The archaeological site is located about 300 meters north of present day Canterbury Cathedral just outside the medieval city wall (Fig. 1). It was in use from the mid-11th to early 16th century, and was excavated between 1988-1991 (Hicks and Hicks, 2001: 1-146). The priory was founded by the Archbishop of Canterbury, Lanfranc, in AD 1084 (Sparks, 2001: 371). Originally it was served by priests, and subsequently by Augustine cannons, who cared for the sick at nearby St John’s hospital and provided free burial for the poor in the cemetery (Brent, 1897; Duncombe, 1785; Somner, 1703; Sparks, 2001: 371).
All skeletons were previously excavated (Anderson and Andrews, 2001: 338-370). A total of 91 burials were recovered from inside and around the priory, which included a male with a chalice and a gold-embroidered monastic-like garment suggesting this was a burial location for clergy (Anderson and Andrews, 2001). However, the presence of children and adult females within the priory indicates that this was not a ‘closed’ monastic community. Instead, these were members of wealthier families (Hicks and Hicks, 2001), who paid for the prestigious burial location, which was a popular way of displaying socio-economic status for wealthy lay people in this period (Daniell, 1997: 96-97).
The cemetery was established just before the priory (Sparks, 1988: 31), and a total of 1342 skeletons were recovered during excavation. Historical textual records indicate that the cemetery served poorer families from local parishes, people who could not afford burial fees, and patients from nearby St. John’s hospital (Brent, 1879; Somner, 1703). It was in constant use until a few years after the priory was dissolved in the 16th century (Sparks, 1988: 32, 2001: 376).
4. Materials and methods
4.1 Sample selection
Microwear values were produced for deciduous maxillary first and second molars from 44 juvenile skeletons aged one to eight years. These skeletons were selected because they retained the skeletal elements needed to estimate age-at-death. We focused upon maxillary molars because they have thicker enamel (Mahoney, 2013), that (usually) has relatively less gross wear compared to their mandibular isomeres. This was important as dentin microwear was not a focus of the present study. The molars selected were also suitable for cleaning and casting for microwear. Microwear values were subdivided into age groups, which were created from skeletal age-at-death (see below) and the timing of dental eruption (Table 2).
Two juvenile skeletons from the priory dated to the earliest Lanfranc period (11th century). The microwear Asfc and epLsar values for these individuals were within the range of microwear values for juveniles from the priory which dated to the 14th-16th centuries. Following this, microwear values were treated as one time period for subsequent analyses. The cemetery burials were not sub-divided by century during excavation.
Deciduous microwear samples
Age in yrs
1Udm1 = maxillary first molar. Udm2 = maxillary second molar
4.2 Preparation and microwear texture data
All teeth were prepared in the Human Osteology Research Lab, University of Kent, using standard methods (e.g., Mahoney, 2006; Nystrom et al., 2004; Schmidt, 2001). The occlusal surface of each tooth was cleaned using 95% ethanol and cotton wool. Impressions of phase II facets were taken using a rubber-based addition-curing silicone (Colténe-Whaledent Lightbody President Jet®). The first impression was discarded and a second impression was taken and used to create the cast. The dental impression was set into dental putty (Colténe-Whaledent, President Putty®). An epoxy resin and hardner (Buehler EpoxiCure®) was poured into the impression to produce a cast of the occlusal surface.
Microwear texture data were produced in the Indiana Prehistory Lab, University of Indianapolis. Resin dental casts were examined using a Sensofar® White Light Confocal Profiler at a magnification of 100x. The microscope collected data from four contiguous areas totalling 276 x 204μm2. After digitally stitching the original four areas together, the final study area was 242 x181μm2. Data came from Phase II wear facets (usually facet 9). Data cloud manipulation was undertaken using Sensoscan® software, where the data were levelled and non-microwear entities (primarily any remaining dirt) were removed. Analysis of the data cloud required the use of Sfrax® and Toothfrax®, which are scale-sensitive fractal analysis programs customized for dental microwear texture analysis. Microwear variables Asfc and epLsar were recorded as scale-dependent relative values (Scott et al., 2006).
4.3. Estimating age-at-death
For the one child aged 1 year, we estimated age-at-death using enamel formation times (Mahoney, 2011). For the rest of the children we used a combination of enamel formation times (Moorrees et al., 1963a,b), timing of dental eruption (Schour and Massler, 1941; Al-Qahtani et al., 2010), long bone length (Hoppa, 1992; Scheuer et. al., 1980), and fusion of cervical vertebra (Scheuer and Black, 2000).
