close

Вход

Забыли?

вход по аккаунту

?

Brief and precarious lives Infant mortality in contrasting sites from medieval and post-medieval England (AD 850Ц1859).

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 134:117–129 (2007)
Brief and Precarious Lives: Infant Mortality in
Contrasting Sites from Medieval and Post-Medieval
England (AD 850–1859)
Mary E. Lewis1* and Rebecca Gowland2
1
School of Human and Environmental Studies, Department of Archaeology, University of Reading, Reading,
Berkshire RG6 6AB, UK
2
Department of Archaeology, University of Durham, South Road, Durham DH1 3LE, UK
KEY WORDS
equations
post-neonatal deaths; neonatal deaths; medieval England; Bays’ theorem; regression
ABSTRACT
This study compares the infant mortality profiles of 128 infants from two urban and two rural
cemetery sites in medieval England. The aim of this paper
is to assess the impact of urbanization and industrialization in terms of endogenous or exogenous causes of death.
In order to undertake this analysis, two different methods
of estimating gestational age from long bone lengths were
used: a traditional regression method and a Bayesian
method. The regression method tended to produce more
marked peaks at 38 weeks, while the Bayesian method produced a broader range of ages and were more comparable
with the expected ‘‘natural’’ mortality profiles.
At all the sites, neonatal mortality (28–40 weeks) outweighed post-neonatal mortality (41–48 weeks) with rural Raunds Furnells in Northamptonshire, showing the
highest number of neonatal deaths and post-medieval
Spitalfields, London, showing a greater proportion of
deaths due to exogenous or environmental factors. Of
the four sites under study, Wharram Percy in Yorkshire
showed the most convincing ‘‘natural’’ infant mortality
profile, suggesting the inclusion of all births at the site
(i.e., stillbirths and unbaptised infants). Am J Phys
Anthropol 134:117–129, 2007. V 2007 Wiley-Liss, Inc.
This study compares the infant mortality profiles of
128 infants from two urban (Christ Church Spitalfields,
St. Helen-on-the-Walls) and two rural (Raunds Furnells,
Wharram Percy) cemetery sites in medieval England with
the aim of interpreting the differences observed in terms
of endogenous or exogenous causes of death. Infant mortality has a profound effect upon the crude death rate of
a population and, consequently, is considered a sensitive
indicator of overall population ‘‘fitness’’ (Adler and
Ostrove, 1999; Saunders and Barrans, 1999; Murray and
Frenk, 2002). This study follows earlier work that examined the health of children from contrasting medieval
environments in an attempt to assess whether urbanization and industrialization had a detrimental effect on the
health of the inhabitants (Lewis, 2002). This previous
study demonstrated a decline in child health and an
increase in child mortality due to the detrimental environmental effects of industrial London. In the current
analysis, it is hypothesized that failure in the population’s
ability to adapt to these environments will be evident in
higher rates of infant mortality, with deaths related to
external factors (post-neonatal mortality) exceeding those
relating to internal causes (neonatal mortality) in the
industrial sample of post-medieval London.
Infant mortality is calculated as the number of infant
deaths in one calendar year, per 1000 live births (Stockwell, 1993). Infant survivability has been found to be
largely dependent on the physical environment into
which the child is born, in particular, the level of sanitation and availability of health care facilities. It is the
improvement in these factors that is largely believed to
have resulted in the rapid decline in infant mortality in
England and Wales during the eighteenth and early
nineteenth centuries, contributing to the ‘‘demographic
transition’’ (Williams and Galley, 1995). Estimated infant
mortality figures from modern preindustrial populations
have been found to vary widely, up to approximately 200
per 1000 live births (Hobbs and Kigguridu, 1992). Infant
mortality rates for many past populations are uncertain;
however, in sixteenth century England, it was estimated
that around 27% of children died before the age of
1 year (Orme, 2001). Graunt (1962) claimed that in
London during nonplague years, 36% of children died
under the age of 6 years, and that urban deaths
exceeded rural ones. However, given the inconsistency
with which stillbirths and neonatal deaths were reported
in the past, historical documents such as the London
Bills of Mortality, which recorded deaths in London from
the sixteenth century, are likely to reflect a minimum
number only (Robertson, 1996).
The absence of infant mortality records in England
prior to the late 1550s means that skeletal remains from
cemetery excavations are the primary source of evidence
for infant death. This archaeological evidence does, of
course, have many limitations, which have been addressed at length in the literature (e.g., Sundick, 1978;
C 2007
V
WILEY-LISS, INC.
C
*Correspondence to: Mary Lewis, School of Human and Environmental Studies, Department of Archaeology, University of Reading,
Reading, Berkshire RG6 6AB, UK. E-mail: m.e.lewis@reading.ac.uk
Received 29 June 2006; accepted 26 March 2007
DOI 10.1002/ajpa.20643
Published online 13 June 2007 in Wiley InterScience
(www.interscience.wiley.com).
118
M.E. LEWIS AND R. GOWLAND
Gordon and Buikstra, 1981; Johnston and Zimmer, 1989;
Goode et al., 1993). Differential burial practices, together
with preservation and recovery biases can limit the
extent to which infants excavated from cemetery populations represent the true proportion of those dying within
a particular society at any one time (Saunders, 2000).
Furthermore, anthropologists are unable to calculate
actual mortality ‘‘rates’’ from this type of information
because data concerning the total population at risk (i.e.,
the number of live births) is not available. It is also frequently impossible to narrow the dates of archaeological
samples into yearly, or even decadal, increments.
In the 1980s, palaeodemographers began to recognize
the influence fertility had on age at death distributions,
and hence the mortality profiles derived from archaeological samples (Sattenspiel and Harpending, 1983;
Buikstra et al., 1986; Milner et al., 1989; Paine and Harpending, 1996). Average age at death within a skeletal
population was demonstrated to be more a reflection of
its birth rate and fertility, rather than a measure of mortality or life expectancy (Sattenspiel and Harpending,
1983). Thus, a large number of infants within a cemetery
sample would reflect the greater number of infants
entering the population, rather than a population struggling to adapt to their environment (Larsen, 1997, p.
339). However, determining fertility in past populations
is problematic and in 1986, Buikstra and coworkers
introduced a method to examine nonadult mortality
structure by comparing the proportion of children dying
between 1 and 5 years (weaning age deaths) with those
who died between 1 and 10 years. Populations with a
high mortality would have a greater proportion of deaths
between 1 and 5 years. However, it does not follow that
a population with large numbers of infants in its mortuary sample is successful; it could also mean that the
high mortality rate of infants within the community
caused shorter birth intervals and a greater number of
pregnancies. The ability of the population to keep these
children alive into adulthood is the true measure of their
success.
