Raising girls and boys in early China: Stable isotope data reveal sex differences in weaning and childhood diets during the eastern Zhou era
Melanie J. Miller Yu Dong Kate Pechenkina Wenquan Fan Siân E. Halcrow
First published:06 March 2020
doi.org/10.1002/ajpa.240331 INTRODUCTION
Growth, maintenance, and reproduction are three imperative functions of a living organism. Resource availability during infancy and childhood shapes the energy allocation toward each of these functions and determines the organism's survival into adulthood, adult health status, as well as the reproductive success of the individual (Gluckman, Hanson, & Beedle, 2007; Gluckman, Hanson, & Buklijas, 2010). The increasing impact of life‐history theory on bioarchaeological research has brought greater attention to studying children and childhood (Beauchesne & Agarwal, 2018; Halcrow & Tayles, 2011; Lewis, 2007; Perry, 2005; Temple, 2019). Studies of diet, health, and disease have been central to understanding childhood in the past, and can reveal critical effects that the care for children has in societies over time and space, and for human evolution at large (Fuller, Richards, & Mays, 2003; Goodman & Armelagos, 1989; Gowland, 2015; Humphrey, 2010; Katzenberg, Herring, & Saunders, 1996; Lewis, 2007; Moffat & Prowse, 2018; Schurr, 1998; Tsutaya, 2017). Dietary studies provide insights to anthropological questions related to how food practices are entangled in the social–political processes within any particular culture and can provide unique insights into the socialization of individuals through food consumption (Chang, 1977; Hastorf, 2017; Ohnuki‐Tierney, 1993; Sterckx, 2005; Twiss, 2007; Weismantel, 1988). Although most archaeological studies of dietary practices using methods such as stable isotope analysis have traditionally focused on adulthood diet (primarily a function of methodological progression and theoretical orientation over time), increasingly bioarchaeologists are recognizing the importance of the childhood period for revealing the biocultural position that food uniquely plays in human existence.
In recent decades anthropological isotopic specialists have increasingly turned to human teeth, which potentially provide resolution of human diets on the level of months to years (Burt, 2015; Eerkens, Berget, & Bartelink, 2011; Miller, Agarwal, & Langebaek, 2018; Reitsema & Vercellotti, 2012; Sealy, Armstrong, & Schrire, 1995; Turner, Kingston, & Armelagos, 2010; Wright & Schwarcz, 1999). Through incremental sampling of specific teeth, which form during discrete periods of development, bioarchaeologists can target particular periods of early life to understand maternal diet, breastfeeding, and weaning patterns in infancy, and dietary shifts during childhood, adolescence and early adulthood (Beaumont, Gledhill, Lee‐Thorp, & Montgomery, 2013; Burt & Garvie‐Lok, 2013; Dal Martello et al., 2018; Fernández‐Crespo, Czermak, Lee‐Thorp, & Schulting, 2018; Halcrow et al., 2018; King et al., 2018). Comparing dental isotopic results to bone samples from the same individual allows us to study dietary patterns over a lifetime and provides detailed individual histories of ancient peoples (Sealy et al., 1995). The results of these analyses can provide unique insights into social processes that are often obscured from other archaeological data, such as the relationships between food access and variables such as sex, gender, age, social status, and biological processes such as how diet and nutrition influence growth, development, and disease.
This study examines the childhood diet of humans who survived into adulthood from two urban mortuary populations excavated from the Eastern Zhou (771–221 BC) city of Zhenghan (郑韩故城), located in the Central Plains region of China (Figure 1). Recent isotopic analysis of bone collagen from adult skeletons of Zhenghan documented significant differences in diets between adult females and males (Dong et al., 2017). Our research studies dietary patterns during early life through the incremental sampling of tooth dentin collagen using the stable isotopes of carbon and nitrogen. We targeted early‐developing permanent dentition, a first molar or canine tooth, to capture periods of dental development and therefore childhood diet, from birth up to ~14 years of age. We test whether these dietary differences between the sexes can be traced to the early phases of childhood and whether sex‐specific cultural practices pertaining to the timing of weaning, weaning foods, and childhood dietary patterns can be identified in early forming dental tissues.
