PLOS ONE
RESEARCH ARTICLE
Growing up in Ancient Sardinia: Infant-toddler
dietary changes revealed by the novel use of
hydrogen isotopes (δ2H)
Saskia E. Ryan ID1¤*, Linda M. Reynard ID1, Elisa Pompianu2, Peter van Dommelen ID3,
Clizia Murgia4, M. Eulàlia Subirà5, Noreen Tuross1*
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Ryan SE, Reynard LM, Pompianu E, van
Dommelen P, Murgia C, Subirà ME, et al. (2020)
Growing up in Ancient Sardinia: Infant-toddler
dietary changes revealed by the novel use of
hydrogen isotopes (δ2H). PLoS ONE 15(7):
e0235080. https://doi.org/10.1371/journal.
pone.0235080
Editor: Dorothée Drucker, Senckenberg
Gesellschaft fur Naturforschung, GERMANY
Received: December 16, 2019
Accepted: June 8, 2020
Published: July 8, 2020
Copyright: © 2020 Ryan et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: All relevant data are
within the manuscript and its Supporting
Information files.
Funding: This research was completed with
fellowship support from the Max Planck-Harvard
Research Center for the Archaeoscience of the
Ancient Mediterranean to SER, and partial funding
from the American School for Prehistoric Research
to NT. The funders had no role in study design,
data collection and analysis, decision to publish, or
1 Department of Human Evolutionary Biology, Harvard University, Boston, Massachusetts, Unitied States of
America, 2 Independent researcher, Marrubiu, Italy, 3 Joukowsky Institute for Archaeology and the Ancient
World, Brown University, Providence, Rhode Island, United States of America, 4 Departament de Biologia
Animal, Unitat d’Antropologia Biològica, Biologia Vegetal i Ecologia, Universitat Autònoma de Barcelona,
Barcelona, Spain, 5 Grup de Recerca en Antropologia Biològica, Unitat d’Antropologia Biològica, Universitat
Autònoma de Barcelona, Barcelona, Spain
¤ Current address: Archéozoologie, Archéobotanique: Sociétés, Pratiques et Environnements, UMR CNRS
7209, Muséum National d’Histoire Naturelle, Sorbonne Université, Paris, France
* saskiaryan@fas.harvard.edu (SER); tuross@fas.harvard.edu (NT)
Abstract
Detailed information about the lives and deaths of children in antiquity is often in short supply. Childhood dietary histories are, however, recorded and maintained in the teeth of both
juveniles and adults. Primary tooth dentinal collagen does not turn over, preserving a
sequential record of dietary changes. The use of nitrogen (δ15N) and carbon (δ13C) isotope
values of incrementally sampled dentin are used in the study of breastfeeding practices but
evidence for the addition of weaning foods, both in terms of mode and, particularly, duration,
has remained analytically inaccessible to date. Here, we demonstrate how the novel use
hydrogen isotope (δ2H) values of sequentially micro-sampled dentin collagen, measured
from individuals excavated from a Punic cemetery, in Sardinia, Italy, can serve as a proxy
for weaning food type and duration in ancient childhood diet. The weaning rate and age,
based on the decline in δ15N and δ13C values of permanent first molars and the concomitant
increase in δ2H, appears to be broadly similar among six individuals. Hydrogen isotopes
vary systematically from a low value soon after birth, rising through early childhood. The
early post-birth values can be explained by the influence of 2H-depleted lipids from mother’s
breastmilk and the later δ2H rise is consistent with, among other things, a substantial portion
of boiled foodstuffs, such as the higher δ2H values observed in porridge. Overall δ2H in dentin shows great promise to elucidate infant and childhood feeding practices, and especially
the introduction of supplementary foods during the weaning process.
Introduction
The interaction of infectious disease and malnutrition have shaped the trajectory of human
evolution. Nowhere are these effects more pronounced than in infant survival. Today, the
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
1 / 13
PLOS ONE
preparation of the manuscript. There was no
additional external funding received for this study.
Competing interests: The authors have declared
that no competing interests exist.
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
global under-5 mortality rate is 3.9% [1], but the extent of infant mortality was much greater
in the past. Even as late as the mid 1800s in Italy, our greater study location, approximately
20% of infants succumbed to a number of diseases, including gastrointestinal disorders, pneumonia and bronchitis, by the age of one [2]. The introduction of weaning foods often exacerbates the exposure to disease-causing microorganisms and extends the high infant mortality
rates into toddlerhood. Weaning-related deaths existed in the past and persist as a source of
childhood mortality to this day. So, while the death rates of infants up to one year of age are
estimated at around 25% for much of human history [3], the second year of life comes with
substantial risk as infants are weaned from human milk [4]. Although some studies have documented the complex nature of infant feeding (e.g.[5]), our knowledge of ancient infant feeding
and nutrition remains limited. When textual sources are lacking, infants and young children
become even less visible. Archaeological studies underrepresent many aspects of infancy and
childhood [6] due in part to the relative scarcity and preservation of skeletal remains. In addition, infants and children who are part of the archaeological record do not necessarily reflect
the healthy child population that survived to adulthood [7–11].
Bioarchaeology has contributed to our knowledge of infant and child lifeways [12,13] and
the incorporation of stable isotopic values in the study of breastfeeding and weaning has
become standard [8,10,14], albeit not without challenges. The empirical observation that infant
bone collagen has elevated amounts of 15N as compared to adults was first reported in 1989
[15], and followed by a study of an archaic population from Florida (USA) that exhibited elevated 15N in bone collagen of children vs the adults [16]. As an isotopic medium to interpret
infant feeding patterns, bone is often hampered by population sample size [8], the time averaging of protein synthesis, and the unknown health status of individuals: all of these issues cause
fine grained determinations of breastfeeding duration from bone to be questionable.
