Stable carbon and nitrogen isotope variations in tooth dentine [630108]

Stable carbon and nitrogen isotope variations in tooth dentine
serial sections from Wharram Percy
B.T. Fullera,b*, M.P. Richardsc, S.A. Maysd
aResearch Laboratory for Archaeology and the History of Art, University of Oxford, 6 Keble Road, Oxford OX1 3QJ, UK
bDepartment of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, UK
cDepartment of Archaeological Sciences, University of Bradford, Bradford, West Yorkshire BD7 1DP, UK
dAncient Monuments Laboratory, English Heritage Centre for Archaeology, Eastney, Portsmouth PO4 9LD, UK
Received 21 February 2003; received in revised form 7 May 2003; accepted 8 May 2003
Abstract
Here we report /afii982913C and /afii982915N measurements of serial sections of human deciduous and permanent tooth dentine from
archaeological samples taken from the medieval village site of Wharram Percy, Yorkshire, UK. We found a pattern of enrichment,for both /afii9829
13C and /afii982915N, where the tooth crown was greater than the cervical part of the root, which in turn was greater than the
apical portion of the root and the associated rib collagen values. This pattern reflects a decrease in the consumption of isotopicallyenriched breast milk and the introduction of less enriched weaning foods in the diet. The (mean /p5SD) di fference between the
deciduous second molar crowns and corresponding rib samples from the same individuals after 2 years of age was 1.2 /p50.4‰ for
/afii9829
13C and 3.2 /p50.8‰ for /afii982915N. The /afii982915N values are as predicted, but as there were no C4plants at Wharram Percy, this 1.2‰
enrichment in /afii982913C represents clear evidence of a carbon trophic level e ffect in collagen from breastfeeding infants. Carbon and
nitrogen results also show that the infant diet among those who died in infancy did not di ffer from those who survived into
childhood. This study demonstrates the promise of using dentine serial sections to study the temporal relationships of breastfeeding,weaning, and dietary patterns of single individuals./p302003 Elsevier Ltd. All rights reserved.
Keywords: Weaning; Palaeodiet; Dentine; Stable isotopes; Carbon; Nitrogen; Wharram Percy
1. Introduction
The duration of breastfeeding and the weaning
process (the introduction of solid foods to an infant’sdiet) have increasingly become a focus of investigationin archaeological populations since the stable isotoperatio of nitrogen ( /afii9829
15N) in body tissues can be used as a
natural tracer to identify the consumption of breast milk[13,18,22,23,28,31,34,35,40,41] . Breastfeeding children
have tissue /afii9829
15N values 2–3‰ higher than their mothers
[14] as a result of the “trophic level e ffect” [32] where
consumer tissue /afii982915N values are elevated by approxi-
mately 2–4‰ over dietary protein. During the weaningprocess, the consumption of supplementary foods resultsin a decline in infant /afii9829
15N values. When a child is
fully weaned (cessation of breastfeeding), its protein/afii982915N values are nearly identical to those of its mother,
assuming similar diets [14].
For isotopic studies, the age of weaning in archaeo-
logical collections has traditionally been estimated bymeasuring the /afii9829
15N of bone collagen (ribs) from di ffer-
ent age classes of infants and children and noting the ageat which the /afii9829
15N values return to those of the adult
population. While this method yields a general time-frame for the duration of breastfeeding, a precise calcu-lation of the introduction of weaning foods is notpossible, as uncertainty exists about the amount of timeneeded for the infant skeleton to fully incorporate theisotopically depleted post-weaning collagen [36,39] .I n
addition, /afii9829
15N values obtained from infants who died
during breastfeeding and weaning need to be interpretedwith caution as these individuals might have been fed
different diets or died of malnutrition, and this has the
potential to alter the expected /afii9829
15N values [20].*Corresponding author. Tel.: +44-0-1865-283-645.
E-mail address: ben.fuller@rlaha.ox.ac.uk (B.T. Fuller).Journal of Archaeological Science 30 (2003) 1673–1684
SCIENCEJournal ofArchaeological
http://www.elsevier.com/locate/jasSCIENCEJournal ofArchaeological
http://www.elsevier.com/locate/jas
0305-4403/03/$ – see front matter /p302003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0305-4403(03)00073-6

The serial sectioning of human skeletal tissues,
enamel and dentine, that are metabolically inert toremodeling [3,15] could circumvent the problems men-
tioned above and allow the study of weaning and dietaryhabits at a more refined level. While a number of studieshave measured bulk isotopic ratios from dentine colla-gen [4,6–8,31,38,41] and noted an o ffset with bone
collagen, little research has been conducted on dentineserial sections. Pioneering research on dentine serialsections from Stellar sea lions showed that isotopicsignals were preserved along annual growth lines andrecorded the dietary and environmental conditions ofthe animal [21]. Drucker et al. [12] used serial sections
from the tooth roots of modern caribou to examineshort-term dietary changes and lichen consumption, anda study of dentine serial sections in modern cattle teethby[5]assessed the duration of weaning and examined a
change from a C
3to C4/C3diet. In this paper, we present
/afii982913C and /afii982915N measurements of human dentine collagen
serial sections from deciduous second molars,permanent canines and third molars. These skeletalsamples are from the medieval site of Wharram Percy,Yorkshire, UK. The dentine collagen results are thencompared to rib collagen /afii9829
13C and /afii982915N values from
the same individual to gain a fuller understanding ofindividual isotopic life histories.