Figure 1
Open in figure viewerPowerPoint
Map showing the location of the modern city of Xinzheng on China's Central Plains with the detailed location of the archaeological sites studied (Xiyasi and Changxinyuan in pink triangles) relative to the ancient Zhenghan city walls (in green segments) and Shuangji river (in blue)
1.1 Reconstructing individual dietary histories using stable isotope analysis
Breastfeeding and weaning are unique processes in humans because they reveal the deep entanglements of biology and culture (Cassidy & Tom, 2015; Maher, 1992; Moffat & Prowse, 2018; Stuart‐Macadam, 1995; Tomori, Palmquist, & Quinn, 2018). How babies are fed so that they can grow and develop healthily is critical to the continuity of a family, culture, and the survival of our species (Dettwyler, 1995, 2004; Humphrey, 2010). Breast milk is the primary dietary component for infants, providing rich nutrition and complex immunological support to the developing body (Andreas, Kampmann, & Mehring Le‐Doare, 2015; Hanson, 1998; Hanson et al., 2002; LaTuga, Stuebe, & Seed, 2014; Martin, 2018; Miller, 2018; Quinn, 2018). Over variable periods of time, babies are weaned so that they are not solely reliant on milk for their caloric and nutritional needs, with other foods introduced at different times during development, and finally cessation of breast milk consumption when the child is fully weaned. Furthermore, feeding children solid foods is biologically risky during early life; with their developing immune systems children are more prone to infection, some of which could be spread through poor food hygiene. The rapidly growing bodies of infants also need adequate caloric intake to ensure “normal” development of the body (such as reaching average height for a particular population). Variations in infant feeding practices (such as length of time breast milk is the exclusive food source) can have significant impacts on health outcomes (Dieterich, Felice, O'Sullivan, & Rasmussen, 2013; Miller, 2018; Moffat & Prowse, 2018; Victora et al., 2008, 2016). The timing and duration of the weaning process and the foods that are introduced to the child at different times, are heavily influenced by cultural beliefs, with potentially great differences between individuals, families, and cultural groups over time and space (Cassidy & Tom, 2015; Maher, 1992; Martin, 2018; Moffat & Prowse, 2018; Sobonya, 2018; Stuart‐Macadam, 1995; Veile & Kramer, 2018). Many of the practices related to feeding children are deeply tied to societal beliefs about how a body is properly nourished to grow and develop, and studying childhood dietary practices can provide significant insights into these aspects of culture. Therefore, studying dietary practices of children can indicate patterns where nutrition and health outcomes may be correlated, producing valuable information not only about the past societies we study, but serve as critical data with implications for modern populations.
Bioarchaeologists are able to contribute to our understanding of the patterns of human diet in ancient populations through methods including studies of skeletal chemistry, paleopathology, dental calculus composition, coprolite analysis, dental microwear, oral health, and more (Larsen, 2015). Bones and teeth record aspects of diet from the time periods they are formed and stable isotope analysis of these tissues can reveal dietary patterns from different periods of an individuals' life (DeNiro, 1985; Eerkens et al., 2011; Price, Schoeninger, & Armelagos, 1985; Sealy et al., 1995). Stable isotopes of carbon (δ13C) and nitrogen (δ15N) are now commonly used in archaeological studies because these elements provide important insights into human history and evolution, especially related to dietary patterns and paleoenvironmental reconstructions (Ambrose, 1993; DeNiro, 1985; Lee‐Thorp, 2008). Carbon and nitrogen stable isotope values relate to the foods people consumed during life, and are often preserved in skeletal tissues (DeNiro & Epstein, 1978; DeNiro & Schoeninger, 1983). Carbon provides information about the broad groups of plants that animals consumed during their life: modern C3 plants have δ13C values averaging around −23‰, while C4 plants have δ13C values averaging around −13‰ (DeNiro & Epstein, 1978; Ehleringer & Cerling, 2002; Farquhar, Ehleringer, & Hubick, 1989; Kohn, 2010; O'Leary, 1988). Nitrogen isotopic data relates to protein consumption and can indicate one's position within a local food web, with a 3–5‰ stepwise enrichment in nitrogen as one rises from the base of the food chain to top‐level predators (DeNiro & Schoeninger, 1983; Minagawa & Wada, 1984; Schoeninger & DeNiro, 1984). Nitrogen isotope values are also influenced by unique environmental conditions such as altitude and aridity, and nitrogenous inputs such as fertilizers (Ambrose, 1991; Bogaard, Heaton, Poulton, & Merbach, 2007; Szpak, 2014; Szpak, Millaire, White, & Longstaffe, 2012).