Many of the difficulties associated with the interpretation of bone isotopes can be overcome
by the use of serial sections of dentin from both children and adults, as primary dentin maintains the dietary inputs of infancy in the form of isotopic trends [14,17–24]. Most archaeological studies of breastfeeding and weaning practices have been entirely focused on the use of
nitrogen and carbon isotopes; less frequently, oxygen isotopes, strontium and calcium concentrations, and calcium isotope ratios have been used (see [14] for a review). Here, we present a
new approach—the hydrogen isotope composition of tooth dentin—to independently document the duration of breastfeeding and, for the first time, provide an isotopic record of the
potential range of foods used in the weaning process from individuals excavated from contexts
dating from the 4th to early 2nd century BCE, from a Punic cemetery in Villamar [25–27], on
the island of Sardinia, Italy.
Hydrogen isotopes are powerful biogeochemical tracers; the hydrogen (and oxygen) isotope
composition of surface and groundwater vary geographically, typically reflecting the isotope
ratio of precipitation [28]. This variation in geographically distributed isotopes can be subsequently reflected in human and faunal tissues [29–32] along with the contribution of δ2H and
δ18O from food [33,34]. Water δ2H in cow’s milk is somewhat higher on average compared
with co-occurring drinking water [35–37], and milk lipids have been found to be uniformly
lower in δ2H values compared to either liquid source [38,39]. This finding highlights the role
of the isotopic composition in different milk components for the interpretation of infant dentin collagen δ2H values. Previous reports suggest that δ2H in bone collagen is related to trophic
level [40–45]; however, this is not universally observed, and because these reports refer to cross
species comparisons, the relevance to very young human infants is unknown.
Another potential controlling factor of dentin δ2H values is food preparation practice. Food
and water sources can be isotopically fractionated through boiling, fermentation, distillation
and cooking, though very few studies have measured the δ2H values of cooked food [46,47].
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
2 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
The introduction of heavily cooked foods, for example, cereal gruel, which has historically
been used as a transitional food during the weaning period, could influence dentin δ2H values.
Barley and wheat have played an instrumental role in the subsistence of Sardinian residents
historically [48]; botanical remains show a continuous record of barley cultivation from at
least the Nuragic Bronze Age to the Middle Ages and beyond (1500 B.C.—A.D. 1100) [49].
Naked wheat (Triticum aestivum-durum) was also found to be in abundance in Punic contexts
elsewhere in Sardinia [50], but it is thought that wheat was exported while barley was the cereal
that was consumed by inhabitants [51]. In addition to the sequential dentin collagen samples,
this study presents δ2H values of both human breastmilk and porridge, a canonical weaning
food, to assess their potential effect on dentin isotope composition.
Materials and methods
Site and individuals
We studied six human individuals and three caprids from the site of Villamar in central Sardinia (Fig 1)(S1 Table for details). Authorization for the use and exportation of the Sardinian
specimens was granted by the Soprintendenza Archeologia Belle Arti e Paesaggio per la città
metropolitana di Cagliari e le province di Oristano e sud Sardegna, with prot. 15667 of 28 July
2017. Official: Chiara Pilo. All necessary permits were obtained for the described study, which
complied with all relevant regulations. The site is a Punic cemetery (4th to early 2nd century
BCE), where 25 of approximately 50 known tombs have been excavated since 2013 [25–27](S2
Table for radiocarbon dates). A total of 299 isotopic values (δ2H of dentin samples = 86 and
bone collagen = 9; δ15N and δ13C of dentin samples = 93 and bone collagen = 9) were produced from these 6 humans and 3 caprids. The cemetery occupies a rocky outcrop of marl and
sandstone into which the graves were cut. The prevalent burial ritual is inhumation, with most
of the deceased having been buried in either chamber tombs or rock-cut niches, accessible via
a vertical and roughly rectangular shaft (See S1 Appendix for further site information).
Incremental dentin sampling
Reconstructing breastfeeding and weaning practices (BWP as defined by Tsutaya & Yoneda
2014[14]) is made possible by micro-sampling dentin and/or enamel [22,54–56]. For this
Fig 1. Villamar in Sardinia, Italy, where the cemetery from which the individuals were excavated is located. The freely available shapefile from the Ancient World
Mapping Centre (AWMC) [52] was used as the basemap and modified with QGIS software [53].
https://doi.org/10.1371/journal.pone.0235080.g001
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
3 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
study, the tissue of choice is dentin collagen from the first permanent molar due to its relatively
straightforward developmental pattern, the size of the sample required for measurement and
the ability to measure hydrogen isotopes in this tissue. Dentin formation commences prior to
enamel development and starts at the crown, extending progressively towards the root [57].
Matching the biological age to the incremental sections of dentin cannot be done with complete accuracy as the dentin grows in stacked cones [58] and serial sections are taken horizontally [14], thus cutting across overlapping dentin layers which have been deposited at different
times. Therefore the values measured from a given section represent a moving average of the
inherent isotope mixing [55]. We have taken the approach here of assigning estimated ages
based on the midpoint of the section, and based on the AlQahtani Atlas (2010) [59]. The age at
which M1 dentin formation begins is approximately 4.5 ± 3 months and the age at apex closure
is approximately 9.5 ± 0.5 years.
The method of sectioning the dentin was adapted from ‘Method 2’ used by Beaumont et al.,
2013 [60], which in turn was based on the work of Kirsanow et al., (2008) [55]. Using a rotary
blade in a dental hand piece, each tooth was cut in half to isolate one complete root and adjoining section of the crown. Samples were photographed, sonicated in deionized (dI) water, and
immersed in 0.5 M EDTA (pH 7.3) which was replaced every 1–2 days until complete demineralization. Dentin was rinsed 15 times in dI water and sequentially cut with a scalpel into thin
parallel sections from the coronal dentin horn to the root cervix in 1 mm increments, perpendicular to the central axis of the root. Demineralized dentin sections were freeze-dried. The
sampling procedure generated between 12 and 22 serial sections per tooth for a total n = 105.