The distribution of each microwear variable for each childhood age group (1-2, 2.1-4, 4.1-6, 6.1-8yrs) was checked with a one sample Kolmogorov–Smirnov test and did not differ significantly from a normal curve. However, sample sizes were unequal. Thus, microwear was compared between the four age groups using a non-parametric Kruskal-Wallis test. Multiple post-hoc pair-wise comparisons of the age groups were undertaken using a Tamhane-2 test. Microwear was compared between the two status groups using a Mann Whitney U test.
Microwear descriptive statistics are summarized in Table 3. Figure 2 illustrates microwear texture surfaces. A Kruskal-Wallis test revealed that the complexity of microwear surfaces differed significantly between the four childhood age groups (H=9.037, p=0.029), but anisotropy did not (H=6.572, p=0.087). Post-hoc tests of pair-wise mean differences using the T2 statistic indicates that children aged 4.1-6 years of age had a significantly lower mean complexity value compared to younger (aged 2.1-4yrs; p=0.017) or older children (6.1-8yrs; p=0.011). Microwear did not differ significantly between higher and lower status children within the age group 2.1-4yrs, or within the age group 4.1-6 yrs.
Deciduous microwear mean values
1Lower = cemetery; higher = priory.
6.1. Weaning amongst one to two year olds.
Microwear was present on molars from seven children in this age group (Table 3). The presence of microwear suggests that mixed-feeding, at least, had commenced. On average, their molar surfaces had a low Asfc and higher epLsar value, which is usually associated with the consumption of tougher foods amongst extant primates (section 1.3). Therefore, at first glance, there appears to be discrepancy between the microwear and the type of diet consumed by the youngest children, as textual accounts indicate that a soft and limited range of foods, such as pap or panda (Orme, 2003:71) would have been consumed. However, a high epLsar can also be indicator of jaw movements during chewing (Ungar et al., 2010). On average, the one to year olds had the most anisotropic texture surfaces compared to all other childhood groups. Thus, the orientation of their microwear was the most organized, reflecting the fewest changes in jaw direction during chewing. This makes sense, when viewed alongside the limited range of foods consumed by this age group. The low mean complexity value for the youngest children, relative to the 2.1-4yr olds, more likely represents the consumption of soft foods (Scott et al., 2012). Flour in pap or panda, contaminated during the milling process is one potential source of the abrasive particles that caused microwear for this age group.
The youngest infant with microwear was aged 1 year, which implies that mixed-feeding might have commenced slightly earlier for some children in Canterbury, compared to children from the contemporary Fishergate House cemetery in the north of England where breast milk continued to be a significant part of the diet until age 18 months (Burt, 2013, 2015). However, there is no change in microwear throughout the course of the year, which might have indicated a transition from mixed-feeding to fully weaned. Instead, a child aged 1.25 years had a similar complexity value compared to another aged 1.75 years (Asfc= 1.71 and 1.69 respectively). This may simply reflect a gradual change in feeding practices that is not detectable from microwear. Alternatively, breast-feeding might have been completely removed from the infant diet at or around the start of the second year after birth. If this was the case for the Canterbury children then their weaning age would lie within the lower end of the age-range recommended for weaning in texts from the period (Fildes, 1995: 115). It would also lie within the lowermost end of the weaning age-range indicated by isotopic studies at contemporary Wharram Percy in the north of England, where breast-feeding ceased between one to two years of age (Mays et al., 2002).
6.2. Variation in diet with age
Dental microwear texture analysis results suggest that the physical properties of diet for children in medieval Canterbury varied from one age group to the next.
6.2.1. Two to four years of age.
Children aged two to four display an increased mean complexity of enamel surfaces combined with a lower mean anisotropy, relative to one to two year olds. When this combination of microwear features are compared with the base-line texture surfaces from extant primates (section 1.3), it implies that the Canterbury children in this age group consumed a range of foods that included relatively harder and more abrasive items. These texture surfaces might be expected, as their diet was probably no longer focused upon just soft infant foods like pap and panda. A more varied diet is also suggested by the lowered mean anisotropy value, indicating that jaw movements were more disorganized during chewing. Increased bite force relative to the infants (Kamegai et al., 2005) might be a factor here as well, driving hard particles deeper into the enamel surface leading to a higher complexity value.
6.2.2. Four to six years of age.
There was a significant change in the physical properties of diet amongst children in this age group. The four to six year olds had significantly less texture complexity than either younger (2.1-4yrs) or older (6.1-8yrs) children. The lowered complexity was matched by a higher mean anisotropy value, which approached significance when compared to the less anisotropic enamel from the 6.1-8 year olds. This combination of microwear features, lower Asfc and higher epLsar (section 1.3), implies that the diet of children in medieval Canterbury had altered, and now included tougher foods.
