Deciduous enamel 3D microwear texture analysis as an indicator of childhood diet in medieval Canterbury, England Patrick Mahoneya, Christopher W. Schmidtb, Chris Detera, Ashley Remyb, Philip Slavinc, Sarah E. Johnsa, Justyna J. Miszkiewiczd, Pia Nystrome.
aSchool of Anthropology and Conservation, University of Kent, UK.
bDepartment of Anthropology, University of Indianapolis, USA.
cSchool of History, University of Kent, UK
dDepartment of Medicine, Imperial College London, UK.
Abstract This study conducted the first three dimensional microwear texture analysis of human deciduous teeth to reconstruct the physical properties of medieval childhood diet (age 1-8yrs) at St Gregory’s Priory and Cemetery (11th to 16th century AD) in Canterbury, England. Occlusal texture complexity surfaces of maxillary molars from juvenile skeletons (n=44) were examined to assess dietary hardness. Anisotropy values were calculated to reconstruct dietary toughness, as well as jaw movements during chewing. Evidence of weaning was sought, and variation in the physical properties of food was assessed against age and socio-economic status. Results indicate that weaning had already commenced in the youngest children. Diet became tougher from four years of age, and harder from age six. Variation in microwear texture surfaces was related to historical textual evidence that refers to lifestyle developments for these age groups. Diet did not vary with socio-economic status, which differs to previously reported patterns for adults. We conclude, microwear texture analyses can provide a non-destructive tool for revealing subtle aspects of childhood diet in the past.
Dental microwear; medieval childhood diet.
Human diet during the 11th to 16th century in medieval England is best understood for adults, higher status families, or ‘closed-communities’ such as monastic settlements (Dyer, 2000: 83; Slavin, 2012: 8; Woolgar, 2010). Knowledge of childhood diet during this period is generally more limited because it was not a focus for medieval writers. Although limited, there is some historical textual evidence that provides weaning and other childrearing advice related to food consumption (Fildes, 1986: 213, 1988: 76). A few isotopic studies have also reported dietary weaning age and subsequent protein consumption for medieval village and urban centres in the north of England (Burt, 2013, 2015; Fuller et al., 2003; Mays et al., 2002; Richards et al., 2002), but not for the south-east. Neither is anything known about the physical properties (hardness, toughness) of medieval childhood diet.
Here, we conduct the first intra-specific three dimensional (3D) dental microwear texture analysis (DMTA) of human deciduous teeth to reconstruct the physical properties of childhood diet in medieval Canterbury, south-east England (Fig.1). DMTA is a non-destructive methodology that provides evidence of the hardness and toughness of foods eaten by an individual (Scott et al. 2005, 2006; Ungar et al., 2003) in the days and weeks preceding death (Grine, 1986). For example, dietary hardness and toughness has been reconstructed from DMTA of permanent tooth enamel for archaeological samples of hunter-gatherers, fossil hominins, and Neanderthals (El Zaatari et al., 2011, El Zaatari and Hublin, 2014; Schmidt et al., in press; Ungar et al., 2008a, 2010). However, few studies have examined microwear surfaces of deciduous enamel (e.g., Bullington, 1991). Our study is the first to apply the 3D methodology to human deciduous teeth.
1.1 Childhood diet in Medieval England
Physicians in sixteenth century Europe advised the introduction of mixed-feeding (gradual introduction of non-milk foods leading to a relative decrease in the contribution of breast milk to total diet: Humphrey, 2014) between seven to nine months of age (Fildes, 1986: 245). Historical records from this period indicate that a child was finally weaned (removal of breast milk) between 12-18 months (Fildes, 1986: her Table 15.3-4; 1995: her Table 4.7). This latter age range is compatible with isotopic evidence from the medieval village of Wharram Percy and the urban Fishergate House cemetery in the north of England, which suggests weaning occurred in the second year after birth (Burt, 2013, 2015; Fuller et al., 2003; Mays et al., 2002; Richards et al., 2002).
Pap (flour, milk, egg yolk) or panada (bread in broth with butter or oil) were popular supplementary foods during mixed-feeding (Fildes, 1986: 213; Orme, 2003: 71). Insights into early childhood diet after mixed-feeding have been gained from historical textual accounts. Grain products were an important component of medieval diet (Slavin, 2012: 169; Stone, 2006:11), and bread with butter, porridge, and gruel, were typical early childhood foods (Orme, 2003: 71-72). However, little is known about dietary variation with age. Isotopic evidence from Wharram Percy indicates that children may have consumed a post-weaning diet that that was lower in protein compared to older individuals (Richards et al., 2002).