Bone sample preparation
Bone was sampled in order to gauge later-life isotope values [61–63]. Bone was demineralized
in 0.5 M EDTA (pH 7.3) and washed 15 times with deionized water [64]. All of the bone collagen was well-preserved based on the criteria of atomic C/N ratios [65,66].
Cooking experiment
Barley (Hordeum vulgare) was cooked in water at a temperature of 100˚C for 135 minutes. The
cooking water, partially-soluble barley fraction and insoluble barley fraction (grains) were
sampled at 15 minute intervals throughout the cooking period and their respective hydrogen
isotope values were measured.
Breastmilk (BM) and urine separation
One mother and infant pair was recruited from the general region near Cambridge, Massachusetts, USA, and the mother gave written permission for sample collection. The University Area Institutional Review Board of Harvard University granted permission for human
tissue from the mother and infant to be used in this study (IRB14-3463). The mother’s diet
was omnivorous and unconstrained. Samples were collected periodically over a ~5 month
period, beginning approximately two weeks after the infant’s birth. Expressed milk (n = 5)
and urine (n = 7) samples were collected from the mother, and feces were collected from
her infant baby (n = 5). The samples were frozen immediately after collection. One millilitre
of defrosted milk was centrifuged at 3500 rpm for 30 min at room temperature to separate
the lipid-protein mixture from the less dense water-carbohydrate components (BM water).
The solid lipid-protein fraction was refrigerated and the lipids (BM lipid-rich) separated
from the remaining mixture (BM lipid-depleted). After both fractions were sampled individually, they were homogenised and re-measured (BM solids). Urine samples were centrifuged at 3500 rpm for 30 min at room temperature and the supernatant (urine water) was
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
4 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
collected for analysis. A fraction of the urine was also freeze-dried to facilitate transfer to silver capsules for analysis (urine solid). Breastmilk fluid and urine were filtered through
0.45 μm syringe filters prior to isotopic measurements.
Isotope measurement
Solid samples—dentin, collagen, barley, various milk non-water fractions, urine solids,
and feces—were measured by continuous-flow mass spectrometry at Harvard University
on a Thermo-Finnegan Delta Plus XP. Dentin and bone collagen δ15N and δ13C values
were obtained via a Costech Elemental Analyser, using USGS 40 and USGS 41 glutamic
acid as standards, and data are reported relative to AIR for δ15N and VPDB for δ13C. The
δ2H values of solid materials were obtained using a Thermo Thermal-Conversion Elemental Analyser (TC/EA), previously described in detail [64]. The δ2H values of bone collagen,
infant feces, breastmilk solids and urine solids were analysed using a chromium (Cr)packed reactor, as described in Reynard et al., 2019 [67]. A subset of dentin samples was
analysed at the United States Geological Survey, also using a Cr-packed reactor (see S3
Table for further information). Tooth dentin δ2H measurements were obtained with a
glassy carbon furnace packing and were converted to equivalent Cr-reactor-run values
using the conversion factor that was also presented by Reynard et al., 2019 [67]. Samples
were analysed in duplicate where possible and all δ2H data are normalised to the
VSMOW-SLAP scale, using VSMOW and SLAP enclosed in sealed silver tubes as standards. The average standard deviation of the duplicate samples of bone collagen δ2H
was ± 3 ‰ and 2 ‰ for tooth dentin (S3 and S4 Tables). The long-term keratin standard
value is -89 ± 4 ‰ for Cr-δ2H (n = 63). Breastmilk water and urine water were analysed on
a Los Gatos Research T-LWIA-45-EP liquid-water analyser for δ2H and δ18O and are
reported relative to SMOW.
Results and discussion
Isotopic changes in dentin collagen during childhood
Each of the six individuals have strikingly similar trends in δ13C and δ15N values in their M1
throughout childhood, suggesting similar weaning practices in this population (Fig 2A and 2B,
respectively). We note that both δ13C and δ15N values progressively fall during the formation
of the M1 dentin indicating that less and less breastmilk was part of the toddler diet with
increasing age. The δ13C and δ15N values gradually plateau in toddlerhood, reaching values
close to those of the bone collagen (representing the years before death–adulthood in most
cases here) (Fig 2D). Two individuals (US 320 ID1 and US 319 CR2), who died before reaching
adolescence, marked by the cessation of root growth, do not have notably different δ15N patterns compared with those who survived into adulthood, suggesting no identifiable N isotope
bias due to childhood mortality.
Hydrogen isotope patterns are dramatically distinct from the corresponding δ15N and δ13C
weaning curves with an inverse pattern seen in the earliest years of dentin deposition (Fig 2D).
From the earliest post-birth time segment in the M1 dentin, δ2H values rise rapidly until they
peak and begin to decrease again, wavering through later childhood, with values in the latter
sections closely corresponding to those of the average bone collagen δ2H value (Fig 2C and
2D). As the isotopic patterns of permanent M1 dentin collagen do not reflect dietary input
right at birth, both the high δ15N and δ13C values may have already declined from a maximum
value. The permanent M1 dentin collagen does, however, capture the total trajectory of nonbreastmilk dietary introduction to the infant/toddler.
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
5 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
Fig 2. The left panels show the separate carbon (A), nitrogen (B) and hydrogen (C) isotope values from each individual (n = 6) from the incremental first permanent
molar dentin sections. A range of ~1–3 ‰ in δ13C values, (A), ~4–5 ‰ in δ15N (B) and ~14–29 ‰ in δ2H values (C) is noted across the total length of tooth dentin.