A change in diet between age four to six could relate in part to a period in which childhood routines started to change (Bailey et al., 2008; Hanawalt, 1977: 64). Greater mobility allowed children to accompany adults outside of their home and into the work place, paradoxically leading to more time spent in adult company (Flemming, 2001; Hanawalt 1977, 1988:158). More time in adult company may have given more access to adult dietary staples, such as a meat or vegetable pottage (e.g., Brears, 2008). A greater component of meat in the diet of the Canterbury children might explain the change in microwear (e.g., El-Zaatari, 2010), especially if this was a permanent supplement to early childhood foods.
Support for the idea that children in this age group accessed ‘tougher’ adult dietary staples, rather than returning to a soft diet similar to the infants, is provided by examining their bite force potential. Children in this age group would have exerted significantly more force during chewing compared to the one to two year olds (Kamegai et al., 2005). If the change in the microwear pattern of the four to six year olds occurred because they re-accessed a soft infant diet, whilst for example caring for a younger sibling (Hanawalt, 1988: 157), then you would expect the enamel of the older children to have a higher complexity value, as abrasive particles from the shared foods would have been driven deeper into their enamel. This idea is not supported by the mean microwear texture values, which show that the older children had a lower, not a higher mean Asfc value, relative to the infants. Neither does ‘teething’ nor a ‘sick-bed’ diet seem likely causal agents. All deciduous teeth would have erupted by around the age of 2.5 years, so pacifiers would not have contributed to the microwear of this age group, or to the preceding age group. A sick-bed diet would not necessarily contribute to the microwear of only this age group.
6.2.3. Six to eight years of age.
Children in this age group had the roughest texture surfaces with many pits and scratches of different sizes overlying each other. The scratches were the least orientated compared to all other childhood age groups, leading to the lowest epLsar value. The reduced range of complexity and anisotropy values for this group indicates that fewer children deviated away from the rough and disorganized wear features. If the tougher diet of the preceding age group marks the introduction of ‘adult foods’, then the increase in food hardness in the eldest children might indicate the addition of hard ‘adult’ foods. This idea is supported by historical textual accounts. From around the age of seven onwards children were treated increasingly like young adults and were given independent tasks outside of their home (Hanawalt, 1977, 1988: 158; Fleming, 2001: 64), including apprenticeships or employment as household servants (Bailey et al., 2008; Dunlop, 1912). It might be expected therefore, that this change in a child’s social network would provide reduced opportunity for a distinct childhood diet as they entered a new environment.
6.3. Childhood status and diet
Mean complexity and anisotropy values for children aged two to four years, or four to six years of age, did not vary consistently with status (Table 3). This finding lends support to the idea that the relationship between status and food consumption for medieval children might be more complex compared to adults (Burt, 2013, 2015; Dawson and Robson Brown, 2013).
This study conducted the first 3D intra-specific dental microwear texture analysis of childhood diet. We searched for evidence of dietary weaning, evaluated variation in the physical properties of diet against age, and compared higher with lower status children. Results indicate that mixed-feeding in Canterbury could commence by the end of a child’s first year. After weaning, and until the age of eight, there was no simple trajectory in the physical properties of the foods that were consumed in the weeks before death. Diet contained abrasives for all age groups. Texture surfaces indicated that, on average, the four to six year olds consumed a diet that included tough foods whilst the eldest children consumed the hardest diet. We related these changes in microwear texture surfaces to medieval textual records that refer to lifestyle developments for these age groups. Our study also lends support to the idea that the relationship between socio-economic status and diet for children in medieval England might not be as clear as it is for adults. We conclude that deciduous dental microwear texture analyses hold great potential for revealing very subtle changes to childhood diet in the past.
Research funded by a British Academy-Leverhulme Trust Research Grant (SG-121921) to PM and CWS. We thank two anonymous reviewers and an Associate Editor for comments that improved the manuscript.
Al-Qahtani, S.J., Hector, M.P., Liversidge, H.M., 2010. Brief Communication: The London atlas of human tooth development and eruption. Am. J. Phys. Anth. 142: 481–490.
Anderson, T., Andrews, J., 2001. The human remains. In: St Gregory’s Priory Northgate Canterbury Excavations 1988-1991. Canterbury Archaeological Trust LTD, Canterbury.