2. Materials
Located in the Yorkshire Wolds about 25 miles
east of the modern city of York, the deserted medievalvillage of Wharram Percy is one of the most thoroughlyexcavated and best preserved English sites of the timeperiod. The Wharram Percy human skeletal collectionconsists of 687 individuals and was chosen for weaningstudies since nearly half of the assemblage representsindividuals under the age of 18. In addition, the individ-uals recovered from the church and churchyard repre-
sent a geographically defined and socially meaningfulpopulation since only those inhabitants of this ruralparish were buried there [26]. The ribs and teeth used in
this research date primarily from the 10th–16th century.
3. Methods
Deciduous second molars ( n=21), permanent third
molars ( n=8), permanent canines ( n=8), and correspond-
ing rib samples were analyzed for this study. Age atdeath for the juveniles sampled was determined usingdental development [33]; for the adults age was esti-
mated using dental wear [27,29] . All teeth studied were
free of caries and were first shot blasted to removesurface debris (soil) and then vertically sectioned in halfwith a low speed diamond saw. One of the tooth halveswas then cut horizontally while the other was retainedfor future analysis. Except in cases of incomplete toothformation, deciduous second molars were sectioned intothree parts corresponding to the crown (defined as thedentine enclosed by the enamel), and the cervical andapical halves of the root. For the permanent thirdmolars, six of the eight teeth were also cut into crown,and cervical and apical halves of the root. The tworemaining third molars (EE36 and EE3) were slightlylarger and were sectioned into four portions correspond-ing to the crown, and cervical, middle and apical thirdsof the root. As a result of their greater length, thepermanent canines were cut into four portions corre-sponding to the crown, and cervical, middle, and apicalthirds of the root, except for G597 which was cut intothree portions comprising the crown and the cervicaland apical root halves. The estimated ages at formationof the dentine serial sections are shown in Tables 1a and
1b. We are aware that tooth dentine does not form
horizontally, but the analytical equipment we usedTable 1a
Estimated age-spans of formation in tooth dentine serial sections
Teeth (3 sections) Approximate timing of dentine development
Crown Cervical root half Apical root half
Deciduous M2 6 months in utero–11 months post-natal 11 months–2 years 2–3 years
Permanent canine 5 months–6 years 6–10 years 10–14 yearsPermanent M3 9.5–13.5 years 13.5–16.7 years 16.7–20 years
Table 1b
Teeth (4 sections) Crown Cervical root third Middle root third Apical root third
Permanent canine 5 months–6 years 6–8.7 years 8.7–11.4 years 11.4–14 years
Permanent M3 9.5–13.5 years 13.5–15.7 years 15.7–17.8 years 17.8–20 years
Note: Dentine development schedule derived from Gustafson and Koch [17], except for third molars, which use data collated from Anderson
et al. [2], Levesque et al. [25]and Garn et al. [16].B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1674

required relatively large samples of collagen (3–5 mg),
and therefore we had to compromise between accuratelysampling dentine along growth increment lines withobtaining enough sample for analysis.
Rib and tooth samples were prepared for isotopic
analysis at the Research Laboratory for Archaeologyand the History of Art, University of Oxford. Collagenwas extracted from the rib samples and the teeth serialsections following a modified Longin procedure, asoutlined in detail elsewhere [30]. As the crown serial
sections were encased in enamel, the demineralizationprocess was much slower and took approximately 4–7days to complete. The resultant solids from the ribs andthe teeth were gelatinized in pH 3 water at 70 (C for
48 h and the final solutions filtered and lyophilized. Thisprocess resulted in the extraction of a number of boneand tooth proteins, but the majority of the extractedmaterial was bone and dentine collagen. The collagenwas placed in tin capsules and combusted to CO
2and N2
in an automated carbon and nitrogen analyzer (CarloErba) coupled to a continuous-flow isotope ratio moni-toring mass spectrometer (PDZ Europa Geo 20/20).Replicate measurement errors are less than /p50.2‰ for
/afii9829
13C and/p50.3‰ for /afii982915N. All results are reported in
units of per mil (‰) with /afii982913C relative to VPDB (Vienna
Pee Dee Belemnite), and with /afii982915N relative to AIR
(Ambient Inhalable Reservoir).
4. Results and discussion
4.1. Deciduous dentition
The stable isotope values for the deciduous second
molar serial sections and corresponding ribs from thesame individual are presented in Table 2 . Deciduous
second molars were chosen to study breastfeeding andweaning since the dentine formation in this tooth spansthe period from about 6 months in-utero until rootcompletion at approximately 3 years of age ( Table 1a ).