We argue that examining mortality in archaeological
samples still has value and explore the age structure of
the infants that died in medieval England as a means of
elucidating the cultural and environmental factors that
led to their death. That is not to say that we are not
aware of taphonomic limitations and/or cultural biases is
burial practice. Although some authors have argued that
the generally low numbers of infant remains in our
archaeological populations may actually reflect true mortality rates (Brothwell, 1986–1987; Panhuysen, 1999),
cemetery samples, such as St. Andrew Fishergate, yielding only 11 infants over a 500 year period, suggest significant under-representation exists (Stroud and Kemp,
1993). In the rare cases where burial records from an
excavated sample survive, the proportion of infants
reported to have died, and those excavated, are at odds.
For example, the low number of infants from the Voegtly
cemetery in Pittsburgh, where 311 infants were reported
to have died but only 200 children under 5 years were
recovered, was suggested by Jones and Ubelaker (2001)
to demonstrate taphonomic bias in preservation. Conversely, Herring et al. (1994) found a greater number of
infants in the skeletal sample from Belleville, Ontario,
than were listed in the burial record. These studies demonstrate both the cultural and taphonomic problems inherent in interpreting infant mortality from both burial
registers and cemetery samples.
Despite these shortcomings, Saunders (2000) and
Saunders and Barrans (1999) have convincingly argued
the necessity and advantages of studying infant mortality rates from archaeological cemetery evidence. It is
unfortunate that in many studies of infant burials it is
issues of infanticide (Smith and Kahila, 1992; Mays,
1993, 2003; Gowland and Chamberlain, 2002) and problems of under-representation in palaeodemography that
are the main focus of attention, and in some cases, the
youngest children are excluded from studies altogether
(Bocquet-Appel, 2002; Paine, 2000; Ribot and Roberts,
1996). One of the key aspects of infant mortality studies
in bioarchaeology is that they provide information on the
population’s ability to adapt to their environment, their
cultural practices, fertility rates, demography, maternal
health, disease epidemics, birthing practices, and social
attitudes towards children.
ENDOGENOUS VERSUS EXOGENOUS
INFANT MORTALITY
Infant mortality rates are traditionally subdivided into
those who die before birth (late fetal or stillbirths), those
dying at birth or within the first 27 days of extrauterine
life (neonatal mortality), and those who die between 28
days and 1 year (post-neonatal mortality). It is an unfortunate drawback of archaeological studies of perinatal
mortality that we are unable to determine whether a
child was premature, stillborn, or survived for the first
few days of life (Saunders and Barrans, 1999). It is generally accepted that, in the past, children under the age
of 28 weeks, or 7 months from conception, were less
likely to survive the extrauterine environment due to
the immaturity of their vital internal structures. In forensic and archaeological science, the presence or absence
of the neonatal line has been used to assess whether a
child was stillborn, or was possibly the victim of infanticide (Smith and Avishai, 2005). Visible as a microscopic
hypoplastic defect on the deciduous teeth, the neonatal
line is thought to represent a period of enamel stasis after the stress of birth. Research on the survivability,
location, and visualization of the neonatal line has raised
questions about the reliability of this feature (Sarnat
and Schour, 1941; Whittaker and MacDonald, 1989);
however, it continues to be a useful marker in reconstructing the life history of the infant (Humphrey et al.,
2005). Nevertheless, it is not always possible to carry
out destructive analysis on collections held in National
museums or Universities, hindering such investigations
on a large scale.
Clinically, the majority of neonatal deaths are considered to reflect the endogenous state of the infant as the
result of genetic and maternal influences (e.g., congenital anomalies, prematurity, low birth weight, birth
trauma), and post-neonatal mortality is seen more as a
consequence of the child’s external environment or exogenous factors (infectious diseases, poor nutrition, poisonings, accidents) (Scott and Duncan, 1999). This ‘‘biometric model’’ developed by Bourgeois-Pichat (1951) works
on the premise that all endogenous deaths occur within
the first month, and that exogenous mortality is proportional to the number of days, after the first month, the
infant survives. The ability of a given population to provide the biocultural means for a child’s survival once it
is born provides an insight into their adaptive success.
Today in industrialized countries, sanitation measures
ensure that the vast majority of infant deaths are due to
American Journal of Physical Anthropology—DOI 10.1002/ajpa
119
INFANT MORTALITY IN MEDIEVAL ENGLAND
TABLE 1. Site information and number of infants included in the study
Site
Period (AD)
Context
Raunds Furnells
St. Helen-on-the-Walls
Wharram Percy
Christ Church, Spitalfields
Total
850–1100
950–1550
950–1500
1729–1859
Rural (early medieval)
Urban (later medieval)
Rural (later medieval)
Industrial (post-medieval)
endogenous factors (Stockwell, 1993), but the reverse
was true in urban centers of the past (Vögele, 1994). The
biometric model has been used by historical demographers to identify, among other things, breastfeeding patterns. Knodel and Kinter (1977) demonstrated that in
areas where breastfeeding was common practice, exogenous mortality remained low until weaning, and where
breastfeeding was uncommon, mortality was high in the
earlier months of life. However, both exogenous and endogenous mortality are affected by the nutritional status
of the mother and baby (Scott and Duncan, 1999).
Historical evidence suggests that neonatal deaths
were higher in rural areas than in urban communities,
where environmental factors have a greater effect on
child mortality, producing higher post-neonatal deaths
(Vögele, 1994). Examining the archaeological evidence,
Saunders et al. (1995) estimated the proportions of neonatal to post-neonatal deaths in infants from Belleville,
Ontario. They found that 74% of infant skeletons were
post-neonates and 26% neonates, reflecting the stronger
exogenous influence of infant death in this industrialized
Canadian sample. However, historical infant mortality
data from preindustrial England (1550–1649) has shown
neonatal mortality (excluding unregistered stillbirths) to
outweigh post-neonatal mortality samples (Hart, 1998).
In a study of birth and death records from Penrith,
Cumbria (1557–1806), Scott and Duncan (1998) found
that both neonatal and post-neonatal mortality fluctuated in accordance with the rise and fall of wheat prices.
However, neonatal mortality lagged behind by 1 year,
suggesting that early mortality was indirectly related to
the nutritional status of the mother during pregnancy,
whereas post-neonatal mortality was directly related to
lactation and weaning during the first year of life.
Previous research has indicated that in later medieval
England, urban and rural differences in health were limited, and that the greatest strain on the children in
these societies occurred during the post-medieval industrial period, where environmental factors resulted in
higher rates of infectious and metabolic disease, and
increased child mortality (Lewis, 2002). This study will
focus on the infant remains from four sites in medieval
and post-medieval England to explore whether infant
mortality patterns can provide more detail about the
contrasting environmental conditions during these periods, and whether inferences about birth practices and
maternal health can also be made from the data.