The right panel (D) shows the average values of the permanent first molars from 6 different individuals. For the purpose of visual representation, each data point from
a 1 mm section is plotted in sequential order. � Approximate age in years is based on the commencement and cessation ages of tooth dentin determined by AlQahtani
et al., (2010) [59], assuming a constant growth rate of dentin, factoring in the length of the tooth. The average length of the M1s sampled is used in the case of ‘319 CR2’
and ‘320 ID1’ whose dentin was not fully mineralised.
https://doi.org/10.1371/journal.pone.0235080.g002
Dietary sources contributing to dentin collagen isotope composition
Nitrogen. The nitrogen source in the diet of monotonously breastfed infants is milk supplied from the mother. This milk source reflects the diet of the mother and in an agricultural
community such as Villamar, the adults of the community were likely omnivores. The δ15N of
bone collagen from Villamar averaged 11.5 ± 0.5 ‰ (n = 6), consistent with an adult diet that
contained meat products. For two reasons, we argue that the likely source of nitrogen in weaning foods were likely plant based. First, the average δ15N of caprid bone collagen (n = 3), is
10.5 ± 0.9 ‰, only 1 ‰ lower than the co-occurring adults at Villamar, making caprid milk or
meat an unlikely major dietary component for human weanlings. Second, the total average
decline from in M1 dentin δ15N collagen was 4.3 ± 0.5 ‰ (average ± SD total range, n = 6),
consistent with a mixed weaning diet that contained largely plant derived products.
Carbon. The δ13C values of dentin mirror those of δ15N (Fig 2D), declining by on average
1.5 ± 0.8 ‰ (n = 6) over the time represented. The drop in δ13C is consistent with an enriched
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
6 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
animal (mother) input followed by the introduction of C3 plant-based weaning foods with
lower δ13C values.
Hydrogen. Far less is known about the hydrogen isotope composition of potential weaning foods, and for that reason we undertook two types of analyses. From a mother-baby pair,
we observe human breastmilk water δ2H values to be slightly higher than that of local drinking
water values (Δaverage of breastmilk water- drinking water = 4.8 ± 1.5 ‰), in agreement with observed
2
H enrichment in cow’s milk water over drinking water [38,68]. In addition, as previously
reported [69], human breastmilk and urine water were isotopically indistinguishable in δ2H.
The non-water fractions are all significantly 2H depleted vs. both the breastmilk water and the
local drinking water (Fig 3), in agreement with previously reported data from cows [38].
The hydrogen isotope composition of the infant/toddler dentin is depleted in 2H in early
infancy (Δaverage of dentin horn section-bone collagen = -15.0 ± 5.8 ‰ of earliest M1 dentin vs. bone
collagen) and steadily rises over the period in which the nitrogen isotope patterns would suggest weaning is taking place. It is likely that lipid-rich human breastmilk is a sole dietary source
during the early stages of infancy and that this monotonous food source significantly contributes to the early depletion in 2H.
The simultaneous high δ15N, δ13C and low δ2H in the early-formed dentin segments demonstrates that δ2H values in dentin are not elevated during exclusive breastfeeding. The dentin
δ2H values are not “trophic” in the sense used in some hydrogen isotopic literature (referring
to herbivores vs. omnivores vs. carnivores) [40–45], but rather reflect early milk inputs.
The rise in δ2H values in the M1 dentin is inconsistent with a diet dominated by animal products, and in agreement with the nitrogen isotopic data which suggest that a substitute animal milk
source was not a major contribution to the infant diet during the weaning period. There are
many other possible foods that could have been used as a substitute food source for weaning
infants of Villamar, including likely soft sources such as fruit which often have elevated δ2H values relative to precipitation [70–72]. We chose to test whether cereal gruel/porridge hydrogen isotope values were consistent with the observed increase in the infant dentine collagen.
The δ2H values of the partially-soluble fraction of barley, i.e. the food fraction that would
likely have been consumed by an infant being weaned, and the cooking water (Fig 4) become
increasingly elevated in 2H over the cooking time (S6 Table). Barley gruel is therefore a potential source of high δ2H dietary inputs which could subsequently be reflected in tooth dentin. A
major factor that controls the starting point and duration of infant weaning is the availability
of a suitable supplementary food that has an appropriate nutrient profile and is easy to swallow
and digest. The type of food that was used for weaning in Punic Sardinia could very likely have
been a grain-based porridge. However, it is likely that any grain treated in the above manner
would result in an increase in δ2H values in both the water and partially solubilized grain.
Brewing, boiling, and stewing have been shown to result in an increase in the δ18O of stew
water of 10.3‰ on average [46]. Tuross et al., (2017) [47] document the change in the δ18O of
biochemical constituents in foods (beef and sweet potatoes) through application of heat, and
showed a change in δ18O of Δ 6‰ pre- and post-cooking in the most extreme case. Here, we
confirm that the fractionation of hydrogen isotopes occurs as a result of cooking both in the
water, as expected [73] and in the partially solubilized grain. Accordingly, one potential cause of
elevated hydrogen isotope values in dentin is the consumption of heavily cooked barley gruel.
Conclusions
The duration of the breastfeeding and weaning period varies widely depending on cultural,
biological and environmental factors [74,75]. This study provides a new powerful method for
examining the decline in the consumption of mother’s breastmilk and the introduction of
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
7 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
Fig 3. Hydrogen isotope values of infant feces, and maternal breastmilk fractions and urine water. See S5 Table for data. Open circles are outliers. Solid samples were
analysed using a Cr-packed reactor.
https://doi.org/10.1371/journal.pone.0235080.g003
weaning foods. Reconstructing breastfeeding and weaning practices of ancient populations
can be accomplished by sequential isotope measurement of incremental tooth dentin from
adult M1 permanent molars, obviating the need for infant skeletal material. The fine-grained
pattern of increasing δ2H and declining δ15N and δ13C in dentin documents the gradual introduction of weaning foods over a duration of one to two years in the Villamar individuals.