Complete serial sectioning (crown and cervical/apicalhalves of the root) of teeth under 3 years of age was notTable 2
/afii982913Ca n d /afii982915N results for dentine serial sections from deciduous
second molars and corresponding rib samples. Measurementerrors=/p50.2‰ for /afii9829
13Ca n d/p50.3‰ for /afii982915N
Sample Serial section Age (years) /afii982913C /afii982915NC : N
G522 crown 1.3 /p119.0 12.1 3.17
G522 rib 1.3 /p119.8 11.6 3.20
NA37 crown 1.3 /p119.6 10.0 3.20
NA37 rib 1.3 /p120.1 9.8 3.23
G327 crown 1.5 /p118.6 12.2 3.10
G327 cervical root 1.5 /p118.6 12.9 3.11
G327 rib 1.5 /p119.1 12.3 3.17
NA28 crown 1.5 /p118.5 13.3 3.20
NA28 rib 1.5 /p118.9 13.1 3.30
G430 crown 1.8 /p119.6 11.1 3.20
G430 cervical root 1.8 /p119.8 9.4 3.05
G430 rib 1.8 /p120.4 9.8 3.18
WCO72 crown 2 /p119.0 9.8 3.14
WCO72 cervical root 2 /p119.0 9.3 3.22
WCO72 rib 2 /p119.6 8.5 3.25
G339 crown 2.5 /p119.6 12.3 3.14
G339 cervical root 2.5 /p119.9 10.0 3.04
G339 apical root 2.5 /p119.7 9.3 3.13
G339 rib 2.5 /p120.4 8.1 3.21
G363 crown 2.5 /p118.7 11.6 3.15
G363 cervical root 2.5 /p119.0 10.6 3.17
G363 rib 2.5 /p120.0 8.8 3.24
NA79 crown 3 /p119.0 10.0 2.99
NA79 cervical root 3 /p119.1 10.8 3.18
NA79 apical root 3 /p119.6 9.6 3.23
NA79 rib 3 /p120.2 8.5 3.23
G576 crown 3.3 /p119.4 11.6 3.05
G576 cervical root 3.3 /p119.5 11.1 3.18
G576 rib 3.3 /p120.5 9.3 3.22
WCO97 crown 5 /p118.3 11.5 3.16
WCO97 cervical root 5 /p118.5 11.1 3.16
WCO97 apical root 5 /p118.6 8.4 3.17
WCO97 rib 5 /p119.4 7.5 3.17
G614 crown 5.5 /p119.1 11.3 3.12
G614 cervical root 5.5 /p119.4 10.0 3.18
G614 rib 5.5 /p120.5 8.5 3.14
NA30 crown 6 /p118.6 12.2 3.16
NA30 cervical root 6 /p119.2 10.6 3.19
NA30 apical root 6 /p119.4 9.7 3.16
NA30 rib 6 /p120.7 8.0 3.35
G424 crown 6.5 /p118.8 12.1 3.14
G424 cervical root 6.5 /p119.1 11.6 3.19
G424 apical root 6.5 /p119.7 9.9 3.21
G424 rib 6.5 /p119.9 8.4 3.20
NA23 crown 7 /p119.1 11.6 3.18
NA23 cervical root 7 /p119.4 10.9 3.19
NA23 apical root 7 /p119.5 8.6 3.19
NA23 rib 7 /p119.8 8.4 3.21
EE65 crown 7.5 /p119.3 11.0 3.17
EE65 cervical root 7.5 /p119.8 9.9 3.21
EE65 apical root 7.5 /p120.0 9.1 3.22
EE65 rib 7.5 /p120.4 7.3 3.08
EE66 crown 8.5 /p118.7 12.8 3.19
EE66 cervical root 8.5 /p119.4 11.7 3.18
EE66 apical root 8.5 /p119.9 9.9 3.25
EE66 rib 8.5 /p120.1 10.1 3.23
WCO140 crown 9 /p118.6 12.4 3.16
WCO140 cervical root 9 /p118.8 12.2 3.21
WCO140 apical root 9 /p119.5 8.5 3.06
WCO140 rib 9 /p119.9 8.8 3.22Table 2 ( continued )
Sample Serial section Age (years) /afii982913C /afii982915NC : N
G500 crown 9.5 /p118.9 12.0 3.17
G500 cervical root 9.5 /p119.6 10.4 3.17
G500 apical root 9.5 /p119.7 9.6 3.19
G500 rib 9.5 /p120.1 8.7 3.25
EE72 crown 10 /p119.4 12.4 3.21
EE72 cervical root 10 /p119.6 10.8 3.20
EE72 apical root 10 /p120.2 9.5 3.31
EE72 rib 10 /p120.1 9.0 3.32
G658 crown 11 /p118.1 10.3 3.14
G658 cervical root 11 /p118.6 9.9 3.18
G658 apical root 11 /p120.0 9.0 3.54
G658 rib 11 /p119.7 7.9 3.18B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1675

possible due to incomplete tooth formation. Thus, for
these teeth the crown was separated from as much of theroot as had formed. All the teeth and ribs sampled hadcollagen yields between 4% and 15% and had C/N ratioswithin the expected range for collagen of 2.9 to 3.6 [10],
indicating good preservation of the skeletal material.
The /afii9829
13C values of the deciduous second molars are
plotted in Fig. 1 a. For nearly all individuals, the follow-
ing13C enrichment pattern is observed: crown>cervical
half of root>apical half of root>rib. Two infants (G327,WCO72) have crown and cervical root portion /afii9829
13C
values that are essentially equal, and the apical half ofthe root from G339 is slightly enriched compared tothe cervical portion but this di fference is within the
analytical error ( /p50.2‰) for carbon. In addition,
the two oldest children (EE72, G658), aged 10 and11 years, have rib /afii9829
13C values that are more enriched
than the apical halves of the deciduous second molarroots by 0.1‰ and 0.3‰, respectively. Again, these /afii9829
13C
values are within the analytical error, and thus it isimpossible to interpret the results as reflecting slightvariations in dietary carbon in later life for theseindividuals.
The /afii9829
15N values for the deciduous second molars are
plotted in Fig. 1 b. The15N enrichment pattern was the
same as for carbon (crown>cervical half of root>apicalhalf of root>rib). That the crown dentine generallyshows greater
15N enrichment than the cervical half of
the root indicates that a relatively greater proportion ofthe crown dentine formed during the time the infant wasbreastfeeding, despite the fact that it incorporates asmall amount of dentine which formed before birth.Given the developmental timing of these parts of thetooth ( Table 1a ), this pattern is as expected if, as inferredfrom previous work [28,31] , breastfeeding was normally
discontinued after about 18 months post-partum as asignificant part of the cervical tooth root sample wouldhave formed after this. There were some variations inthis general pattern. Individual G430 (1.5–2 years) has arib/afii9829
15N value that is elevated by 0.4‰ compared to the
cervical portion of the deciduous second molar root, andtwo children (EE66, WCO140) have rib values that areslightly enriched in
15N. In particular, one infant (G327)
aged about 1.5 years has a /afii982915N value for the cervical
portion of the deciduous second molar root that is 0.8‰elevated over that of the crown. The higher value forthe cervical part of the (incomplete) tooth root is asexpected if this individual was breast-fed until time ofdeath, as under these circumstances all of the (incom-pletely formed) root sample would have formed duringthe period this individual was breastfeeding, whereas,as discussed above, a small proportion of the crowndentine forms prior to birth and so would be relativelydepleted in
15N.