MATERIALS
The four sites were originally selected for their geographical location, large nonadult sample sizes, and contrasting settlement character (Table 1, Fig. 1). Raunds
Furnells, the earliest of the settlements, dates between
the ninth to twelfth centuries (Boddington, 1996) and
thus, predates the industrial developments that occurred
in England during the thirteenth century. The skeletons
Total
nonadults
Total infants
(% nonadults)
%
Nonadults
142
200
303
186
831
40
12
64
47
163
28.1
6.0
21.1
25.2
19.6
Fig. 1. Map of England showing the location of the study
sites.
from St. Helen-on-the-Walls (St. Helens), Aldwark (AD
950–1550) in later medieval York represent infants from
an urban sample. Medieval York was at its height in
terms of wealth and influence between the eleventh and
fifteenth centuries and was the focus of social, political,
and religious activities for the North of England
(Andrews, 1984). However, Aldwark was documented to
have been one of the poorest parishes in York (Palliser,
1980), and during the fourteenth century experienced a
shift in its demography from mostly craftsmen to
females, some of whom are suspected to have been working as prostitutes (Hall et al., 1988). Wharram Percy
(AD 950–1550), a rural settlement in the Yorkshire
Wolds, was contemporaneous with St. Helens and situated eight miles from the city of York (Bell and Beresford, 1987). Because of trading practices and its location,
individuals from Wharram Percy would probably have
been exposed to similar environmental pathogens as
their urban counterparts in York (Hall et al., 1988;
Milne and Richards, 1992). Finally, infants excavated
from the London crypt of Christ Church, Spitalfields,
dating from AD 1729 to 1859, were selected to represent
the industrial sample. The church is located in east London, and parish records reveal that the individuals
interred in the crypt were the middle to upper class
members of the congregation (Cox, 1996). Georgian society did not have the sophisticated sewage, water supply,
and disposal systems of the Victorian Age, but the people
of Spitalfields did benefit from the Lighting and Cleansing Act of 1759 and the Paving Act of 1778. In the
1700s, the population of London was estimated to be
American Journal of Physical Anthropology—DOI 10.1002/ajpa
120
M.E. LEWIS AND R. GOWLAND
20,000 and, whatever their status, the residents of Spitalfields would have experienced the effects of overcrowding, air pollution, and inadequate sewage disposal
(Cox, 1996).
raw long bone data from each of the sites in an appendix
so that it is available for further study using other
methods.
Neonatal and post-neonatal mortality
METHODS
Estimating age-at-death
Dental development is generally considered the most
accurate method of estimating age-at-death of nonadults.
However, the developing tooth crowns of perinates are
extremely fragile and this, together with the fact that
they are situated within relatively large crypt spaces in
the mandible and maxilla, means that they are frequently lost. Age estimates of perinatal skeletons are,
therefore, usually based on approximations from long
bone diaphyseal lengths (Clement and Kósa, 1992).
Errors in the relationship between diaphyseal lengths
and gestational age are well known, but difficult to control for in archaeological populations. In addition to factors pertaining directly to the mother, such as ethnic
group, age, height, and parity, external factors such as
season, social class, and pollutants have all been shown
to affect the size and weight of the infant at birth (Adair,
2004; Hauspie et al., 1994). It has also been proposed
that female fetuses mature earlier than do males in
terms of both leg length and weight (Lampl and Jeanty,
2003), potentially resulting in longer diaphyseal lengths,
and hence, older age estimates in female dominated
samples.
Despite these problems, long bone growth is still considered to be a useful index to estimate the age of perinatal skeletons because of the rapidity of growth during
this period (Jeanty and Romaro, 1984). Several wellknown standards have been produced for use on perinatal skeletal samples (e.g., Fazekas and Kósa, 1978; Scheuer et al., 1980). Similar studies have been made of long
bone diaphyseal length based on in utero ultrasound
measurement of long bones (e.g., Jeanty et al., 1981,
1982; Jeanty and Romaro, 1984). All of these methods
have a common element in their construction; they use
regression equations to derive age from long bone length.
A number of authors have demonstrated that ageing
techniques based on regression equations are statistically biased towards producing an age structure akin to
that of their reference population (Bocquet-Appel and
Masset, 1982, 1996; Konigsberg and Frankenberg, 1992;
Lucy et al., 1996; Aykroyd et al., 1997, 1999; Konigsberg
et al., 1997). Recent work in paleodemography has
emphasized the utility of Bayesian inference in age-atdeath estimation of skeletal populations (Hoppa and
Vaupel, 2002, p. 2). Bayesian data analysis allows us to
make inferences from data using probability models for
observable quantities (known age data) and for quantities that are unknown (archaeological data), but that we
wish to learn about (Gelman et al., 1995). Consequently,
this study of neonatal and post-neonatal mortality in
medieval England has employed two methods of age estimation. The first is the regression method produced by
Scheuer et al. (1980)—the most commonly used method
for ageing perinatal skeletons in the UK—and the second, a Bayesian method for ageing perinatal skeletons
produced by Gowland and Chamberlain (2002). While
some consensus on appropriate methods for estimating
age-at-death is being attained, debates continue (see
Hoppa and Vaupel, 2002; Chamberlain, 2006; Samworth
and Gowland, 2006). Consequently, we have included the
In historical demographic studies, cumulative exogenous mortality is proportional to the log 3 (n+1) where n
is the age in days within the first year of life. This cumulative death rate is then plotted on a graph, with the
vertical axis representing the number of deaths, and the
horizontal axis the days in a year, spaced according to
the above formula. The points representing the cumulative total of deaths will lie on a straight line or very
close to it. Projecting a horizontal line to the left until it
cuts the vertical axis, enables exogenous and endogenous
deaths to be separated (Wrigley, 1977). For archaeological populations, as we do not know the number of infant
deaths per year, in the long time span of our cemeteries,
or can we age our skeletons to as accurately as days,
such statistics cannot be applied. However, a crude estimate of exogenous and endogenous deaths can be established by looking at the numbers of individuals dying
between the ages of 28–40 weeks of gestation compared
with 41–48 weeks. Hence, the percentage of neonatal
deaths, including possible stillbirths (28–40 weeks) and
post-neonatal deaths (41–48 weeks), aged by both methods, were compared between rural and urban samples to
see if different patterns of infant mortality could be discerned.
RESULTS
Regression statistics
Of a total of 163 infant skeletons (under 48 weeks),
128 (78.5%) were aged using the available diaphyseal
lengths (humerus, radius, ulna, femur, tibia) for each
individual (Tables 2 and 3). In 1929, the Infant Life
Preservation Act set 28 weeks gestation as the period at
which a child was developed enough to be viable outside
the mother’s womb, although with modern medical intervention this period has been reduced to 24 weeks
(Knight, 1996). In the past, it is logical that the later a
child was born the more likely it was to survive. Traditionally, a full-term fetus is aged between 38 and 40
weeks gestation (Saunders, 2000) and for the purpose of
our study 40 weeks was taken as the time of birth. Any
individuals below this age were considered to have been
premature, with a reduced chance of survival, spontaneously aborted, or stillborn. This method would clearly
place any small-for-gestational-age (SGA) babies within
a younger, or premature, age category as methods to
identify these children are limited (Lewis, 2007).