Bones and teeth are composed of two interwoven matrices, the organic collagen, and inorganic hydroxylapatite matrices, and each provides different isotopic data and reflects different aspects of diet. The organic collagen component has been found to preferentially incorporate dietary proteins, with amino acids routed directly from protein sources into the body's collagen matrix (Ambrose & Norr, 1993; Tieszen & Fagre, 1993). Dietary studies of hydroxylapatite (bone apatite, dental enamel) provide carbon isotope data (from carbonate incorporated into the structure) and this carbon has been found to be pulled from all carbon sources within the diet (so all major dietary components can contribute carbon to the inorganic fractions). Additionally, there are quantified offsets between the diet an organism consumes and the isotopic composition of bodily tissues, with collagen‐diet spacing of approximately +5‰ for δ13C (collagen δ13C value is +5‰ more positive than average diet) (Lee‐Thorp, Sealy, & van der Merwe, 1989; Sullivan & Krueger, 1981; Van Der Merwe & Vogel, 1978), and apatite‐diet spacing has been found to range from +9 to +12‰ for δ13C (Krueger & Sullivan, 1984; Lee‐Thorp et al., 1989). This study focuses on the organic collagen fraction of bone and tooth samples.
Babies exclusively consuming breast milk show a positive offset of +2–3‰ for δ15N and about +1‰ for δ13C compared with their mothers, as revealed in studies of fingernail samples from modern mother–baby pairs (Fogel, Tuross, & Owsely, 1989; Fuller, Fuller, Harris, & Hedges, 2006). Growing infants require dietary supplementation around the age of 6 months, as energetic requirements begin to outpace the nutritional content of breast milk (Kramer & Kakuma, 2002; Moffat & Prowse, 2018). As other foods are introduced and the baby is progressively weaned off breast milk the child's isotopic values will change, usually with a noted decline in δ15N, and at least a slight shift in δ13C (Eerkens et al., 2011; Fuller et al., 2003; Wright & Schwarcz, 1999), though other isotopic patterns related to a lack of breastfeeding and physiological stress have been noted in archaeological populations (Beaumont & Montgomery, 2016; King et al., 2018; Kwok, Garvie‐Lok, & Katzenberg, 2018). Previous bioarchaeological research has used patterned changes in δ15N (a drop of 2–4‰) and δ13C (decline by around 1‰) in tooth and bone samples to demarcate the completion of the weaning process and the shift to a diet without significant amounts of breast milk (Beaumont & Montgomery, 2016; Burt & Garvie‐Lok, 2013; Eerkens et al., 2011; Eerkens, de Voogt, Dupras, Francigny, & Greenwald, 2018). Therefore, we expect to see changes to the diet around 6 months of age with the introduction of new food sources, and likely expansion of dietary components as the child ages, with the underlying assumption that children eventually consume a diet that is very similar to the adults around them (Tsutaya, 2017).
Interpretations of estimated age when breast milk consumption is terminated are most successful in cases where childhood and adulthood diets were relatively homogeneous (i.e., where children are transitioned onto diets that appear very similar to adulthood diets), and where diets were usually isotopically constrained with very minimal mixing across C3, C4, terrestrial, and aquatic resources (Burt & Garvie‐Lok, 2013; Eerkens et al., 2018; Tsutaya, 2017). Identifying a specific age when weaning concludes can be difficult when individuals (in this case usually assumed as lactating mothers and their breastfeeding infant) eat across both C3 and C4 food groups (causing greater change in δ13C than the usual 1‰ expected decline) and where various protein sources may be introduced to children's diets (potentially confounding the drop in δ15N expected with the cessation of breast milk). Additionally, if children are weaned onto diets that are unique to the childhood period (and are different from the dietary values observed in adulthood for the same population), then isotopic profiles will not match the isotopic values of that same individual in later adult bone collagen samples, or the specified adult group dietary mean values (such as the average values of reproductive‐age adult females from the population), potentially confounding interpretation of an exact age where breast milk was no longer consumed. Finally, identifying smaller contributions by particular food groups will be isotopically masked by the more dominant food(s), such that if breast milk is retained in the diet but consumed in very small amounts (or consumed irregularly) it likely will not register in dentin chemistry. Therefore, we can at best make an estimate of a general time period where breast milk was significantly reduced and therefore other foods became more primary in the child's diet (Halcrow et al., 2018).