Socio-economic status could determine the quality, variety, and type of foods consumed by adults (e.g., Dyer, 2006: 201-9; Powell et al., 2001: 298; Woolgar, 2006: 196; Woolgar et al., 2006: 270). Outside of periods of religious observance (primarily Advent and Lent) wealthier lay households and monastic communities regularly consumed meat, but other than pork, it contributed less to the peasant diet (DeWitte and Slavin, 2013; Dyer, 2000: 84-86; Powell et al., 2001: 308). Higher social strata preferred white bread made from wheat, while those of lower socio-economic status usually consumed coarser whole grain bread (Campbell, 2010; Stone, 2006: 17; Slavin, 2012: 180).
It is unclear if the relationship between adult status and food consumption extends to children from this period (Burt, 2013). In medieval York, lower status children consumed higher status and more expensive foods after weaning (Burt, 2015). Furthermore, a study of gross dental wear on deciduous teeth from medieval sites in the south of England, including Canterbury, reported no differences between higher and lower status burials of similarly aged children (Dawson and Robson Brown, 2013). Thus, the relationship between food consumption and status for children in this period might be more complex than that reported for adults.
1.2. Study Aims
This study conducts the first intra-specific microwear texture analysis of human deciduous teeth to reconstruct the physical properties of childhood diet in Medieval Canterbury (Fig. 1). All dental samples were from human juvenile skeletons (n=44) aged between one to eight years of age, which were recovered during excavation of St Gregory’s priory and cemetery (11th to 15th Century AD) in Canterbury (Hicks and Hicks, 2001). The site is unique in south-east England as it contained a large number of well-preserved juveniles. It has two burial areas, a priory and a cemetery, which correspond with higher and lower socio-economic status respectively (see section three).
The study aims are, 1) to search for microwear evidence of dietary weaning in the youngest children. 2) Determine if variation in the physical properties of diet correlates with age. 3) Compare microwear from those buried in the higher status priory to those buried in the lower status cemetery. Prior to these analyses, we conduct a preliminary experimental study to explore microwear texture formation processes on human deciduous enamel compared to permanent enamel.
1.3. Dental microwear texture analysis
Microscopic wear in the form of scratches and pits is laid down on the occlusal surface of tooth enamel as hard particles are sheared between or compressed into opposing crowns as the jaw moves through the chewing cycle (Gordon, 1982). Food contaminated by grit that is harder than enamel, such as quartz inclusions, is one microwear causal agent (Lucas et al., 2013; Peters, 1982; Teaford and Lytle, 1996). Two dimensional (2D) dental microwear analyses have been used since the 1950’s to explore jaw movements of extinct mammals and modern humans (Butler, 1952; Dahlberg, 1960; Mills, 1955). Subsequent 2D studies described microwear patterns by their frequency, size, and orientation in extant and fossil mammals (Grine, 1981; Puech, 1979; Walker, 1976; Walker et al., 1978) leading to a range of quantitative studies that sought to infer aspects of diet in past human populations (e.g., Mahoney, 2006; Pastor, 1993; Schmidt, 2001; Teaford et al., 2001). Methodological developments led to DMTA, the 3D characterization of microwear surfaces (Scott et al. 2006). This automated quantification of microwear in three dimensions minimizes inter-observer measurement error (Grine et al., 2002), and thus holds great potential for the future of dietary reconstruction in an archaeological context.
Dental microwear texture analysis is based upon the principle that an enamel surface can look different when observed at different scales. A surface may appear smooth when observed at a coarse scale but can appear rough at a finer scale. The texture of an enamel surface can be quantified in three dimensions by combining white-light confocal profilometry with scale-sensitive fractal analysis (Scott et al. 2005, 2006; Ungar et al., 2003). In this study we focus upon two texture variables that have been previously related to dietary hardness and toughness in extant primates:
Area-scale fractal complexity(Asfc) (Scott et al., 2005). Values for complexity measure changes in surface topography across different scales. Enamel with pits and scratches of different sizes superimposed onto each other, or a surface that is heavily pitted, would typically have higher complexity values (Ungar et al., 2008b, 2010). Consumption of hard abrasive foods, which are ‘crushed’ between opposing enamel surfaces, and correlated with frequent dental pits in 2D microwear analyses (Teaford, 1985; Teaford and Walker, 1984), tends to produce relatively higher Asfc values (and lower levels of anisotropy) in some primate hard seed and hard fruit eaters (Scott et al., 2005, 2006, 2012). Thus, a tooth surface that is dominated by dental pits with a high Asfc value has been used to infer a hard and abrasive diet (Scott et al., 2012; Ungar et al., 2010).