The age at death of the infants based on regression
equations is presented in Figures 2–5. It is clear that
there is a stark under-representation of infants in the
St. Helens sample, and this has been a matter of debate
in the literature before. Only 12 (6%) infants were recovered from this site in a period spanning 500 years, with
one child found within the pelvis of an adult female
(Grauer, 1991; Lewis, 2002). The peak at 38 weeks using
regression equations, may suggests that children were
dying at birth, and has prompted speculation that these
infants were being killed by their impoverished mothers.
Documentary evidence from St. Helens has indicated
that many of the female tenants were involved in prostitution (Palliser, 1980), and it is possible that the high
American Journal of Physical Anthropology—DOI 10.1002/ajpa
121
INFANT MORTALITY IN MEDIEVAL ENGLAND
TABLE 2. Diaphyseal lengths (in millimeters)
Mean age
Humerus
Raunds Furnells
33.7
34.4
35.3
59.9
35.3
58.2
35.6
60
36.1
58.5
36.2
60.9
36.3
62.1
36.6
36.7
37.0
63.8
37.4
64.7
37.7
65.4
37.9
63.9
38.2
64.5
38.4
38.7
65.7
38.7
66.6
38.9
66
38.9
64.5
39.2
66.5
39.5
67.4
39.7
68.2
43.3
73.5
43.7
76.5
Wharram Percy
29.1
44.8
29.9
30.1
47.4
31.3
50.0
32.7
53.1
33.0
54.3
33.1
33.1
54.3
33.6
54.6
34.3
55.3
34.3
56.1
35.0
57.5
35.0
58.4
35.1
57.8
35.2
35.3
58.3
35.4
59.6
35.7
59.2
35.8
36.0
60.5
36.1
36.2
57.4
36.3
60.4
36.9
61.5
37.2
64.2
37.6
65.0
38.2
64.6
38.3
64.5
38.5
38.7
68.5
39.9
41.5
71.8
42.5
74.8
42.7
74.4
Radius
Ulna
Femur
48.6
61.2
63.2
65.9
46.6
49
47.8
53.9
49.6
56.1
61.2
52.7
51.7
55.5
58.3
57.9
61.8
58.6
66.8
68.4
68.7
68.9
69.9
70.1
71.1
72.3
73.2
Tibia
54.6
58.7
60
58.9
60.6
59.9
60.2
61.6
62
64.1
65.4
75.3
56.5
51.2
52.8
54.6
55.6
56.5
62.4
61.7
59.2
62.9
64.7
70.6
41.8
42.3
43.0
43.2
44.7
49.2
49.0
49.9
49.6
76.3
76.8
77
77.6
79.1
90.0
49.5
54.9
53.6
57.9
58.9
63.0
65.1
46.1
53.1
52.1
54.3
51.8
52.7
48.9
47.5
55.2
53.9
46.8
49.2
54.5
57.3
54.0
51.4
62.3
59.3
59.4
52.9
61.9
59.4
66.7
44.8
47.1
52.3
54.4
59.3
51.3
53.0
46.9
48.8
69.8
67.1
65.3
65.7
66.2
56.0
59.2
59.4
56.2
57.9
67.3
68.0
68.4
68.7
60.5
57.9
70.9
71.7
72.7
62.8
61.1
75.0
75.6
76.4
79.9
62.0
62.3
87.7
74.8
73.2
58.9
68.1
number of neonatal deaths represent their disposal by
women who needed to remain unburdened by babies to
continue in their trade. However, this pattern may also
reflect the burial practices of the medieval period, where
unbaptised infants were either disposed of away from
consecrated ground, or buried in a specific, unexcavated
area of the cemetery. Unfortunately, detailed plans of the
Mean age
42.8
43.5
43.7
44.0
44.1
44.5
44.6
44.6
44.7
45.0
45.0
46.1
46.2
46.8
47.0
47.5
47.6
Humerus
Radius
Ulna
Femur
Tibia
90.9
91.4
92.2
79.8
60.1
76.9
78.0
78.1
78.3
78.6
80.4
79.0
79.3
79.3
81.3
82.0
83.2
83.8
84.7
85.0
Christ Church Spitalfields
30.6
48.0
32.0
51.0
33.7
52.0
34.3
35.6
61.0
36.0
61.0
36.7
37.1
62.0
37.3
62.0
37.3
62.0
37.3
38.0
64.0
38.0
65.0
39.0
67.0
39.3
67.0
39.3
69.0
39.3
67.0
39.6
70.0
39.8
68.0
40.6
40.9
71.0
40.9
71.0
40.9
69.0
41.3
71.0
41.8
42.2
74.0
42.9
73.5
43.6
77.0
43.7
43.9
79.0
44.2
76.0
44.6
45.5
47.2
80.5
47.3
84.5
47.9
83.0
47.9
82.0
47.9
80.0
48.0
48.2
85.0
48.5
87.0
49.4
89.0
67.2
91.5
65.3
69.6
74.0
94.0
95.0
60.8
65.5
65.2
73.5
70.4
98.8
96.0
97.2
85.0
83.6
102.7
87.0
76.1
38.0
42.0
42.0
42.0
48.0
49.0
48.0
54.0
49.0
50.0
51.0
54.0
50.0
53.0
52.0
54.0
54.0
54.0
58.0
55.0
52.0
56.0
58.0
59.0
61.5
57.0
59.0
60.0
63.0
61.0
61.0
65.0
61.0
47.0
49.0
61.0
63.0
67.0
71.0
60.0
61.0
60.0
72.0
72.0
74.0
74.0
77.0
78.0
78.0
78.0
79.0
63.0
63.0
63.0
66.0
83.0
83.0
84.0
67.0
69.0
70.0
67.0
69.0
75.0
71.0
72.5
71.0
69.0
66.0
68.0
87.0
89.9
91.0
91.5
92.0
93.0
94.0
97.0
102.0
75.0
73.0
80.0
76.0
80.0
79.0
82.5
84.0
85.0
64.0
61.5
67.0
71.0
70.0
104.0
104.0
104.0
64.0
64.0
68.0
73.0
71.0
72.0
105.0
110
111
84.0
90.0
85.0
87.0
88.0
89.0
92.5
57.0
56.0
58.0
57.0
55.0
61.5
60.0
61.0
58.0
62.0
60.0
62.0
76.0
82.0
64.0
64.0
63.0
67.0
64.0
62.0
67.0
67.0
68.0
excavation are not available and it is not possible to
identify where within the cemetery the recovered infants
were located. What is clear is that the tiniest of babies
(under 38 weeks) are not in the sample and it is possible
that these bones were overlooked at the time of excavation. While the under-enumeration provides interesting
information in itself regarding attitudes towards infant
American Journal of Physical Anthropology—DOI 10.1002/ajpa
122
M.E. LEWIS AND R. GOWLAND
deaths, the impossibly small sample size meant that
data from this site was not examined further.
