch
Wilber J, Thompson G. 1998. The multiply injured child. In:
Green N, Swiontkowski M, editors. Skeletal trauma in children, 2nd ed. Philadelphia: W.B. Saunders Company. p 71
Woodward P, Davies S, Graham A. 1993. Excavations at Greyhound Yard, Dorchester 1981-4. In: Draper J, editor. Dorchester: Dorset.