Dentin laid down in early infancy has a lower δ2H value consistent with breastmilk intake.
In the following dentin sections δ2H rises sharply with the concurrent decline in δ15N and
δ13C values. All three isotopic ratios are consistent with a substitute weaning food dominated
by plant source(s), and we provide evidence for the source of the increase in δ2H during the
later first, second and third year of life as consistent with a food such as cooked grains, and
inconsistent with substantial animal products. Given the large amplitude shifts observed in
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
8 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
Fig 4. The time progression (mins) of hydrogen isotopes of cooking water and partially-soluble barley fractions of an experimental barley-water gruel. Both the
cooking water and the partially-soluble barley fraction become elevated in 2H as a result of cooking. Data points are straight-line joined. The hydrogen isotopes of the
insoluble barley fraction (not shown) are not progressively fractionated over the cooking duration (See S6 Table for data).
https://doi.org/10.1371/journal.pone.0235080.g004
δ2H values in sequential dentin samples, consistent with likely dietary inputs (breastmilk and
cooked barley porridge), this isotopic tracer reveals more detailed and potentially new and otherwise untraceable dietary sources in infancy and early childhood.
Supporting information
S1 Appendix. Villamar site details.
(PDF)
S1 Table. List of individuals with key information pertaining to the sample analyses from
each.
(DOCX)
S2 Table. Radiocarbon dates of selected Villamar individuals.
(DOCX)
S3 Table. Hydrogen, nitrogen and carbon isotope values, standard deviations and sample
numbers of individual dentin sections taken from the first permanent molars of six individuals from Villamar. � δ2H analysed at Harvard were converted to equivalent Cr-reactor
run values using a conversion factor presented by Reynard et al., 2019[67].
(DOCX)
S4 Table. Bone collagen hydrogen, nitrogen and carbon isotope values of Villamar humans
and caprids.
(DOCX)
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
9 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
S5 Table. Hydrogen isotope values (‰) of infant feces, maternal breastmilk and urine
(chromium (Cr)-packed reactor), and local meteoric water which were collected over a
period of ~5 months.
(DOCX)
S6 Table. Hydrogen isotope values (‰) of barley cooking experiment fractions. � Glassy-C
reactor configuration.
(DOCX)
Acknowledgments
We thank the volunteer mother-baby pair who provided samples. In addition, we thank
Samantha Lash and Richard Waldbauer for assistance in the field.
Author Contributions
Conceptualization: Saskia E. Ryan, Noreen Tuross.
Data curation: Saskia E. Ryan, Linda M. Reynard, Noreen Tuross.
Formal analysis: Saskia E. Ryan, Linda M. Reynard, Elisa Pompianu, Clizia Murgia.
Funding acquisition: Noreen Tuross.
Investigation: Saskia E. Ryan, Elisa Pompianu, Peter van Dommelen, Noreen Tuross.
Methodology: Saskia E. Ryan, Linda M. Reynard, Noreen Tuross.
Resources: Elisa Pompianu, Peter van Dommelen, M. Eulàlia Subirà, Noreen Tuross.
Supervision: M. Eulàlia Subirà, Noreen Tuross.
Validation: Saskia E. Ryan, Linda M. Reynard, Noreen Tuross.
Visualization: Saskia E. Ryan, Noreen Tuross.
Writing – original draft: Saskia E. Ryan, Linda M. Reynard, Noreen Tuross.
Writing – review & editing: Saskia E. Ryan, Linda M. Reynard, Elisa Pompianu, Peter van
Dommelen, Clizia Murgia, M. Eulàlia Subirà, Noreen Tuross.
References
1.
World Health Organization. World Health Statistics 2019: Monitoring Health for the SDGs. World Health
Organization. 2019. https://doi.org/10.1007/978-1-349-04787-1_12
2.
Del Panta L. Infant and child mortality in Italy, eighteenth to twentieth century: long-term trends and territorial differences. Infant and child mortality in the past. 1997.
3.
Volk T, Atkinson J. Is child death the crucible of human evolution? J Soc Evol Cult Psychol. 2008;247.
https://doi.org/10.1037/h0099341
4.
Gordon JE, Wyon JB, Ascoli W. The second year death rate in less developed countries. Am J Med Sci.
1967; 254: 357–380. PMID: 6054538
5.
Pozzi L. The determinants of infant and childhood mortality: A complex tangle in the historical research.
Atti della XLI Riun Sci Milano. 2002; 5–7.
6.
Baxter JE. The Archaeology of Childhood. Annu Rev Anthropol. 2008; 37: 159–175. https://doi.org/10.
1146/annurev.anthro.37.081407.085129
7.
Beaumont J, Montgomery J, Buckberry J, Jay M. Infant mortality and isotopic complexity: New
approaches to stress, maternal health, and weaning. Am J Phys Anthropol. 2015; 157: 441–457.
https://doi.org/10.1002/ajpa.22736 PMID: 25773670
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
10 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
8.
Reynard LM, Tuross N. The known, the unknown and the unknowable: Weaning times from archaeological bones using nitrogen isotope ratios. J Archaeol Sci. 2015; 53: 618–625. https://doi.org/10.1016/j.
jas.2014.11.018
9.
Prowse TL, Saunders SR, Schwarcz HP, Garnsey P, Macchiarelli R, Bondioli L. Isotopic and dental evidence for infant and young child feeding practices in an imperial roman skeletal sample. Am J Phys
Anthropol. 2008; 137: 294–308. https://doi.org/10.1002/ajpa.20870 PMID: 18615573
10.
Howcroft R. Weaned Upon A Time: Studies of the Infant Diet in Prehistory. Stockholm: Stockholm University; 2013.