At Raunds Furnells, the majority of perinatal deaths
occur around the time of birth (36%). At Wharram Percy
and Spitalfields the deaths are more evenly distributed,
although with the majority of deaths occurring at birth
in the London site (19%) and much earlier, between 34
and 35 weeks (19.6%), at Wharram Percy. The Kolmogorov–Smirnov statistic (Siegel, 1956) indicates that
none of the overall distributions differ significantly from
each other at the 95% confidence level. At all the sites
with the exception of Spitalfields, neonatal mortality is
greater than post-neonatal mortality (Fig. 6).
Bayesian statistics
The use of Bayesian statistics in the estimation of
skeletal age-at-death has been strongly advocated in
TABLE 3. Number of perinates with measurable diaphyseal
lengths in the study sample, aged using data from
Scheuer et al. (1980)
Gestational
age (weeks)
28–29.9
30–31.9
32–33.9
34–35.9
36–37.9
38–39.9
40–41.9
42–43.9
44–45.9
46–47.9
48–49.9
Total
Raunds
Furnells
St.
Helens
Wharram
Percy
Christ
Church
Spitalfields
0
0
1
4
9
9
0
2
0
0
0
25
0
0
0
0
1
6
0
0
1
2
0
10
2
2
5
10
7
5
1
5
8
6
0
51
0
1
2
2
6
8
6
5
3
5
4
42
recent years (Hoppa and Vaupel, 2002). Fundamental to
Bayesian analysis is the use of prior probabilities; the
choice of which is important as it will have some impact
upon the results, particularly so for those skeletal indicators that have a poor correlation with chronological
age (Aykroyd et al., 1997). There are several approaches
that one can adopt when determining a prior. The
‘‘Rostock Manifesto,’’ as outlined by Hoppa and Vaupel
(2002), recommends that this value should be derived
from the archaeological data itself, by modeling the age
distribution most likely to have generated the distribution of skeletal indicators in the archaeological sample.
However, this requires that the archaeological sample be
of a sufficient size and of unbiased composition in order
to provide useful estimates (Boldsen et al., 2002).
Another approach is to assume that the value of the
prior is equal for all age categories (i.e., a uniform prior
probability). This is sometimes referred to as an ‘‘uninformative prior.’’ However, because population mortality
profiles would rarely have equal proportions of deaths
across all age categories this would not normally provide
the most appropriate values (Chamberlain, 2006, p. 114–
115). The approach adopted here is to use prior probabilities based on model life tables of natural perinatal
mortality as outlined in the method by Gowland and
Chamberlain (2002). Gowland and Chamberlain (2002)
produced posterior probabilities of age, given long bone
length, from a large sample of known age perinatal long
bone length data.
The age distributions we produced using this method
were compared to a ‘‘natural’’ profile derived from the
1958 Perinatal Mortality Survey (Butler and Alberman,
1969) of all 17,000 newborns dying within a one week
period in England and Scotland (estimated 98% of all
births; Figs. 7–9). At Raunds, the results provide a more
even spread of deaths from 32 to 44 weeks, with infants
now appearing at the 40 week (birth) category and at 44
week gestational weeks. There is still a paucity of chil-
Fig. 2. Number of infant deaths in each gestational week at Raunds estimated using regression statistics.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
INFANT MORTALITY IN MEDIEVAL ENGLAND
123
Fig. 3. Number of infant deaths in each gestational week at St. Helens estimated using regression statistics.
Fig. 4. Number of infant deaths in each gestational week at Wharram Percy estimated using regression statistics.
dren aged from 28 to 30 weeks compared to the
‘‘natural’’ mortality profile, but it follows it much more
closely than with the regression statistics. One might
expect this because the prior probability is based on natural perinatal mortality, though that is not to say that
the prior absolutely determines the outcome as evidenced by the different mortality curves produced. At
Wharram Percy, the profile changed from two peaks at
34 and 44 weeks, to a much more gradual peak of infant
deaths between 40 and 42 weeks, with many of the
infants shifting to the 40 weeks age category, again
reflecting the ‘‘natural’’ profile. The previous two peaks
have disappeared, but there are a number of deaths
occurring at 48 weeks, out of step with the ‘‘natural’’ profile. Finally, Spitalfields still shows an even spread of
infant deaths from 30 to 48 weeks, but the peak has
shifted from 36–38 to 40 weeks gestation, and the large
number of deaths in the 46 week category has disappeared. As with Wharram Percy, there is an increase in
the number of deaths at 48 weeks.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
124
M.E. LEWIS AND R. GOWLAND
Fig. 5. Number of infant deaths in each gestational week at Spitalfields estimated using regression statistics.
Fig. 6. Percentage of neonatal and post-neonatal deaths for each site using regression statistics.
The analysis of neonatal and post-neonatal mortality
has also altered with more children dying in the postneonatal period at all of the sites (Fig. 10). At Raunds,
however, neonatal deaths remain high (76%), and at
Wharram Percy they outweighed post-neonatal deaths.
The results from both of these rural sites are in tune
with documentary evidence for the medieval period
where deaths due to endogenous factors exceed those as
the result of exogenous causes. At Spitalfields, neonatal
and post-neonatal deaths are almost equal suggesting
that environmental factors were beginning to influence
the pattern of infant death in this industrial sample.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
INFANT MORTALITY IN MEDIEVAL ENGLAND
125
Fig. 7. Proportion of infant deaths in each gestational week at Raunds, compared to a ‘‘natural’’ profile. Estimated using Bayesian statistics.
Fig. 8. Proportion of infant deaths in each gestational week at Wharram Percy, compared to a ‘‘natural’’ profile. Estimated using
Bayesian statistics.
DISCUSSION
The aim of this study was to examine the impact of
rural and industrial environments on neonatal and postneonatal deaths in post-medieval England, using
archaeological infant remains and by employing two contrasting statistical methods. The analysis of infant skeletal material from archaeological sites is always challenging. Preservation issues, the skill of the excavator and
differential burial practices can all determine the number of infants recovered from a site, making direct com-
parisons of infant mortality problematic. While infant
mortality rates tell us something about the population’s
ability to adapt to their environment, examining the
ratio of neonatal and post-neonatal deaths within each
site has the potential to provide us with information on
the causes of death.
Data from burial records in England have shown that
during the later medieval period, infant mortality was
highest in the urban areas, and that in rural areas the
majority of deaths occurred due to ‘‘endogenous’’ factors,
resulting in high numbers of children dying at birth, or
American Journal of Physical Anthropology—DOI 10.1002/ajpa
126
M.E. LEWIS AND R. GOWLAND
Fig. 9. Proportion of infant deaths in each gestational week at Spitalfields, compared to a ‘‘natural’’ profile. Estimated using
Bayesian statistics.