The 3 year old (NA79) shows the same15N enrich-
ment as G327, cervical half of root>crown. However, inthis case the low crown /afii9829
15N value (10‰) points to the
fact that this child was not exclusively breastfed or haddifficulty breastfeeding soon after birth. This low con-
sumption of breast milk in the early diet appears to havebeen increased in later life since the cervical half of theroot is enriched in
15N by 0.9‰ compared to the crown.
The crown /afii982915N values for NA37 and WCO72 are 10‰
and 9.8‰, respectively ( Fig. 1 b). These results are well
below the mean of the other crown sections(11.6/p51.0‰; mean /p5SD) and could suggest that these
individuals were not or only minimally breastfed. Asbreastfeeding has been proven to confer importantFig. 1. /afii982913C (a) and /afii982915N (b) values for dentine serial sections from deciduous second molars and ribs plotted for individual samples. Also shown
is the adult mean /p5SD.B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1676

nutritional and health benefits to infants such as a
decreased risk of infection, diarrhea, and mortality[9,11,24] , this possible lack of breastfeeding for NA37
and WCO72 might have contributed to their earlydeaths. The 11 year old (G658) also displays a lowcrown /afii9829
15N value (10.3‰) ( Fig. 1 b). While the di fference
in/afii982915N between the crown and the rib (2.4‰) indicates
that this individual was breastfed, the low crown valuessuggests that some supplementary foods were added tothe diet at an earlier age thereby diluting the full
15N
enrichment of the breast milk. It is also possible that thelow crown /afii9829
15N value reflects the fact that the diet and
thereby the breast milk from the mother of thisinfant was lower in /afii9829
15N compared to the rest of the
population.
Most studies of the duration of breastfeeding in past
populations use /afii982915N values from bone collagen. Such
studies, involve determining /afii982915N values from infants
and children who died over a range of di fferent ages.
Weaning practices are therefore reconstructed fromchildren who failed to survive. Breastfeeding practices asreconstructed from those who died in infancy may notwholly typify regimes for those who did survive infancybut rather they may identify unsuccessful strategies thatincreased risk of infant death. To a limited extent, thecurrent data permit investigation of whether those whodied in infancy at Wharram Percy generally experienceddifferent breast-feeding practices to those who survived
into later childhood.
Dentine formation in the crown of the deciduous
second molar begins at about 6 months in-utero, and thecrown is complete by about 11 months after birth ( Table
1a). Previous research [31] has determined that breast-
feeding continued for 1–2 years after birth in this
population. Thus, most of the second molar crowndentine formed at a time during which most infants werebeing breastfed at Wharram Percy. Our hypothesis isthat if infant feeding/weaning practices di ffered between
those who died aged 2 or under and those who survivedbeyond 2 years (i.e. beyond what we believe to be thenormal duration of breastfeeding), then the /afii9829
13C and
/afii982915N values of the crown dentine might also di ffer
between these two age groups.
Carbon and nitrogen stable isotope results from
deciduous second molars for 6 individuals aged 2 orunder and 15 individuals aged over 2 years are listed inTable 2 . Through the use of t-tests, no significant
difference in either /afii9829
13Co r /afii982915N was determined for the
two age groups ( Table 3 ). Although the numbers are
small, the current data o ffer no support for the notion
that, at Wharram Percy, infant diet among those whodied in infancy generally di ffered from that of those
who survived.
The mean /afii9829
13C and /afii982915N results for the deciduous
second molar serial sections and ribs are presented inFig. 2 . This data is only plotted for individuals agedgreater than two years so that the magnitude of the
dietary transition from breast milk to weaning foodscan be assessed. A depletion pattern (mean /p5SD) with a
very strong linear correlation (R
2=0.976) is observed
(crown>cervical half of root>apical half of root>rib)in both carbon ( /p118.9/p50.4‰>/p119.3/p50.4‰>/p1
19.7/p50.4‰>/p120.1/p50.4‰) and nitrogen (11.7 /p5
0.8‰>10.8 /p50.7‰>9.3 /p50.5‰>8.5 /p50.7‰). This dualTable 3
/afii982913Ca n d /afii982915N values (‰) in those aged 2 years or less ( N=6) and
those aged over 2 years at death ( N=15)
Age group Mean SD
/afii982913C <=2 years 19.1 0.5
>2 years 18.9 0.4
/afii982915N <=2 years 11.4 1.4
>2 years 11.7 0.8
t-tests <=2 years vs. >2 years: /afii982913C,t=0.7; /afii982915N,t=0.5. In each
case Pw0.2 at 19 degrees of freedom, i.e. non-significant.
Fig. 2. /afii982913Ca n d /afii982915N results (mean /p5SD) for all deciduous second
molar dentine serial sections: crown, cervical root half, apical root halfand rib samples from individuals greater than 2 years old.
Fig. 3. /afii982913Ca n d /afii982915Nd ifferences between the deciduous second molar
crowns and ribs plotted against age of death.B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1677

Table 4
/afii982913Ca n d /afii982915N results for dentine serial sections from permanent canines (C), permanent third molars (M3) and corresponding rib samples.