11.
Beaumont J. The Whole Tooth and Nothing but the Tooth: Or why Temporal Resolution of Bone Collagen May Be Unreliable. Archaeometry. 2020; https://doi.org/10.1111/arcm.12544
12.
Perry MA. Redefining Childhood through Bioarchaeology: Toward an Archaeological and Biological
Understanding of Children in Antiquity. Archeol Pap Am Anthropol Assoc. 2006; 89–111. https://doi.org/
10.1525/ap3a.2006.15.89
13.
Stuart-Macadam P, Stuart-Macadam P. Breastfeeding in Prehistory. Breastfeeding. Routledge;
2017. pp. 75–100. https://doi.org/10.4324/9781315081984-3
14.
Tsutaya T, Yoneda M. Reconstruction of breastfeeding and weaning practices using stable isotope and
trace element analyses: A review. American Journal of Physical Anthropology. 2014. pp. 2–21. https://
doi.org/10.1002/ajpa.22657 PMID: 25407359
15.
Fogel ML, Tuross N, Owsley DW. Nitrogen isotope tracers of human lactation in modern and archaeological populations. Annu Rep Dir Geophys Lab 1988–1989 Carnegie Inst Washington. 1989;
16.
Tuross N, Fogel ML. Exceptional molecular preservation in the fossil record: the archaeological, conservation and scientific challenge. Archaeometry of pre-Columbian sites and artifacts: proceedings of a
symposium Los Angeles: Getty Conservation Institute. 1994.
17.
King CL, Millard AR, Gröcke DR, Standen VG, Arriaza BT, Halcrow SE. A comparison of using bulk and
incremental isotopic analyses to establish weaning practices in the past. STAR Sci Technol Archaeol
Res. Taylor & Francis; 2018; 8923: 1–9. https://doi.org/10.1080/20548923.2018.1443548
18.
Czermak A, Schermelleh L, Lee-Thorp J. Imaging-assisted time-resolved dentine sampling to track
weaning histories. Int J Osteoarchaeol. 2018; 28: 535–541. https://doi.org/10.1002/oa.2697
19.
Tsutaya T, Miyamoto H, Uno H, Omori T, Gakuhari T, Inahara A, et al. From cradle to grave: multi-isotopic investigations on the life history of a higher-status female from Edo-period Japan. 2016; 124: 185–
197. https://doi.org/10.1537/ase.161029
20.
Beaumont J, Montgomery J. The Great Irish Famine: Identifying Starvation in the Tissues of Victims
Using Stable Isotope Analysis of Bone and Incremental Dentine Collagen. PLoS One. 2016; 11: 1–21.
https://doi.org/10.1371/journal.pone.0160065 PMID: 27508412
21.
Burt NM, Garvie-Lok S. A new method of dentine microsampling of deciduous teeth for stable isotope
ratio analysis. J Archaeol Sci. 2013; 40: 3854–3864. https://doi.org/10.1016/j.jas.2013.05.022
22.
Eerkens JW, Berget AG, Bartelink EJ. Estimating weaning and early childhood diet from serial microsamples of dentin collagen. J Archaeol Sci. 2011; 38: 3101–3111. https://doi.org/10.1016/j.jas.2011.07.
010
23.
Fuller BT, Richards MP, Mays SA. Stable carbon and nitrogen isotope variations in tooth dentine serial
sections from Wharram Percy. J Archaeol Sci. 2003; 30: 1673–1684. https://doi.org/10.1016/S03054403(03)00073-6
24.
Scharlotta I, Goude G, Herrscher E, Bazaliiskii VI, Weber AW. “Mind the gap”—Assessing methods for
aligning age determination and growth rate in multi-molar sequences of dietary isotopic data. Am J Hum
Biol. 2018; 30: 1–13. https://doi.org/10.1002/ajhb.23163 PMID: 30129288
25.
Pompianu E. Nuovi scavi nella necropoli punica di Villamar (2013–2015). FOLD&R it. 2017; 395: 1–28.
Available: http://www.fastionline.org
26.
Pompianu E. Cartagine in Sardegna. Nota su alcuni contesti con incinerazioni dalla necropoli di Villamar. Folia Phoenicia. 2019; 3: 99–116.
27.
Guirguis M, Pla Orquı́n R, Pompianu E. Premature deaths in Punic Sardinia. Perception of childhood in
funerary contexts from Monte Sirai and Villamar. In: Tabolli J, editor. From Invisible to Visible New data
and Methods for the Archaeology of infant and Child Burials in Pre-Roman Italy and Beyond. Studies in
Mediterranean Archaeology 149, Astrom Editions, Nicosia; 2018. pp. 207–215.
28.
Dansgaard W. Stable isotopes in precipitation. Tellus. 1964; 16: 436–468. https://doi.org/10.3402/
tellusa.v16i4.8993
29.
Bowen GJ. Isoscapes: spatial pattern in isotopic biogeochemistry. Ann Rev Earth Planet Sci. 2010; 38:
161–187.
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
11 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
30.
Ehleringer JR, Bowen GJ, Chesson L a., West AG, Podlesak DW, Cerling TE. Hydrogen and oxygen
isotope ratios in human hair are related to geography. Proc Natl Acad Sci U S A. 2008; 105: 2788–2793.
https://doi.org/10.1073/pnas.0712228105 PMID: 18299562
31.
Sharp ZD, Atudorei V, Panarello HO, Fernández J, Douthitt C. Hydrogen isotope systematics of hair:
Archeological and forensic applications. J Archaeol Sci. 2003; 30: 1709–1716. https://doi.org/10.1016/
S0305-4403(03)00071-2
32.