Exact proportion length-scale anisotropy (epLsar) (Scott et al., 2005). Values for the anisotropy of microwear texture surfaces measure the orientation of surface features. Enamel dominated by scratches all orientated in the same direction produces a high anisotropy value (Ungar et al., 2008a). Low anisotropy values indicate low similarity in wear feature orientation. Tougher foods which are ‘sheared’ between opposing enamel surfaces, and correlated with frequent dental scratches in 2D microwear analyses (Teaford, 1993; Teaford and Walker, 1984), can produce comparatively higher epLsar values (and lower levels of Asfc) in some primate species that consume leaves, stems and other tough fibrous foods (Scott et al., 2005, 2006, 2012). Therefore, enamel covered with scratches mainly orientated in the same direction with a high epLsar value has been used to infer consumption of tough abrasive foods (Scott et al., 2012; Ungar et al., 2010). Jaw movement also has been reconstructed from dental scratches (Butler, 1952; Gordon, 1982; Scott et al., 2006; Young and Robson, 1987), whereby high epLsar values indicate more consistent rather than varied jaw movements during chewing (Ungar et al., 2010).
While texture values within a species are variable, and texture surfaces for harder or tougher diets will often overlap (Strait et al., 2013), the key correlations between microwear and the physical properties of a diet established in the 1980s (Teaford, 1985; Teaford and Oyen, 1989; Teaford and Walker, 1984), have been confirmed more recently in studies of texture surfaces from mammals, and in experimental studies (e.g., Schubert et al., 2010; Schultz, 2013, Xia et al., 2015; Hua et al., 2015). Thus, DMTA distinguishes between extant primates of known diet, and these correlations provide a base-line from which to infer diet in historic and pre-historic populations.
Potential sources of deciduous microwear texture variation in Medieval Canterbury.
Breast feeding will produce no microwear and the introduction of abrasive foods should produce tooth wear. After weaning, flour prepared using traditional milling methods could introduce hard abrasive grit into cereal foods, which has been identified as a source of microwear in 2D studies (e.g., Teaford and Lytle, 1996). In medieval Canterbury, cereal foods such as these came from regional farmlands, demesnes, and local grain traders (Campbell, 2010; Slavin, 2012: 52-55, 2014). During the Middle Ages, grain was ground and prepared for consumption by local mills in Canterbury, one of which was owned by St. Gregory’s priory (Hastead, 1800; Somner, 1703). Mills in medieval Kent often used limestone and sandstones querns for milling (Farmer, 1992; Keller, 1989), which can introduce a residue of grit into foods (Teaford and Lytle, 1996).
Consumption of meat can alter microwear texture surfaces (El Zaatari, 2010). 19th century Fuegian hunter-gatherers with a diet that consisted mainly of meat had a lower mean Asfc but higher epLsar value, relative to other hunter-gatherer populations (El Zaatari, 2010). Chewing tough meat that contained some abrasives would require repetitive shearing motions of the jaw, leading to many scratches orientated in the same direction. The consumption of meat in medieval England varied by status amongst adults (above). If childhood status, or age, also determined access to meat, then this might contribute variation to epLsar values amongst the Canterbury children.
Beyond hard and abrasive, and tough foods, there are several other potential microwear formation processes that should be considered when interpreting deciduous enamel textures. First, bite force potential will differ significantly between younger and older children, as the muscles of mastication gain size and strength (Kamegai et al., 2005). As such, more force exerted during chewing would provide more opportunity for hard particles to be driven into enamel as microwear accumulates for the first time. Thus, variation in microwear texture surfaces between children of different ages might relate in part to differences in bite force. Lateral movement of the mandible will also increase with age, as the mandible increases in size. Greater lateral movement, as the mandible moves through the chewing cycle, might produce longer scratches, though this would not necessarily alter an anisotropy value.