Fig. 10. Proportion of neonatal and post-neonatal deaths for each site using Bayesian statistics.
within the first few days of life (Hart, 1998). In urban
areas, post-neonatal deaths tended to outweigh neonatal
deaths due to ‘exogenous’ causes, such as infections arising from encounters with the external environment
(Landers, 1990; Malhotra, 1990; Vögele, 1994). In nineteenth century Poland, Budnik and Liczbinska (2006)
reported higher neonatal death rates in the rural areas
as a result of ‘‘birth frailty,’’ possibly referring to prema-
turity or low birth weight in these children. In 1891, a
government report examined differences in infant mortality between three rural counties and three industrial
towns in England. It found that the overall risk of death
for infants born and reared in towns was more than double the rural picture (Williams and Galley, 1995). The
majority of urban deaths were caused by diarrheal diseases, which killed eight times more in the urban area
American Journal of Physical Anthropology—DOI 10.1002/ajpa
INFANT MORTALITY IN MEDIEVAL ENGLAND
(40%) than in the rural (5%), and crowd diseases such as
measles and scarlet fever were three times more prevalent, with tuberculosis double that of the rural areas.
While we know stillborn and unbaptised babies were
officially excluded from burial in the latter part of the
medieval period (Gilchrist and Slone, 2005), we have no
information on parental or cultural attitudes towards
the children who survived for a few days after birth
(neonates), as opposed to children who lived for weeks or
months before their death, and who may have made a
deeper emotional connection with their parents. If neonates were buried in a different location or manner to
post-neonates, this would affect the proportions of these
children recovered. At Raunds Furnells, whatever statistic is used, neonatal deaths exceed post-neonatal deaths,
indicating no such discrimination existed, and that endogenous factors such as congenital defects have more
influence on the death rate than environmental ones.
Infants make up 28% of the nonadult sample at this site,
but the very young infants (28–30 weeks) and older
infants (44–48 weeks) are few in number. At Wharram
Percy, the age distribution was similar to that expected
from a normal mortality profile when both stillborn and
neonatal deaths are accorded similar burials. The age
distribution for this site, however, is flatter than
expected, with a slight peak at 34–35 weeks, indicating
that a greater number of premature births were included
at this site. This suggests that there may have been a
greater tradition within rural areas of burying stillborn
or unbaptised infants within the main cemetery. The situation at Spitalfields may have been different. From documentary evidence we know that urban infants who
died while being wet-nursed in the countryside would
not have been returned home for burial and this may
have had some affect on the age profile of the infants in
the crypt. However, Molleson and Cox (1993, p. 132)
refer to the burial of at least one stillborn child, suggesting that midwives were performing baptisms in utero.
The spread of infant deaths is again similar to the
‘‘natural’’ mortality profile, indicating that the majority
of babies were buried within the crypt, with children
aged between 30 and 48 weeks all represented, and
infants making up a quarter (25%) of the nonadult
sample.
Overall, a much greater spread of gestational ages was
obtained at each site when the infants were aged using
the Bayesian method as opposed to the regression
method. This is what one would expect, as the Bayesian
technique does not assign all individuals with the same
long bone length to the same gestational age (as the
regression technique does), instead it takes into account
the fact that individuals with the same long bone length
may fall within a range of gestational ages (Gowland
and Chamberlain, 2002). It does this by incorporating a
probability distribution of gestational age given the long
bone length based on observations of a large sample of
known age data. Therefore, while this technique does
not provide ages for individuals, it does generate a probability distribution of gestational ages from the observed
long bone lengths, and as such would seem to be better
suited to the analytical purposes of this study. In addition to the broader range of gestational ages observed
using the Bayesian method, the peaks in ages at death
tend to be less pronounced and occur at 40 weeks as
opposed to 38 weeks when using the regression method
(see Gowland and Chamberlain, 2002 for a discussion of
the methodological reasons for this). Using the Bayesian
127
data, all sites with the exception of Raunds, show an
increase in infant deaths at the 48 week cut-off point.
This pattern is likely to have occurred because the
Bayesian method does not provide probabilities for those
infants dying over the age of 48 weeks. Hence, the upper
limit of this method is effectively truncated, and some of
the infants falling within this 48 week age category may
in fact have died at the slightly older ages of 50, 51, or
52 weeks.
At all sites, the regression and Bayesian results show
neonatal deaths outweighing post-neonatal deaths. However, the Bayesian results increase the number of children dying after 41 weeks, and at Spitalfields, the percentage of neonatal and post-neonatal deaths is almost
equal (see Figures 6 and 10). The London Bills of Mortality for the period of the Spitalfields crypt use, show
that infant mortality rates were high and fluctuated
between 341 per 1000 live births (1725–1749) to 151
(1825–1829), above the national rate, and greater than
the rural areas with rates of 191 and 144 respectively
(Galley and Shelton, 2001, p. 69). The greater number of
children surviving birth at Spitalfields may be explained
by the apparent wealth of the Spitalfields mothers who
were mostly skilled silk-weavers, and medical advances,
such as the introduction of the forceps and man-midwives in the early eighteenth century (Loudon, 1992;
Wear, 1992). The age of enamel hypoplasia formation
and cribra orbitalia in the Spitalfields sample, compared
to the other sites, suggest that weaning may have
occurred by around 7 months (Lewis, 2002), increasing
the child’s risk of exposure to contaminated food, diarrhea, and malnutrition, and contributing to the greater
post-neonatal mortality rates.
CONCLUSIONS
This study has analyzed and compared the mortality
patterns of infants excavated from both urban and rural
medieval sites in England, with the aim of interpreting
the differences observed in terms of endogenous or exogenous causes. In order to undertake this analysis, two
different methods of estimating gestational age from
long bone lengths were used: a traditional regression
method and a Bayesian method. The regression method
tended to produce more marked peaks at 38 weeks,
while the Bayesian method produced a broader range of
ages, although the differences in age distributions did
not affect the interpretation of the results significantly.
Of the three sites, it is Wharram Percy that provides the
most convincing evidence for a total infant sample, with
all infants represented and the expected peak between
38 and 42 weeks, indicating that both stillborn and neonatal deaths were accorded burial within the main cemetery. Raunds also showed a ‘‘natural’’ mortality pattern,
although fewer infants of younger gestational ages were
buried than at Wharram Percy. At Spitalfields a greater
proportion of infants were of post-neonatal age with
fewer very young, possibly stillborn, infants. It is likely
that there was a more strict observance of the exclusion
of unbaptised infants within this urban crypt than at
the rural burial sites of Wharram Percy and Raunds.
This pattern may also indicate the success of medical
advances during childbirth, coupled with the greater
impact of exogenous causes on infant death within this
urban setting, as the result of poor sanitation and early
weaning ages.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
128
M.E. LEWIS AND R. GOWLAND
ACKNOWLEDGMENTS
Part of this research was funded by a University of
Bradford Studentship. Thanks are also due to the people
and institutions that allowed ML access to the skeletal
material. Firstly, to the Yorkshire Museum and York
Archaeological Trust, for making the St. Helen-on-theWalls material available, to Northants County Council
for access to the Raunds Furnells collection via Bradford
University and, to Simon Mays (English Heritage) for
access to the Wharram Percy skeletons. ML is also
grateful to Louise Humphrey (Natural History Museum)
for access to the Spitalfields collection. We thank the
reviewers and Clark Spenser Larsen, whose useful comments helped to improve this paper.