Bailey, B.G., Bernard, M.E., Carrier, G., Cherise Elliott, L., Langdon, J., Leishman, N, Mlynarz, M., Mykhed, O. and Sidders, L.C., 2008. Coming of age and the family in medieval. England. J. Fam. Hist. 33: 41-60.
Brears, P., 2008. Cooking and dining in medieval England. Prospect Books., London.
Brent, J., 1879. Canterbury in the olden time. Simpkin, Marshall and Co., London.
Bullington, J., 1991. Deciduous dental microwear of prehistoric juveniles from the lower Illinois valley. Am. J. Phys. Anth. 84: 59-73.
Burt, N.M., 2013. Stable isotope ratio analysis of breastfeeding and weaning practices of children from Medieval Fishergate House York. Am. J. Phys. Anth. 152: 407-16.
Burt, N.M., 2015. Individual dietary patterns during childhood: an archaeological application of a stable isotope micro-sampling method for tooth dentin. J. Arch. Sci. 53: 277-290.
Butler, P.M., 1952. The milk-molars of Perissodactyla with remarks on molar occlusion. Proc. Zoo. Soc. Lond. 121: 777-812.
Campbell, B.M.S., 2010. Agriculture in Kent in the high middle ages. In: Sweetinburgh S, (Ed.), Later medieval Kent 1220-1540. Boydell Press, Woodbridge. pp. 25-53.
Dahlberg, A. 1960. Microscopic studies of abrasion and attrition on tooth surfaces. J. Dent. Res. 39(4): 713-4.
Daniell, C., 1997. Death and burial in Medieval England 1066–1550. Routledge, London.
Dawson, H., Robson Brown, K., 2013. Exploring the relationship between dental wear and status in late medieval subadults from England. Am. J. Phys. Anth. 150: 433–441.
DeWitte, S., Slavin, P., 2013. Between famine and death. Physiological stress and dairy deficiency in England on the eve of the Black Death (1315-50): New evidence from paleoepidemiology and manorial accounts. J. Interdis. Hist. 44: 37-60.
Duncombe, J., 1785. The history and antiquities of the three archiepiscopal hospitals and other charitable foundations at and near Canterbury. Bibliotheca Topographica Britannica No XXX, London.
Dunlop, O.J., 1912. English apprenticeship and child labour: a history. T Fisher Unwin, London.
Dyer, C., 2000. Everyday life in Medieval England. Hambledon and London, London.
Dyer, C., 2006. Seasonal patterns in food consumption in the latter Middle Ages. In: Woolgar, C.M., Serjeantson, D., Waldron, T., (Eds), Food in medieval England. Oxford University Press, Oxford. pp. 201-14.
El-Zaatari, S., 2010. Occlusal microwear texture analysis and the diets of historical/prehistoric hunter-gatherers. Int. J. Osteoarchaeol. 20: 67–87.
El-Zaatari, S., Grine, F.E., Ungar, P.U., Hublin, J-J., 2011. Ecogeographic variation in Neanderthal dietary habits: Evidence from occlusal molar microwear texture analysis. J. Hum. Evol. 61:411-424
El-Zaatari, S., Hublin, J-J., 2014. Diet of upper Paleolithic modern humans: evidence from microwear texture analysis. Am. J. Phys. Anth. 153, 4: 570-81.
Farmer, D., 1992. Millstones for medieval manors. Agr. Hist. Rev. 40: 97-111.
Fildes, V.A., 1986. Breasts, bottles and babies: a history of infant feeding. Edinburgh University Press, Edinburgh.
Fildes, V.A., 1988. Wet nursing: a history from antiquity to the present. Basil Blackwell, New York.
Fildes, V.A., 1995. Chapter 4: The culture and biology of breastfeeding: an historical review of Western Europe. In: Stuart-Macadam, P., Dettwyler, K., (Eds), Breastfeeding: biocultural perspectives. Aldine De Gruyter, New York. pp. 101-126.
Flemming, P., 2001. Family and household in Medieval England. Palgrave Macmillan, Basingstoke.
Fuller, B.T., Richards, M.P., May, S.A., 2003. Carbon and Nitrogen isotope variations in tooth dentine serial sections from Wharren Percy. J. Arch. Sci. 30: 1673–1684.
Goldberg, G., 1992. Women, work, and life cycle in a medieval economy: Women in work and Yorkshire c. 1300-1520. Oxford University Press, Oxford.
Gordon, K.D., 1982. A study of microwear on chimpanzee molars: implications for dental microwear analysis. Am. J. Phys. Anth. 59: 195-215.