‘Teething’, and the use of pacifiers by young children, could contribute microwear that was unrelated to diet. Dental eruption in humans typically commences around the sixth post-natal month as deciduous central incisors emerge through the gum line (Hillson, 2014: his Table 4). The second molar is the last deciduous tooth to erupt, usually towards the start of the third post-natal year (Hillson, 2014: his Table 4). Infants biting on pacifiers might scratch the enamel surface. For example, in medieval England, a child might be given a piece of coral during teething (Hanawalt, 1993:52). This potential source of microwear is more likely in younger infants, and more likely to accumulate on early erupting incisors. Selecting later erupting teeth can reduce the potential for pacifiers to obscure a diet-microwear relationship.
2. Preliminary experimental study
The efficacy of DMTA as an indicator of the physical properties of diet has been demonstrated numerous times using adult permanent teeth (see Section 1.3). Its value, however, has not been demonstrated to the same extent on deciduous teeth. Would a single microwear formation process produce similar microwear texture values for both deciduous and permanent enamel? Or instead, would these two enamel types differ in an unexpected way when exposed to the same formation process. For example, deciduous enamel is more porous and relatively softer than permanent enamel (e.g., Wilson and Beynon, 1989). To investigate microwear formation processes on deciduous relative to permanent enamel we undertook an experimental study before we examined microwear texture surfaces of juveniles from St Gregory’s priory and cemetery.
Complexity and anisotropy values were calculated for thirteen deciduous incisors. The deciduous teeth were experimentally abraded, complexity and anisotropy values were re-calculated, and compared to those from before the experiment. We repeated the experiment on six permanent premolars, and compared the results to the deciduous teeth.
Thirteen deciduous mandibular incisors were donated by former students to the Indiana Prehistory Laboratory, University of Indianapolis, knowing these teeth would be used in wear experiments. These teeth were included in the experimental study because the labial surface of each tooth showed no signs of gross dental wear. The six permanent maxillary premolars were from an early 20th century cemetery population in Indiana. The premolars were selected because the mesial surface of each tooth was unworn. There was no reason to suspect that the enamel microstructure of the dental samples used in the experimental study would influence microwear formation processes in a way that would differ to the dental samples from Canterbury.
2.2 Experimental procedures
Microwear preparation and analytical procedures are described in section four. Before each tooth was experimentally abraded, a target area was identified, and the complexity and anisotropy of that area was recorded. The same target area was then abraded experimentally and the texture values were recorded again. Thus, we produced “before” and “after” experimental data sets.
We tested the null hypothesis that there would be no difference between the complexity and anisotropy post-experimental abrasion values from deciduous enamel, when compared to permanent enamel, using a Mann Whitney U test.
2.3 Experimental results
Table 1 shows the experimentally induced microwear texture values. Mean Asfc increased by 5.25 for deciduous teeth, and by 4.64 for permanent teeth, from before to after the experimental abrasion. Mean anisotropy increased by 0.0053 for deciduous teeth, and by 0.0062 for permanent teeth, from before to afterwards. The post-experimental complexity and abrasion values did not differ significantly when compared between deciduous and permanent teeth (p= 0.844, p=0.116, respectively). Therefore, the null hypothesis was retained.
Mean experimentally induced microwear texture values.
2.4 Discussion of experimental results
The experimentally created wear was statistically indistinguishable when compared between deciduous and permanent teeth, indicating that microwear forms in a similar way when these two enamel types are subjected to the same force applied in the same direction. However, there were slight differences in the mean values from the two enamel types. Deciduous enamel accumulated a slightly more complex surface with fewer similarly orientated scratches during the course of the experiment, relative to permanent enamel. We observed that incisor enamel surfaces touching the abrasive paper were more curved compared to premolars. So, these slight differences in the degree to which the microwear values changed could be an artefact of this experiment, as force would have been applied to a smaller area on the incisors compared to the premolars. Future studies can explore this in more detail, to determine if facet size plays a key role in microwear formation. Overall, it is clear that deciduous and permanent enamel produce similar microwear texture surfaces when subjected to the same force applied in the same direction.
2.5 Limitations of experimental study
Results from the experimental study underscore the efficacy of the DMTA variables employed here. However, other DMTA variables commonly used in studies of dietary inference were not examined. Scale of maximum complexity, textural fill volume, and heterogeneity are yet to be analysed. Moreover, our study only included experimental wear generated in a single direction with a single force. It may be possible that a threshold exists whereby extreme force, or the direction of a force, can distinguish between adult and deciduous microwear texture surfaces.