LITERATURE CITED
Adair L. 2004. Fetal adaptations to maternal nutritional status
during pregnancy. Am J Phys Anthropol 38:50.
Adler NE Ostrove JM. 1999. Socioeconomic status and health:
what we know and what we don’t. Ann New York Acad Sci
896:3–15.
Andrews G. 1984. Archaeology in York: an assessment. In:
Addyman PV, Black VE, editors. Archaeological papers from
York: Presented to M.W. Barley. York: York Archaeological
Trust. p 173–187.
Aykroyd R, Lucy D, Pollard A, Roberts C. 1999. Nasty, brutish,
but not necessarily short: a reconsideration of the statistical
methods used to calculate age at death from adult human
skeletal and dental age indicators. Am Antiquity 64:55–70.
Aykroyd R, Lucy D, Pollard A, Solheim T. 1997. Technical note:
regression analysis in adult age estimation. Am J Phys
Anthropol 104:259–265.
Bell RD, Beresford M. 1987. Wharram Percy: the Church of St.
Martin. London: Society for Medieval Archaeology.
Bocquet-Appel JP. 2002. Paleoanthropological traces of a neolithic demographic transition. Curr Anthropol 43:637–650.
Bocquet-Appel JP, Masset C. 1982. Farewell to paleodemography. J Hum Evol 11:321–333.
Bocquet-Appel JP, Masset C. 1996. Paleodemography: expectancy and false hope. Am J Phys Anthropol 99:571–583.
Boddington A. 1996. Raunds Furnells. The Anglo-Saxon church
and churchyard. London: English Heritage.
Boldsen JL, Milner GL, Konigsberg LW, Wood JW. 2002. Transition analysis: a new method for estimating age from skeletons. In: Hoppa RD, Vaupel JW, editors. Paleodemography:
age distributions from skeletal samples. WP: Cambridge:
Cambridge University Press. p 73–106.
Bourgeois-Pichat J. 1951. La measure de la mortalité infantile:
I. Principes et méthodes. Population 6:233–48.
Brothwell DR. 1986–7. The problem of the interpretation of
child mortality in earlier populations. Anthropol Port 4–5:
135–143.
Budnik A, Liczbinska G. 2006. Urban and rural differences in
mortality and causes of death in historic Poland. Am J Phys
Anthropol 129:294–304.
Buikstra J, Konigsberg L, Bullington J. 1986. Fertility and the
development of agriculture in the prehistoric midwest. Am
Antiquity 51:191–204.
Butler NR, Alberman ED. 1969. Perinatal problems: the second
report of the 1958 perinatal mortality survey. Edinburgh: Livingstone.
Chamberlain A. 2006. Demography in archaeology. Cambridge:
Cambridge University Press.
Clement J, Kósa F. 1992. The fetal skeleton. In: Clark D, editor.
Practical forensic odontology. London: Butterworth-Heinemann,
p 43–52.
Cox M. 1996. Life and death in Spitalfields 1700 to 1850. York:
Council for British Archaeology.
Fazekas IG, Kósa F. 1978. Forensic fetal osteology. Budapest:
Academic Press.
Galley C, Shelton N. 2001. Bridging the gap: determining longterm changes in infant mortality in pre-registration England
and Wales. Popul Studies 55:65–77.
Gelman A, Carlin J, Stern H, Rubin D. 1995 Bayesian data
analysis. London: Chapman & Hall.
Gilchrist R, Slone B. 2005. Requiem: the monastic cemeteries in
Britain. London: Museum of London.
Goode H, Waldron T, Rogers J. 1993. Bone growth in juveniles:
a methodological note. Int J Osteoarch 3:321–323.
Gordon CC, Buikstra J. 1981. Soil pH, bone preservation,
and sampling bias at mortuary sites. Am Antiquity 48:566–
571.
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 29:677–685.
Grauer AL. 1991. Life patterns of women from medieval York.
In: Walde D, Willows ND, editors. Proceedings of the 22nd
Annual Chacmool Conference. University of Calgary, Chacmool: The Archaeological Association of the University of Calgary. p 407–413.
Graunt J. 1662. Natural and political observations. London:
Thomas Roycroft.
Hall RA, MacGregor H, Stockwell M. 1988. Medieval tenements
in Aldwark, and other sites. London: Council for British
Archaeology.
Hart N. 1998. Beyond infant mortality: gender and stillbirth in
reproductive mortality before the twentieth century. Popul
Studies 52:215–229.
Hauspie R, Chrzastek-Spruch H, Verleyen G, Kozlowska M,
Suzsanne C. 1994. Determinates of growth in body length
from birth to 6 years of age: a longitudinal study of Dublin
children. Int J Anthropol 9:202.
Herring D, Saunders S, Boyce G. 1994. Bones and the burial
registers: infant mortality in a 19th century cemetery from
Upper Canada. C Northeast Hist Arch J 20:54–70.
Hobbs C, Kigguridu M. 1992. A global analysis of life expectancy and infant mortality. Canada: Carleton University
Press.
Hoppa R, Vaupel J. 2002. The Rostock manifesto for paleodemography: the way from age to stage. In: Hoppa R, Vaupel J,
editors. Paleodemography: age distributions from skeletal
samples. Cambridge: Cambridge University Press. p 1–8.
Humphrey L, Dean M, Jeffries T. 2005. Identification of the neonatal line using LA-ICP-MS. Am J Phys Anthropol Suppl.
40:119–120.
Jeanty P, Dramaix-Wilmet D, van Kerkem J, Schwers J. 1982.
Ultrasonic evaluation of fetal limb growth. Part I. Radiology
143:751–754.
Jeanty P, Kirkpatrick C, Dramaix-Wilmet D, Struyven J. 1981.
Ultrasonic evaluation of fetal limb growth: Part II. Radiology
140:165–168.
Jeanty P, Romaro R. 1984. Obstetrical ultrasound. London:
McGraw Hill.
Johnston FE, Zimmer LO. 1989. Assessment of growth and age
in the immature skeleton. In: Iscan MY, Kennedy KAR, editors: Reconstruction of life from the skeleton. New York: Alan
R. Liss. p 11–21.
Jones E, Ubelaker D. 2001. Demographic analysis of the Voegtly
cemetery sample, Pittsburgh, Pennsylvania. Am J Phys
Anthropol 32:86.
Knight B. 1996. Forensic pathology, 2nd ed. London: Arnold.
Knodel J, Kintner H. 1977. The impact of breast feeding patterns on the biometric analysis of infant mortality. Demography 14:391–409.
Konigsberg L, Frankenberg S, Walker R. 1997. Regress what on
what? Palaeodemographic age estimation as a calibration
problem. In: Paine RR, editor. Integrating archaeological
demography: Multidisciplinary approaches to prehistoric populations. Carbondale: Center from Archaeological Investigations, Southern Illinois University. p 64–88.