Measurement errors= /p50.2‰ for /afii982913C and/p50.3‰ for /afii982915N
Sample Serial section Age (years) Sex /afii982913C /afii982915NC : N
EE36 (C) crown 18–21 Female /p119.0 9.9 3.18
EE36 (C) cervical root 18–21 Female /p119.3 8.9 3.19
EE36 (C) middle root 18–21 Female /p118.3 8.4 3.17
EE36 (C) apical root 18–21 Female /p118.8 8.5 3.24
EE36 (M3) crown 18–21 Female /p118.7 9.1 3.17
EE36 (M3) cervical root 18–21 Female /p119.0 7.9 3.13
EE36 (M3) middle root 18–21 Female /p119.0 8.0 3.16
EE36 (M3) apical root 18–21 Female /p118.6 8.4 3.19
EE36 rib 18–21 Female /p119.2 9.0 3.29
CN28 (C) crown 22–24 Female /p119.4 8.0 3.03
CN28 (C) cervical root 22–24 Female /p119.6 8.1 3.15
CN28 (C) middle root 22–24 Female /p119.6 7.3 3.14
CN28 (C) apical root 22–24 Female /p119.2 7.8 3.19
CN28 (M3) crown 22–24 Female /p119.4 8.1 3.16
CN28 (M3) cervical root 22–24 Female /p119.2 8.7 3.21
CN28 (M3) apical root 22–24 Female /p119.0 9.2 3.20
CN28 rib 22–24 Female /p119.4 9.5 3.26
G597 (C) crown 22–25 Female /p119.4 9.3 3.18
G597 (C) cervical root 22–25 Female /p119.4 7.6 3.14
G597 (C) apical root 22–25 Female /p119.6 7.7 3.19
G597 (M3) crown 22–25 Female /p119.5 8.2 3.16
G597 (M3) cervical root 22–25 Female /p119.1 8.2 3.18
G597 (M3) apical root 22–25 Female /p119.0 8.9 3.22
G597 rib 22–25 Female /p119.9 7.8 3.28
CN2 (C) crown 22–25 Female /p119.6 11.7 3.06
CN2 (C) cervical root 22–25 Female /p119.3 11.2 3.16
CN2 (C) middle root 22–25 Female /p119.5 10.6 3.19
CN2 (C) apical root 22–25 Female /p119.3 10.6 3.16
CN2 (M3) crown 22–25 Female /p119.5 11.2 3.16
CN2 (M3) cervical root 22–25 Female /p119.3 10.0 3.13
CN2 (M3) apical root 22–25 Female /p119.3 10.5 3.26
CN2 rib 22–25 Female /p119.4 10.5 3.13
EE3 (C) crown 25–35 Female /p120.0 11.0 3.19
EE3 (C) cervical root 25–35 Female /p119.7 8.6 3.15
EE3 (C) middle root 25–35 Female /p119.5 8.0 3.17
EE3 (C) apical root 25–35 Female /p119.5 8.0 3.17
EE3 (M3) crown 25–35 Female /p119.1 9.4 3.13
EE3 (M3) cervical root 25–35 Female /p120.7 9.0 3.12
EE3 (M3) middle root 25–35 Female /p119.7 10.8 3.23
EE3 (M3) apical root 25–35 Female /p120.0 8.6 3.39
EE3 rib 25–35 Female /p119.8 10.3 3.29
NA59 (C) crown 35–45 Male /p119.5 9.8 3.16
NA59 (C) cervical root 35–45 Male /p119.8 9.2 3.17
NA59 (C) middle root 35–45 Male /p119.6 9.3 3.23
NA59 (C) apical root 35–45 Male /p120.0 8.6 3.18
NA59 (M3) crown 35–45 Male /p119.9 9.2 3.15
NA59 (M3) cervical root 35–45 Male /p119.6 9.5 3.20
NA59 (M3) apical root 35–45 Male /p119.8 9.9 3.20
NA59 rib 35–45 Male /p120.1 8.9 3.12
EE67 (C) crown 30–50 Male /p119.6 8.3 3.07
EE67 (C) cervical root 30–50 Male /p119.1 9.3 3.19
EE67 (C) middle root 30–50 Male /p119.3 9.4 3.18
EE67 (C) apical root 30–50 Male /p119.2 9.5 3.20
EE67 (M3) crown 30–50 Male /p119.4 9.8 3.17
EE67 (M3) cervical root 30–50 Male /p119.2 9.5 3.14
EE67 (M3) apical root 30–50 Male /p118.5 10.4 3.17
EE67 rib 30–50 Male /p119.7 9.1 3.27
G746 (C) crown 30–50 Female /p119.4 9.6 3.15
G746 (C) cervical root 30–50 Female /p119.2 9.3 3.16
G746 (C) middle root 30–50 Female /p119.3 8.8 3.12
G746 (C) apical root 30–50 Female /p119.6 8.7 3.18B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1678

sequence of enrichment indicates that infants display
both a carbon and nitrogen trophic level shift duringbreastfeeding. The mean
15N enrichment of the crown
dentine over the ribs is 3.2 /p50.8‰ ( Fig. 3 ). This ob-
served 3.2‰ enrichment is in good agreement with theresults of previous studies of infant bone collagen[13,18,22,23,34] . In addition, the mean
13C enrichment
of the crown dentine over the ribs is 1.2 /p50.4‰ ( Fig. 3 ).
Enrichment in13C (0.5–1.4‰) has been previously noted
in infant collagen by other researchers [13,22,23,31,41]
but never with the consistency and clarity of the presentresults ( Fig. 3 ). Since many of these past studies were
conducted in regions where there is an input of both C
3
and C4plants, this increase in /afii982913C has sometimes been
interpreted as infants being weaned onto C4dietary
components such as maize gruel or milk from animalsfed millet. In addition, the only longitudinal study ofmodern breastfeeding infants did not find a change in/afii9829
13C during suckling [14]. Thus, there has been confu-
sion about the existence of this carbon trophic leveleffect and its exact magnitude for breastfeeding infants.