Bowen GJ, Wilkinson B. Spatial distribution of delta O-18 in meteoric precipitation. Geology. 2002; 30:
315–318. https://doi.org/10.1130/0091-7613(2002)030<0315:sdooim>2.0.co;2
33.
Kirsanow K, Tuross N. Oxygen and hydrogen isotopes in rodent tissues: Impact of diet, water and
ontogeny. Palaeogeogr Palaeoclimatol Palaeoecol. 2011; 310: 9–16. https://doi.org/10.1016/j.palaeo.
2011.03.022
34.
Podlesak DW, Torregrossa AM, Ehleringer JR, Dearing MD, Passey BH, Cerling TE. Turnover of oxygen and hydrogen isotopes in the body water, CO2, hair, and enamel of a small mammal. Geochim
Cosmochim Acta. 2008; 72: 19–35. https://doi.org/10.1016/j.gca.2007.10.003
35.
Lin GP, Rau YH, Chen YF, Chou CC, Fu WG. Measurements of δD and δ18Ostable isotope ratios in
milk. J Food Sci. 2003; 68: 2192–2195. https://doi.org/10.1111/j.1365-2621.2003.tb05745.x
36.
Abeni F, Petrera F, Capelletti M, Prà AD, Bontempo L, Tonon A, et al. Hydrogen and oxygen stable isotope fractionation in body fluid compartments of dairy cattle according to season, farm, breed, and reproductive stage. PLoS One. 2015;10. https://doi.org/10.1371/journal.pone.0127391 PMID: 25996911
37.
Chesson LA, Valenzuela LO, O’Grady SP, Cerling TE, Ehleringer JR. Hydrogen and Oxygen Stable
Isotope Ratios of Milk in the United States. J Agric Food Chem. 2010; https://doi.org/10.1021/jf904151c
PMID: 20104849
38.
Ehtesham E, Hayman A, Van Hale R, Frew R. Influence of feed and water on the stable isotopic composition of dairy milk. Int Dairy J. 2015; https://doi.org/10.1016/j.idairyj.2014.09.006 PMID: 25908898
39.
Dunne J, Rebay-Salisbury K, Salisbury RB, Frisch A, Walton-Doyle C, Evershed RP. Milk of ruminants
in ceramic baby bottles from prehistoric child graves. Nature. 2019; 574: 246–248. https://doi.org/10.
1038/s41586-019-1572-x PMID: 31554964
40.
Reynard LM, Hedges REM. Stable hydrogen isotopes of bone collagen in palaeodietary and palaeoenvironmental reconstruction. J Archaeol Sci. 2008; 35: 1934–1942. https://doi.org/10.1016/j.jas.2007.12.004
41.
Birchall J, O’Connell TC, Heaton THE, Hedges REM. Hydrogen isotope ratios in animal body protein
reflect trophic level. J Anim Ecol. 2005; 74: 877–881. https://doi.org/10.1111/j.1365-2656.2005.00979.x
42.
Topalov K, Schimmelmann A, David Polly P, Sauer PE, Lowry M. Environmental, trophic, and ecological factors influencing bone collagen δ2H. Geochim Cosmochim Acta. 2013; 111: 88–104. https://doi.
org/10.1016/j.gca.2012.11.017
43.
Gröcke DR, Sauer PE, Bridault A, Drucker DG, Germonpré M, Bocherens H. Hydrogen isotopes in
Quaternary mammal collagen from Europe. J Archaeol Sci Reports. 2017; 11: 12–16. https://doi.org/10.
1016/j.jasrep.2016.11.020
44.
Bowen GJ, Ehleringer JR, Chesson LA, Thompson AH, Podlesak DW, Cerling TE. Dietary and physiological controls on the hydrogen and oxygen isotope ratios of hair from mid-20th century indigenous
populations. Am J Phys Anthropol. 2009; 139: 494–504. https://doi.org/10.1002/ajpa.21008 PMID:
19235792
45.
van der Sluis LG, Reimer PJ, Ogle N. Adding Hydrogen to the Isotopic Inventory—Combining δ13C,
δ15N and δ2H Stable Isotope Analysis for Palaeodietary Purposes on Archaeological Bone. Archaeometry. 2019; 61: 720–749. https://doi.org/10.1111/arcm.12441
46.
Brettell R, Montgomery J, Evans J. Brewing and stewing: the effect of culturally mediated behaviour on
the oxygen isotope composition of ingested fluids and the implications for human provenance studies. J
Anal At Spectrom. 2012; 27: 778. https://doi.org/10.1039/c2ja10335d
47.
Tuross N, Reynard LM, Harvey E, Coppa A, McCormick M. Human skeletal development and feeding
behavior: the impact on oxygen isotopes. Archaeol Anthropol Sci. Archaeological and Anthropological
Sciences; 2017; https://doi.org/10.1007/s12520-017-0486-5
48.
Spano Giammellaro A. The Phoenicians and the Carthaginians: the Early Mediterranean Diet. In: Flandrin J-L, Montanari M, editors. Food: A Culinary History from Antiquity to the Present. Columbia University Press 1999; 1999.
49.
Bakels C. Plant remains from Sardinia, Italy, with notes on barley and grape. Vegetation History and
Archaeobotany. 2002. https://doi.org/10.1007/s003340200334
50.
Pérez-Jordà G, Hurley J, Ramis D, van Dommelen P. Iron Age botanical remains from nuraghe S’Urachi, Sardinia. Antiquity. 2020; 1–8.
51.
Attene G, Ceccarelli S, Papa R. The barley (Hordeum vulgare L.) of Sardinia, Italy. Genet Resour Crop
Evol. 1996; 43: 385–393.
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
12 / 13
PLOS ONE
Dietary changes through childhood revealed by the novel use of hydrogen isotopes in tooth dentin
52.