Due to differential renewal/turnover rates of bodily tissues, we can chemically analyze various tissues to document dietary inputs from particular periods of time during the life‐course. For example, teeth develop during very specific windows of childhood and record the chemical signature of the diet consumed during that time of growth (Beaumont et al., 2013; Burt & Garvie‐Lok, 2013), and then do not remodel after they are formed, creating a static tissue (however, localized chemical changes can occur due to disease or trauma). Additionally, since teeth develop at generally well‐defined chronological ages with deciduous teeth beginning in utero, and permanent third molars completing formation usually in late teens to early 20s, we can select particular teeth to target specific periods of growth and development to answer questions about early life dietary practices (Beaumont et al., 2013; Burt & Garvie‐Lok, 2013; Eerkens et al., 2011). Using teeth, we can address questions of maternal‐child interactions (breastfeeding and weaning) and child‐rearing related to food practices, and also look at dietary changes over different periods of youth. Human bone continues to remodel over the entire lifetime, and chemically reflects the average diet of at least the decade prior to death (Hedges, Clement, Thomas, & O'Connell, 2007). If an adult is sampled, their bone can be compared to one or more of their teeth, and potentially very detailed isotopic life histories can be produced, documenting dietary change or stability over the lifetime (Sealy et al., 1995; Turner et al., 2010). Additionally, biosocial experiences, such as sex, gender, age, and status can also be studied in relation to dietary patterns and can inform us of how these aspects of humanity were experienced in different cultures across time and space (Eerkens & Bartelink, 2013; Miller et al., 2018; Reitsema & Vercellotti, 2012; White, 2005).
Bioarchaeologists have generally used two strategies to study childhood dietary practices: analysis of bones from individuals who died during youth (usually sampling across different age groups from infants to adults to capture a cross‐section of dietary patterns); while others analyze teeth from humans of any age to capture discrete periods of early life and compare that to a bone sample to look at diachronic intra‐individual dietary patterns (see Beaumont et al., 2018 for a recent report on comparing these data sets). For example, Xia et al. (2018) examined rib and long bone samples from individuals from the Western Zhou Dynasty (1122–771 BC) site of Boyangcheng (薄阳城) from Anhui Province, China, for carbon and nitrogen isotope data to study dietary habits from different periods of life, including identifying breastfeeding and weaning practices. Testing a cross‐section of the mortuary population ranging from age 2 to 45+ years at time of death, the authors found that adults consumed diets with mixtures of C3 and C4 foods (individuals older than 14 years of age average δ13C –17.6‰ ±2.0‰; average δ15N 10.6‰ ±0.6‰) (Xia et al., 2018). They used the range of adult values as the comparative group for the young children, and found that by around ages 3 to 4 years their δ15N values dropped to values similar to the adults, and therefore the authors propose this is the age where weaning finished (Xia et al., 2018). They noted that children between 2 to 10 years of age also had diets slightly distinct from the adult population (lower in both δ13C and δ15N), suggesting that children ate more plant‐based diets.
The other method for studying childhood diet uses dentition to study discrete periods of time during early life, from either deciduous dentition or permanent dentition depending on the research question. For example, Yi et al. (2018) used both bone and tooth samples from a Late Neolithic population from the Gaoshan Ancient City (高山古城) in Sichuan Province, China. Their results found dietary variation over the lifetime of the individuals studied: the Neolithic Gaoshan adulthood diets ranged from δ13C –19.6 to −18.0‰, therefore dominated by C3 foods (rice), while their childhood diets ranged from δ13C –19.6 to −15.0‰, containing an unexpected signal indicating some consumption of C4 foods (millets) (Yi et al., 2018). Additionally, their incremental analysis of dentin showed that weaning occurred in that population between 2.5 and 4 years of age. These two studies show the potential for these different life‐course approaches to reveal detailed dietary patterns for individuals and populations (Xia et al., 2018; Yi et al., 2018).
Our study sampled permanent dentition from individuals who died as adults (defined here as completion of longitudinal growth as indicated by epiphyseal fusion of long bones) along with a bone sample from the same person. By studying individuals who died in adulthood, variables such as sex could be inferred (from the examination of sexual dimorphism of the pelvis and skull) and used as a factor to study how sex may relate to weaning patterns, childhood diet, and adulthood diet. We employ the incremental method of sampling human dentition, which divides a tooth into small segments from crown to root with associated age estimations for each section (Beaumont & Montgomery, 2015). This approach has demonstrated that dental incremental samples can provide dietary data over identifiable chronological periods on the order of months to years during specific times of childhood (Beaumont & Montgomery, 2016; King et al., 2018).