The data from Wharram Percy clearly show a change
in both /afii9829
13C and /afii982915N from the crown to the apical root
of the second molar teeth. Since Wharram Percy had noknown C
4plants that could be used as supplements to
breast milk during weaning, the data presented here(Fig. 3 ) strongly suggests that the 1.2‰ enrichment in
13C is entirely the result of a carbon trophic level e ffect
[32] of breastfeeding infants. Thus, it is likely that the
observed13C enrichment in infant skeletons from pre-
vious studies [13,22,23,41] are the result of a breast milk
trophic level e ffect and not due to the input of C4
weaning foods. Ultimately, this debate over the existence
of a13C enrichment in the protein of breastfeeding
infants will have to be resolved by modern humanlongitudinal studies, which are currently underway. Pre-liminary results do indeed suggest a 1‰
13C enrichment
in the hair and nails of breastfeeding infants comparedto maternal /afii9829
13C values (Fuller et al., in prep.).
4.2. Permanent dentition
The results for the permanent canines, permanent
third molars and ribs are presented in Table 4 . The
timing of the development of these teeth means thatthe majority of the dentine in the canine, and all of thedentine in the third molar, was laid down after thecessation of breastfeeding. As there is significant overlap
in the formation of the permanent canines and perma-
nent third molars, it is expected that the isotopic signa-tures will track each other during these periods, and thisis generally observed in the results ( Fig. 4 a–h). These
specific teeth were chosen as a pilot study to examine thepossibility of tracking changing individual dietary habitsduring childhood and adolescence. All the teeth and ribsexamined had collagen yields between 4% and 15% andhad C/N ratios within the expected range for collagen of2.9 to 3.6 [10].
The /afii9829
13C results for the permanent canines, perma-
nent third molars, and ribs from the same individualsare plotted in Fig. 4 a–h. Unlike the deciduous second
molars, no discrete patterning was observed between theserial sections of the canines and the third molars for/afii9829
13C. The majority of the collagen /afii982913C values ( /p119‰
to/p120‰) indicate a diet based largely on C3plant and
terrestrial animal protein with little input from marineprotein sources [36]. As there are only small variations in
the /afii9829
13C values, the sources of dietary carbon were
relatively constant during the lifetime of these individ-uals. However, the female (EE36) shows a
13C enrich-
ment in her middle ( /p118.3‰) and apical ( /p118.8‰)
canine roots and in her third molar crown ( /p118.7‰)
which could suggest some marine food consumptionduring late childhood or early adolescence ( Fig. 4 a). In
addition, the male (EE67) also displays increases in both/afii9829
13C(/p118.5‰) and /afii982915N (10.4‰) for the apical half of
the third molar root ( Fig. 4 g). Since /afii982913C and /afii982915N
values are elevated in marine ecosystems [36], this dual
enrichment might be the result of a small input ofmarine protein to the diet of this male during his lateteenage years. As the consumption of marine food wasminor in the Wharram Percy population, it is possiblethat these individuals travelled and lived near the coastduring this time. The crown and the cervical third of thethird molar root from EE3 show the largest variation in/afii9829
13C (1.6‰), suggesting that the diet of this individual
was more variable during early adolescence ( Fig. 4 e).
Fig. 4 a–h also illustrate the /afii982915N values for perma-
nent canines, permanent third molars, and ribs fromthe same individuals. In contrast to the /afii9829
13C results, the
/afii982915N values show much more variation in both the
canines and the third molars. This reflects the di fferent
dietary protein sources where the enrichment in15No f
collagen is primarily the result of an increase in animalTable 4 ( continued )
Sample Serial section Age (years) Sex /afii982913C /afii982915NC : N
G746 (M3) crown 30–50 Female /p119.8 9.2 3.17
G746 (M3) cervical root 30–50 Female /p119.7 8.7 3.18
G746 (M3) apical root 30–50 Female /p119.2 9.0 3.17
G746 rib 30–50 Female /p119.4 9.6 3.23B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1679

B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1680

Fig. 4 (a–f)B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1681

protein consumption versus plant protein consumption
but has also been attributed to physiological andenvironmental factors [1,20,36,37] . Six of the canine
crowns (EE36, G597, CN2, EE3, NA59, G746) areenriched in
15N over the cervical halves of the roots,
which is likely the result of breastfeeding. This is asexpected as canine dentine development starts at about 5months of age ( Table 1a ). Five of the rib samples (EE36,
CN28, CN2, EE3, G746) have /afii9829
15N values nearly equal
to or greater than apical portions of the third molarroots ( Fig. 4 a, b, d, e, h), illustrating for these individ-
uals that the consumption of animal protein was con-stant or had increased during the later part of life as ribcollagen reflects the diet from approximately the last 5 to10 years of life [39]. In contrast, both of the males
studied (NA59, EE67) show a decrease in /afii9829
15N from the
apical halves of the third molar roots to the ribs by 1.1‰and 1.3‰, respectively ( Fig. 4 f, g). This depletion in15N
can likely be attributed to decreased animal proteinconsumption in later life. With such a small study (8individuals) over such a large period of time (approxi-mately 600 years), the present results clearly may not berepresentative of the entire Wharram Percy population,yet they show that the dietary habits of these inhabitantswere not static but changed with age. Analyses of dentalmaterial from further skeletons would be necessary toestablish if these dietary trends were representative ofthis community as a whole.