Ancient World Mapping Center. “Map Tiles”. [Internet]. [cited 6 Nov 2018]. Available: http://awmc.unc.
edu/wordpress/tiles/
53.
QGIS Development Team. QGIS Geographic Information System [Internet]. Open Source Geospatial
Foundation. 2018. http://www.qgis.org/
54.
Balasse M. Reconstructing dietary and environmental history from enamel isotopic analysis: Time resolution
of intra-tooth sequential sampling. Int J Osteoarchaeol. 2002; 12: 155–165. https://doi.org/10.1002/oa.601
55.
Kirsanow K, Makarewicz CA, Tuross N. Stable oxygen (d18O) and hydrogen (dD) isotopes in ovicaprid
dentinal collagen record seasonal variation. J Archaeol Sci. 2008; 35: 3159–3167. https://doi.org/10.
1016/j.jas.2008.06.025
56.
Eerkens JW, Bartelink EJ. Sex-biased weaning and early childhood diet among middle holocene
hunter-gatherers in Central California. Am J Phys Anthropol. 2013; 152: 471–483. https://doi.org/10.
1002/ajpa.22384 PMID: 24127159
57.
Hillson S. Teeth. Cambridge: Cambridge University Press.; 1986.
58.
Smith TM. Incremental dental development: Methods and applications in hominoid evolutionary studies.
J Hum Evol. 2008; https://doi.org/10.1016/j.jhevol.2007.09.020 PMID: 18045649
59.
AlQahtani SJ, Hector MP, Liversidge HM. Brief communication: The London atlas of human tooth development and eruption. Am J Phys Anthropol. 2010; 142: 481–490. https://doi.org/10.1002/ajpa.21258
PMID: 20310064
60.
Beaumont J, Gledhill A, Lee-Thorp JA, Montgomery J. Childhood diet: A closer examination of the evidence from dental tissues using stable isotope analysis of incremental human dentine*. Archaeometry.
2013; 55: 277–295. https://doi.org/10.1111/j.1475-4754.2012.00682.x
61.
Hedges REM, Clement JG, Thomas CDL, O’Connell TC. Collagen turnover in the adult femoral midshaft: Modeled from anthropogenic radiocarbon tracer measurements. Am J Phys Anthropol. 2007;
133: 808–816. https://doi.org/10.1002/ajpa.20598 PMID: 17405135
62.
Sealy JC. Body tissue chemistry and Palaeodiet. In: Brothwell DR, Pollard AM (Eds), Handbook of
Archaeological Science. John Wiley & Sons, Chichester; 2001. pp. 269–279.
63.
Ambrose SH, Norr L. Experimental evidence for the relationship of the carbon isotope ratios of whole
diet and dietary protein to those of bone collagen and carbonate. In, Lambert J B & Grupe G (Eds) Prehistoric Human Bone–Archaeology at the Molecular Level. Springer Verlag; 1993. pp. 1–38.
64.
Tuross N, Warinner C, Kirsanow K, Kester C. Organic oxygen and hydrogen isotopes in a porcine controlled dietary study. Rapid Communications in Mass Spectrometry. 2008. pp. 1741–1745. https://doi.
org/10.1002/rcm.3556 PMID: 18446769
65.
DeNiro MJ. Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to
palaeodietary reconstruction. Nature. 1985; 317: 806–809. https://doi.org/10.1038/317806a0
66.
Van Klinken GJ. Bone collagen quality indicators for palaeodietary and radiocarbon measurements. J
Archaeol Sci. 1999; 26: 687–695. https://doi.org/10.1006/jasc.1998.0385
67.
Reynard LM, Ryan SE, Tuross N. The interconversion of δ2H values of collagen between thermal conversion reactor configurations. Rapid Commun Mass Spectrom. 2019; 33: 678–682. https://doi.org/10.
1002/rcm.8396 PMID: 30681206
68.
Masud Z, Vallet C, Martin GJ. Stable Isotope Characterization of Milk Components and Whey Ethanol.
J Agric Food Chem. 1999; 47: 4693–4699. https://doi.org/10.1021/jf9900027 PMID: 10552874
69.
Wong WW, Lee LS, Klein PD. Deuterium and oxygen-18 measurements on microliter samples of urine,
plasma, saliva, and human milk. Am J Clin Nutr. 1987; 45: 905–913. https://doi.org/10.1093/ajcn/45.5.
905 PMID: 3578092
70.
Ingraham NL, Caldwell EA. Influence of weather on the stable isotopic ratios of wines: Tools for
weather/climate reconstruction? J Geophys Res Atmos. 1999; https://doi.org/10.1029/98JD00421
71.
Bong YS, Lee KS, Shin WJ, Ryu JS. Comparison of the oxygen and hydrogen isotopes in the juices of
fast-growing vegetables and slow-growing fruits. Rapid Commun Mass Spectrom. 2008; 22: 2809–
2812. https://doi.org/10.1002/rcm.3663 PMID: 18712704
72.
Bricout J. Composition en isotopes stable de l’eau des jus de fruits. Ann des Falsif l’expertise Chim Toxicol. 1973;
73.
Sharp Z. Principles of Stable Isotope Geochemistry, 2nd Edition [Internet]. 2017. Available: https://doi.
org/10.25844/h9q1-0p82
74.
Rebay-Salisbury K. Breast is best–and are there alternatives? Feeding babies and young children in
prehistoric Europe. Mitteilungen der Anthropologischen Gesellschaft in Wien. 2017. pp. 13–29.
75.
Fulminante F. Infant feeding practices in Europe and the Mediterranean from prehistory to the middle
ages: A comparison between the historical sources and bioarchaeology. Child Past. 2015; 8: 24–47.
https://doi.org/10.1179/1758571615Z.00000000026
PLOS ONE | https://doi.org/10.1371/journal.pone.0235080 July 8, 2020
13 / 13