1Anthropologists are interested in the breastfeeding and weaning practices of ancient peoples in part because of its significant influence on infant and child health and hence the survival and growth of a population. Variables such as the length of exclusive breastfeeding, age of weaning food introduction, length and velocity of the weaning process, and types of weaning foods, are well-documented as affecting the adaptive success of historic and modern populations (Stuart-Macadam, Dettwyler, 1995; WHO, UNICEF 2003), as many chapters in this book clearly illustrate. Chemical methods, particularly stable nitrogen and carbon isotope analysis of bone, have permitted reconstruction of these variables in many prehistoric populations (e.g. Bocherens et al., 2007; Clayton et al., 2006; Fogel et al., 1989; Howcroft et al., 2014; Pearson et al., 2010; Prowse et al., 2008; Schurr, Powell 2005; Turner et al, . 2007; Waters-Rist et al., 2011; Williams et al., 2005; Wright, Schwarcz, 1999).

2The chapter by Herrscher explains the biogeochemical basis of stable isotope methodology (see chapter 1, in the third part of this book) and a review by Humphrey (2014) outlines how stable isotopes can be applied to understand dietary transitions in early life. Briefly, stable nitrogen isotopes, and to a lesser extent stable carbon isotopes, undergo trophic level shifts, from plant to herbivore to carnivore. Because breastfeeding infants are consuming their mother’s tissue they undergo a further trophic level shift (Fogel et al., 1989; Fuller et al., 2006). During life, these stable isotope values are recorded in an individual’s tissues, such as their bone and teeth, and continually improving methods allow researchers to accurately detect these values many thousands of years later. Over twenty-five years of stable isotope research on human skeletons from archaeological sites to reconstruct breastfeeding and weaning practices has contributed a vast amount to our understanding of this important life-history phase.

3This case-study is about ancient foragers (hunter-fisher-gatherers) who lived in the western Lake Baikal region of Siberia around 8,000 to 5,000 years ago (fig. 1). It is presented in more detail in Waters-Rist et al. (2011). Two successive groups lived in this area: the earlier one, called the Kitoi, existed from about 8,000 to 7,000 years ago, and is represented by the mortuary sites Shamanka II and Lokomotiv; the later one, called the Serovo, is represented by skeletons from approximately 6,000 to 5,000 years ago at the site of Ust’-Ida I (fig. 1). Extensive research by the international and interdisciplinary Baikal Archaeology Project has documented numerous cultural and biological differences between these populations, including in mortuary practices (Weber et al., 2002), dietary patterns (Katzenberg et al., 2010), health and disease (Lieverse, 2010; Waters-Rist, 2011), activity patterns (Lieverse et al., 2011), and mitochondrial DNA (Mooder et al., 2010). One major question is what happened to the Kitoi population, and the nature of the subsequent population replacement or admixture event. Demographic data have suggested that the Kitoi’s population size was declining, whereas the later Serovo were possibly growing (Link, 1999). As well, dental enamel markers of physiological stress, formed from significant bouts of malnutrition or disease, are higher in the Kitoi population suggesting the infants and children may have been in poorer health (Waters-Rist, 2011).

4Did differences in breastfeeding and weaning practices influence the size and/or health of these populations? Longer periods of breastfeeding can cause a longer period of postpartum amenorrhea wherein there is a longer period of time between births and hence a fewer number of children born per woman, and thus a smaller population (reviewed by Ellison, 1995; WHO, 1998). Considered more of a ‘folk tale’ until Henry’s 1961 publication, it is now well established that breastfeeding can exert a considerable effect on female fecundity via lactational amenorrhea (also see Vitzhum, 1994). Research has found this effect is tied most strongly to the duration and frequency of suckling (i.e. every 4-6 hours) (Howie, McNeilly, 1982; McNeilly, 2001; Stern et al., 1986) and the mother’s nutritional intake and energy balance (e.g. Ellison, Vitzhum, 1994; Worthman et al., 1993); thus it is expected that the effect will be strongest in the first months to year of an infant’s life before weaning foods are introduced and while breastmilk is the only nutritional source. In addition however, there is ample evidence that natural fertility populations who frequently nurse older infants have longer inter-birth intervals than those that nurse less frequently or wean at an earlier age (e.g. Delvoye et al., 1977; Konner, Worthman, 1980; Panter-Brick, 1991; Vitzhum, 1989; Wood et al., 1985). So, while certainly not the only variable at play, it is reasonable to consider if differences in breastfeeding and weaning were possibly contributing to the purportedly different fertility levels in the ancient Cis-Baikal hunter-gatherers. Basically, were the Kitoi breastfeeding more intensively over a comparatively longer amount of time than the Serovo? Breastfeeding and weaning length and velocity, as well as the types of weaning foods and the age at which they are given, can also markedly influence infant/child nutritional and immunological health and disease susceptibility (Sellen, 2006; WHO, UNICEF, 2003). Thus, did differences in these regards influence the Kitoi or Serovo populations? These questions can be answered using stable nitrogen and carbon isotope analyses, in particular with a novel sampling strategy that produces a quasi-longitudinal record of an infant/child’s feeding pattern.