Konigsberg LW, Frankenberg SR. 1992. Estimation of age structure in anthropological demography. Am J Phys Anthropol
89:235–256.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
INFANT MORTALITY IN MEDIEVAL ENGLAND
Lampl M, Jeanty P. 2003. Timing is everything: a reconsideration of fetal growth velocity patterns identifies the importance
of individual and sex differences. Am J Hum Bio 15:667–680.
Landers J. 1990. Age patterns of mortality in London during
the ‘long eighteenth century’: a test of the ‘high potential’
model of metropolitan mortality. Soc Hist Med 3:27–60.
Larsen CS. 1997. Bioarcheology: Interpreting behavior from the
human skeleton. Cambridge: Cambridge University Press.
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:
211–223.
Lewis M. 2007. The bioarchaeology of children. Perspectives
from biological and forensic anthropology. Cambridge: Cambridge University Press.
Loudon I. 1992. Death in childbirth: An international study of
maternal care and maternal mortality 1800–1950. Oxford:
Clarendon Press.
Lucy D, Aykroyd R, Pollard A, Solheim T. 1996. A Bayesian
approach to adult human age estimation from dental observations by Johansson’s age changes. J For Sci 41:189–194.
Malhotra KC. 1990. Changing patterns of disease in India with
special reference to childhood mortality. In: Swedlund AC,
Armelagos GJ, editors. Disease in populations in transition:
Anthropological and epidemiological perspectives. New York:
Bergin and Garvey.
Mays S. 1993. Infanticide in Roman Britain. Antiquity 67:883–
888.
Mays S. 2003. Comment on ‘‘A Bayesian approach to aging perinatal skeletal material from archaeological sites: implications
for the evidence for infanticide in Roman Britain’’ by Gowland
RL, Chamberlain AT. J Arch Sci 30:1695–1700.
Milne G, Richards J. 1992. Two Anglo-Saxon buildings and associated finds. York: York University Archaeological Publications 9.
Milner G, Humpf D, Harpending H. 1989. Pattern matching of
age-at-death distributions in paleodemographic analysis. Am
J Phys Anthropol 80:49–58.
Molleson T, Cox M. 1993. The Spitalfields Project, volume II–
The Middling Sort. York: Council for British Archaeology,
Research Report 86.
Murray C, Frenk J. 2002. Summary measures of population
health in the context of the WHO framework for health system performance assessment. In: Murray CJ, Salomon JA,
Mathers CD, Lopex AD, editors. Summary measures of population health. Concepts, ethics, measurement and application.
Geneva: World Health Organization, p 1–12.
Orme N. 2001. Medieval children. New Haven: Yale University
Press.
Paine R. 2000. If a population crashes in prehistory, and there
is no palaeodemographer there to hear it, does it make a
sound? Am J Phys Anthropol 112:181–190.
Paine R, Harpending H. 1996. Assessing the reliability of paleodemographic fertility estimators in stimulated skeletal distributions. Am J Phys Anthropol 101:151–159.
Palliser DM. 1980. Location and history. In: Magilton J, editor.
The archaeology of York: The medieval walled city north-east
of the Ouse. London: Council for British Archaeology. p 2–14.
Panhuysen R. 1999. Child mortality in early medieval Maastricht: missing children? J Paleopathol 11:94.
Ribot I, Roberts CA. 1996. A study of non-specific stress indicators and skeletal growth in two mediaeval subadult populations. J Arch Sci 23:67–79.
129
Robertson J. 1996. Reckoning with London: interpreting the
bills of mortality before John Graunt. Urban Hist 23:325–350.
Samworth RJ, and Gowland RL. 2006. Estimation of adult skeletal age-at-death: statistical assumptions and applications.
Int J Osteoarch 16:1–15.
Sarnat B, and Schour I. 1941. Enamel hypoplasia (chronological
enamel aplasia) in relation to systemic disease: a chronologic,
morphologic and etiologic classification. J Am Dent Ass
28:1989–2000.
Sattenspiel L, Harpending H. 1983. Stable populations and skeletal age. Am Antiquity 48:489–498.
Saunders SR. 2000. Subadult skeletons and growth-related
studies. In: Katzenberg M, Saunders S, editors. Biological
anthropology of the human skeleton. New York: Wiley-Liss.
p 1–20.
Saunders SR, Barrans L. 1999. What can be done about the
infant category in skeletal samples? In: Hoppa RD, Fitzgerald
CM, editors. Human growth in the past: studies from bones
and teeth. Cambridge: Cambridge University Press. p 183–
209.
Saunders SR, Herring DA, Boyce G. 1995. Can skeletal samples
accurately represent the living populations they come from?
The St. Thomas’ cemetery site, Belleville, Ontario. In: Grauer
AL, editor. Bodies of evidence: reconstructing history through
skeletal analysis. New York: Wiley-Liss. p 69–89.
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 Human Biol 7:257–265.
Scott S, Duncan C. 1998. Human demography and disease.
Cambridge: Cambridge University Press.
Scott S, Duncan CJ. 1999. Malnutrition, pregnancy and infant
mortality: a biometric model. J Interdis Hist 30:37–60.
Siegel S. 1956. Nonparametric statistics for the behavioral sciences. London: McGraw-Hill.
Smith P, Avishai G. 2005. The use of dental criteria for estimating postnatal survival in skeletal remains of infants. J Arch
Sci 32:83–89.
Smith P, Kahila G. 1992. Identification of infanticide in archaeological sites: a case study from the Late Roman-Early Byzantine Periods at Ashkelon, Israel. J Arch Sci 19:667–675.
Stockwell E. 1993. Infant mortality. In: Kiple K, editor. The
Cambridge world history of human disease. Cambridge: Cambridge University Press. p 224–229.
Stroud G, Kemp RL, editors. 1993. Cemeteries of the Church
and Priory of St Andrew, Fishergate. The archaeology of York,
Vol. 12: The medieval cemeteries. York: Current British
Archaeology.
Sundick RI. 1978. Human skeletal growth and age determination. Homo 29:228–249.
Vögele JP. 1994. Urban infant mortality in Imperial Germany.
Soc Hist Med 7:401–426.
Wear A. 1992. Making sense of health and the environment in
early modern England. In: Andrew W, editor. Medicine in society. Cambridge: Cambridge University Press. p 119–147.
Whittaker D, MacDonald D. 1989. A colour atlas of forensic dentistry. London: Wolfe.
Williams N, Galley C. 1995. Urban-rural differentials in infant
mortality in Victorian England. Popul Studies 49:401–420.
Wrigley E. 1977. Births and baptisms: the use of Anglican baptism registers as a source of information about numbers of
births in England before the beginning of civil registration.
Popul Studies 31:281–312.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
Документ
Категория
Без категории
Просмотров
0
Размер файла
280 Кб
Теги
infant, site, medieval, england, 850ц1859, brief, contrasting, post, precarious, live, mortality
1/--страниц
Пожаловаться на содержимое документа