5. Conclusions
We have conducted a study on the deciduous and
permanent dentine serial sections of human teeth, usingmaterial from the medieval village site of WharramFig. 4. (a–h). /afii982913C and /afii982915N values for dentine serial sections from permanent canines (Can), permanent third molars (M3) and ribs from the same
individuals.B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1682

Percy. For deciduous second molars and ribs, the
following isotopic enrichment pattern was observed forboth
13C and15N: crown>cervical part of root>apical
part of root>ribs. This pattern is the result of decreasedbreast milk consumption and an increased intake ofweaning foods. After the age of weaning (2 years old),the mean di fference between the dentine crowns and the
ribs was 1.2 /p50.4‰ for /afii9829
13C and 3.2 /p50.8‰ for /afii982915N.
While this13C enrichment has been seen elsewhere
[13,22,23,41] , these results are unique as they can be
clearly attributed to a trophic level e ffect of breast-
feeding infants, as there were no known C4plants at
Wharram Percy.
The comparison of the deciduous second molar
crowns from individuals who died before weaning withthose who survived past the age of weaning (about 2years old) failed to find a dietary di fference between the
two age groups. While the sample size was small, thislack of a di fferent feeding regime between those who
survived beyond infancy/early childhood and those whodid not suggests that earlier inferences concerning infantfeeding patterns [28,31] based on analyses of non-
survivors are likely to be a reasonable indication ofusual feeding practices in this population rather thandefining some abnormal, unsuccessful strategy. Theserial sectioning of the permanent dentition indicatedthat there was little change in the sources of dietarycarbon during life, but two individuals seem to showevidence of minor inputs of marine protein in their diets.The /afii9829
15N results were more variable with some individ-
uals likely consuming equal or more animal protein inlater life.
Finally, this research has demonstrated the potential
for using isotopic results from human dentine serialsections to deduce individual dietary habits. Futurework in this area should focus on the sectioning of teethalong the lines of growth for dentine. As the formationrates for teeth are well characterized [19], the meticulous
sampling of dentine along the lines of growth has theability to determine weaning patterns and dietary habitsat the refined level of the individual. Ideally, thesedentine /afii9829
13C and /afii982915N measurements should be com-
bined with enamel carbonate /afii982913C and /afii982918O results so
that fuller isotopic life histories of past populations canbe developed and interpreted.
Acknowledgements
Robert Hedges is thanked for his discussion of the
results and the conclusions. Theya Molleson is thankedfor her discussion of breastfeeding and weaning prac-tices in England. David Harris is thanked for providingstatistical help. This manuscript was greatly improvedby the editing of Christina Tsou and by the comments oftwo anonymous reviewers. We would also like to thankKen Neal for his technical expertise and assistance in
preparing and measuring some of the samples.
References
[1] S.H. Ambrose, E ffects of diet, climate and physiology on nitrogen
isotope abundances in terrestrial food webs, Journal of Archaeo-logical Science 18 (1991) 293–317.
[2] D.L. Anderson, G.W. Thompson, F. Popovich, Age of
Attainment of Mineralisation Stages of the PermanentDentition, Journal of Forensic Sciences 21 (1976) 191–200.
[3] J.L. Bada, In vivo racemization in animal proteins, Methods in
Enzymology 106 (1984) 98–115.
[4] J.L. Bada, R.O. Peterson, A. Schimmelmann, R.E.M. Hedges,
Moose teeth as monitors of environmental isotopic parameters,Oecologia 82 (1990) 102–106.
[5] M. Balasse, H. Bocherens, A. Mariotte, S.H. Ambrose, Detection
of dietary changes by Intra-tooth carbon and nitrogen isotopicanalysis: An experimental study of dentine collagen of cattle ( Bos
taurus ), Journal of Archaeological Science 28 (2001) 235–245.
[6] H. Bocherens, M. Fizet, A. Mariotti, Is collagen from teeth or
bone equivalent for isotopic (
13C,15N) diet investigations? Fifth
North American Paleontological Convention (NAPC V),Chicago. Special Publication Series of the Paleontological Society6 (1992) 30.
[7] H. Bocherens, M. Fizet, A. Mariotte, Diet, physiology and
ecology of fossil mammals as inferred from stable carbon andnitrogen isotope biogeochemistry: implications for Pleistocenebears, Palaeogeography, Palaeoclimatology, Palaeoecology 107(1994) 213–225.
[8] H. Bocherens, M. Fogel, N. Tuross, M. Zeder, Trophic structure
and climatic information from isotopic signatures in Pleistocenecave fauna of southern England, Journal of ArchaeologicalScience 22 (1995) 327–340.
[9] A.S. Cunningham, D.B. Jelli ffe, E.F.P. Jelli ffe, Breastfeeding and
health in the 1980s: a global epidemiological review, Journal ofPediatrics 118 (1991) 659–666.
[10] M.J. DeNiro, Post-mortem preservation and alteration of in vivo
bone collagen isotope ratios in relation to paleodietaryreconstruction, Nature 317 (1985) 806–809.
[11] K.A. Dettwyler, C. Fishman, Infant feeding practices and growth,
Annual Review of Anthropology 21 (1992) 171–204.
[12] D. Drucker, H. Bocherens, A. Pike-Tay, A. Mariotti, Isotopic
tracking of seasonal dietary change in dentine collagen:preliminary data from modern caribou, Comptes Rendus De LAcademie Des Sciences Serie II Fascicule A—Sciences De LaTerre Et Des Planetes 333 (2001) 303–309.
[13] T.L. Dupras, H.P. Schwarcz, S.I. Fairgrieve, Infant feeding and
weaning practices in Roman Egypt, American Journal of PhysicalAnthropology 115 (2001) 204–211.
[14] M.L. Fogel, N. Tuross, D. Owsley, Nitrogen isotope tracers of
human lactation in modern and archaeological populations,Annual Report of the Director, Geophysical Laboratory,1988–1989, Carnegie Institution of Washington, Washington,DC, 1989, pp. 111–116.