5Small samples (<5 grams) of bone were taken from all individuals less than ten years of age, for a total of 33 Kitoi and 16 Serovo. A novel sampling strategy was designed to compare a piece of bone that had recently formed prior to death, and thus would have recorded diet most recently prior to death, to a piece of bone that had formed at an earlier age and thus reflected diet from an earlier period of life. Long bones such as those in the arms and legs grow in length (longitudinal growth) via the process of endochondral ossification wherein new bone is deposited at the ends of the bone, called the metaphyses. Bone in the middle of the bone, the diaphysis, is older, although growth in width and thickness (appositional growth) also adds a small amount of new bone to the outside surface. Bone is a cellularly active tissue, always turning over, such that over the course of many months and years older bone gets replaced by newer bone. Thus, comparison of the stable isotope values of bone samples from the metaphyses versus the diaphysis can reveal the presence of a dietary shift, such as from when the infant was still in the womb to breastfeeding, or from breastfeeding to weaning. Results from subadults up to four years of age are shown in table 1. These types of sampling methods, that analyze tissue formed at different ages of an individual’s life, are being increased used because they circumvent many of the limitations posed by the heterogeneity of a population’s stable isotope values and variability arising from non-dietary factors (e.g. Beaumont et al., 2013, 2015; Eerkens et al., 2011; Howcroft et al., 2014; Montgomery et al., 2013; Sandberg et al., 2014).

6There are similarities and differences in the Kitoi and Serovo infant feeding patterns. In both populations breast-milk was a significant source of nutrition until the age of three years. Breastfeeding to this age has many health and dietary benefits for the infant/child (Prentice, 1991). Indeed, the weaning process need not be physiologically stressful for the infant/child, despite assumptions that this is usually the case (Katzenberg et al., 1996). On average the Kitoi breastfed to a slightly older age, until three-and-a-half to four years. The Serovo infants were completely weaned by the age of three. While this difference in weaning completion is rather small, it is not the only difference. There is also a difference in the age when breastmilk appears to become a less important protein source (versus when aquatic or terrestrial protein became more important). This happened at an older age in the Kitoi (not until the age of two) and a younger age in the Serovo (already evident in a one-year-old and well underway in the two year olds). This may imply that while the Kitoi breastfed for a longer period of time, there was a shorter shift from breastmilk being the most important protein to non-breatmilk protein being most important.

Table 1 – Cis-Baikal Forager Intra-Long Bone Stable Nitrogen Isotope Ratios for Subadults Birth to Four Years of Age (n=24) with Suggested Feeding Status

7It is not possible to detect the introduction of weaning foods with the sample analyzed in this study, but the types of weaning foods can be discerned from the stable carbon and nitrogen data. The Kitoi and Serovo had diets that differed in the types of fishes that were regularly consumed, as well as possibly in the contribution of terrestrial mammals (not surprising as they lived in slightly different ecozones) (Katzenberg et al., 2010; Weber et al., 2011). These differences in population diet were clear in the weanling diet. One suggested important weaning food, for both populations, was fat. Fat is a very nutritious weaning food and is advantageous because it is easily masticated and digested by the infant. It would have been easily derived from fatty fishes in Lake Baikal or its surrounding rivers, the freshwater ‘nerpa’ seal, and/or terrestrial mammals such as deer and moose.

8So, could these differences in infant feeding have affected the adaptive success of the Kitoi and Serovo populations? Possibly, yes, even a difference of a half year to year could influence the average number of children born to a woman and hence contribute to a smaller Kitoi population versus a larger Serovo population. The differences in certain components of the weaning process, those being the age at and length of time over which breastmilk consumption significantly decreased, may also have had implications for the morbidity and mortality of the infants. A shorter or more sudden weaning phase could have been more physiologically disruptive. It is also possible that the Kitoi mothers breastfed their infants for longer because food was less secure, perhaps related to annual periods of food shortage in the late Winter/early Spring, which Waters-Rist (2011) suggests were frequent amongst the Kitoi but not the Serovo. We will never know all the variables that influenced the infant feeding practices of these populations, but this stable isotope research has demonstrated both general similarities and more subtle differences in how infants were fed in a region of Siberia many thousands of years ago. Despite its clear importance, there are many prehistoric cultures from many areas of the world that are unexplored in regards to infant feeding practices: we should remedy this in order to better understand the cultural variability and biological constraints that have shaped the evolution of this critical phase of life (Dettwyler, 2004).