[15] J.P. Gage, M.J.O. Francis, J.T. Tri ffitt, Collagen and Dental
Matrices, Wright, London, 1989.
[16] S.M. Garn, A.B. Lewis, B. Bonne, Third molar formation and its
development course, The Angle Orthodonist 32 (1962) 270–279.
[17] G. Gustafson, G. Koch, Age estimation up to 16 years of age
based on dental development, Odontologisk Revy 25 (1974)297–306.
[18] D.A. Herring, S.R. Saunders, M.A. Katzenberg, Investigating
the weaning process in past populations, American Journal ofPhysical Anthropology 105 (1998) 425–439.B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1683

[19] S. Hillson, Dental Anthropology, Cambridge University Press,
Cambridge, 1996.
[20] K.A. Hobson, R.T. Alisauskas, R.G. Clark, Stable nitrogen
isotope enrichment in avian tissues due to fasting and nutritionalstress: implications for isotopic analyses of diet, Condor 95 (1993)388–394.
[21] K.A. Hobson, J.L. Sease, Stable isotope analyses of tooth annuli
reveal temporal dietary records: an example using Stellar sealions, Marine Mammal Science 14 (1998) 116–129.
[22] M.A. Katzenberg, S.R. Saunders, W.R. Fitzgerald, Age di ffer-
ences in stable carbon and nitrogen isotope ratios in a populationof prehistoric maize horticulturists, American Journal of PhysicalAnthropology 90 (1993) 267–281.
[23] M.A. Katzenberg, S. Pfei ffer, Nitrogen isotope evidence for
weaning age in a nineteenth century Canadian skeletal sample, in:A.L. Grauer (Ed.), Bodies of Evidence: Reconstructing Historythrough Skeletal Analysis, New York, Wiley-Liss, Cambridge,1995, pp. 221–235.
[24] M.A. Katzenberg, D.A. Herring, S.R. Saunders, Weaning and
infant mortality: evaluating the skeletal evidence, Yearbook ofPhysical Anthropology 39 (1996) 177–199.
[25] G-Y. Levesque, A. Demirjan, R. Tanguay, Sexual dimorphism in
the development, emergence and agenesis of the mandibular thirdmolar, Journal of Dental Research 60 (1981) 1735–1741.
[26] S.A. Mays, The human remains, in: A. Clark (Ed.), Wharram
Percy: The Churchyard. Society for Mediaeval Archaeology (inpreparation).
[27] S. Mays, C. de la Rua, T. Molleson, Molar crown height as a
means of evaluating existing wear scales of estimating age atdeath in human skeletal remains, Journal of ArchaeologicalScience 22 (1995) 659–670.
[28] S.A. Mays, M.P. Richards, B.T. Fuller, Bone stable isotope
evidence for infant feeding in mediaeval England, Antiquity 76(2002) 654–656.
[29] A.E.W. Miles, The dentition in the assessment of individual age in
skeletal material, in: D.R. Brothwell (Ed.), Dental Anthropology,Permamon Press, Oxford, 1963, pp. 191–209.
[30] M.P. Richards, R.E.M. Hedges, Stable isotope evidence for
similarities in the types of marine foods used by Late Mesolithichumans at sites along the Atlantic coast of Europe, Journal of
Archaeological Science 26 (1999) 717–722.
[31] M.P. Richards, S. Mays, B.T. Fuller, Stable carbon and nitrogen
isotope values of bone and teeth reflect weaning age at theMediaeval Wharram Percy Site, Yorkshire, U.K, AmericanJournal of Physical Anthropology 119 (2002) 205–210.
[32] M. Schoeninger, M.J. DeNiro, Nitrogen and carbon isotopic
composition of bone collagen from marine and terrestrialanimals, Geochimica et Cosmochimica Acta 48 (1984) 625–639.
[33] I. Schour, M. Massler, The development of the human dentition,
Journal of the American Dental Association 28 (1941) 1153–1160.
[34] M.R. Schurr, Stable nitrogen isotopes as evidence of age at
weaning at the Angel Site: A comparison of isotopic and demo-graphic measures of weaning age, Journal of ArchaeologicalScience 24 (1997) 919–927.
[35] M.R. Schurr, Using stable nitrogen isotopes to study weaning
behavior in past populations, World Archaeology 30 (1998)327–342.
[36] H.P. Schwarcz, M.J. Schoeninger, Stable isotopic analyses in
human nutritional ecology, Yearbook of Physical Anthropology34 (1991) 283–321.
[37] H.P. Schwarcz, T.L. Dupras, S.I. Fairgrieve,
15N-enrichment in
the Sahara: in search of a global relationship, Journal Archaeo-logical Science 26 (1999) 629–636.
[38] J.C. Sealy, R. Armstrong, C. Schrire, Beyond lifetime averages:
tracing life histories through isotopic analysis of di fferent calcified
tissues from archaeological human skeletons, Antiquity 69 (1995)290–300.
[39] M.J. Stenhouse, M.S. Baxter, The uptake of bomb
14Ci n
humans, in radiocarbon dating, in: R. Berger, H.E. Suess (Eds.),Radiocarbon Dating, University of California Press, Berkeley,1979, pp. 324–341.
[40] L.E. Wright, H.P. Schwarcz, Stable carbon and oxygen isotopes
in human tooth enamel: identifying breast-feeding in prehistory,American Journal of Physical Anthropology 106 (1998) 1–18.
[41] L.E. Wright, H.P. Schwarcz, Correspondence between stable
carbon, oxygen and nitrogen isotopes in human tooth enamel anddentine: Infant diets at Kaminaljuyu, Journal of ArchaeologicalScience 26 (1999) 1159–1170.B.T. Fuller et al. / Journal of Archaeological Science 30 (2003) 1673–1684 1684