Journal of Archaeological Science (2001) 28, 12471255

doi:10.1006/jasc.2001.0698, available online at http://www.idealibrary.com on

Isotopic Comparison of Hair, Nail and Bone:

Modern Analyses
T. C. OConnell and R. E. M. Hedges
Research Laboratory for Archaeology and the History of Art, University of Oxford, 6 Keble Road,
Oxford OX1 3QJ, U.K.

M. A. Healey
Dept of Archaeological Science, University of Bradford, Bradford, BD7 1DP and Chemical Treatment Unit,
Surrey Police Headquarters, Guildford, Surrey, GU3 1HG

A. H. R. W. Simpson
Department of Orthopaedic Surgery, University of Edinburgh, Princess Margaret Rose Hospital, Fairmilehead,
Edinburgh, EH10 7ED
(Received 15 September 2000, revised manuscript accepted 8 February 2001)
This paper presents a comprison of the isotopic values of eight pairings of hair keratin and bone collagen and 12
pairings of hair keratin and nail keratin taken from living humans resident in the U.K., with the aim of examining
whether modern human isotopic data can be directly compared to archaeological isotopic data.
Results showed that bone collagen was enriched relative to hair keratin from the same individual by +14 in
13C
and +086 in
15N, with some small degree of variability. Isotopic comparison of hair keratin and nail keratin from
the same individual showed that there is no significant dierence between hair and nail keratin in
13C, but that nail
keratin is on average +065 more enriched in
15N than hair.
Dierences in amino acid composition between hair keratin and bone collagen may account for the carbon isotopic
dierences between the two proteins, and there is no significant overall carbon isotopic dierence between hair and nail.
However there are significant isotopic dierences for nitrogen in the two pairings, that dierences in amino acid
composition and turnover times cannot explain. We suggest that these results indicate that constancy of isotopic values
between tissues, even for similar proteins, cannot be taken for granted.
 2001 Academic Press
Keywords: CARBON, NITROGEN, DIET, PROTEIN, COLLAGEN, KERATIN, ARCHAEOLOGY.

Introduction

he isotopic analysis of archaeological human

tissue is often the only quantitative and objective technique available for the reconstruction
of palaeodiet (Schwarcz & Schoeninger, 1991). In
principle, carbon and nitrogen isotopic values can
reveal the position of an individual within the food
chain, the types of plants consumed and whether the
diet was terrestrial or marine-based (Ambrose, 1993).
At present, the interpretation of archaeological human
isotopic data is largely based on comparisons with
archaeological faunal isotopic data and other archaeological human populations, and on theories developed
via laboratory-based animal studies (Ambrose & Norr,
1993; DeNiro & Epstein, 1978, 1981; Tieszen & Fagre,

1993). However, the study of the eects of various diets

on modern human isotopic values can provide useful
information on the detailed isotopic eects of particular dietary variations, and can be used to test, refine
and extend palaeodietary theories (Webb, Minson &
Dye, 1980; Nakamura et al., 1982; Minagawa, 1992;
Yoshinaga et al., 1996; OConnell & Hedges, 1999a).
Direct comparison of isotopic data from modern
and archaeological humans is hampered since dierent
tissues are routinely analysed from the two groups:
usually bone from archaeological populations, and
usually hair keratin from modern populations. If
an enduring isotopic relationship may be described
between bone collagen and hair keratin in humans,
then this will facilitate the comparison of archaeological and modern human isotopic values.

1247
03054403/01/111247+09 $35.00/0

 2001 Academic Press

1248 T. C. OConnell et al.

Isotopic values of hair keratin and bone collagen

have been measured in controlled animal feeding
experiments. These studies showed that the two proteins do not have widely diering carbon and nitrogen
isotopic values, and that the isotopic values of both
hair keratin and bone collagen are related to dietary
protein isotopic values. The carbon isotopic values of
bone collagen and hair keratin correlate well with that
of total diet, with the collagen enriched by approximately 5, and the keratin being enriched by +1 to
+3 relative to diet (Ambrose & Norr, 1993; DeNiro
& Epstein, 1978; Jones et al., 1981; Katzenberg &
Krouse, 1989; Minson, Ludlow & Troughton, 1975;
Tieszen, 1983; Tieszen & Fagre, 1993). The nitrogen
isotopic values of most body proteins including bone
collagen, hair keratin and muscle protein also correlate
well with diet, generally all being enriched by +23
relative to dietary protein
15N (Nakagawa et al., 1985;
Sealy et al., 1987; Hare et al., 1991). However, bone
collagen and hair keratin have only been directly
compared in one animal study (three individuals), in
which no significant dierence was found between their

15N values (DeNiro & Epstein, 1981).
Theoretical consideration of these results fits with
the limited experimental data. The isotopic values of
the proteins collagen and keratin are defined by the
isotopic composition of their constituent amino acids.
Large dierences in isotopic values between components amino acids have been observed (Gaebler, Vitti
& Vukmirovich, 1966; Macko et al., 1987; Hare et al.,
1991), reflecting their very dierent metabolic, synthetic and catabolic processes. Collagen and keratin
isotopic values can therefore be expected to dier, since
the two proteins have very dierent amino acid compositions. In particular, one third of the amino acid
residues of collagen are glycine, which is known to be
enriched in carbon isotopic values relative to other
amino acids, which would lead to collagen itself being
more enriched in
13C relative to other tissues: this is
observed in the controlled animal feeding experiments.
In terms of nitrogen isotopic values, the isotopic eect
of deamination and transamination of amino acids
in the body is not fully understood, however, it is
observed that most non-essential amino acids have
similar nitrogen isotopic values, irrespective of the
source tissue (Gaebler, Vitti & Vukmirovich, 1966).
Therefore bone collagen and hair keratin might be
expected to have similar nitrogen isotopic values.
It is important to know whether similar isotopic
patterns are observed in the tissues of free-living
humans as in samples from animals in a laboratorybased study, to ensure that palaeodietary theories
derived from animal experiments can be extrapolated
to humans. But replicating results from animal feeding
experiments in studies on humans is dicult, due to
ethical considerations and restrictions on sample
availability.
An earlier published paper (OConnell & Hedges,
1999b) compared the isotopic values of hair and bone

samples taken from a series of individuals from two

archaeological human populations, to explore whether
similar results to those seen in animal experiments
could be observed in uncontrolled human samples.
The results showed that bone collagen was enriched
relative to hair keratin by between 0 and 1 in
13C,
and by between 0 and 2 in
15N, but the magnitudes
of the dierences were quite variable. These results,
particularly the dierence in nitrogen isotopic values,
do not agree with the animal data, or with the theoretical expectations. However, an archaeological population may not be the best source of material for such
a comparison. For example, lack of information about
individuals diet, state of health, nutrition, and cause of
death, meant that it was not possible to discount those
who died after a long illness, or those who might be
expected to have altered their diet in the last weeks/
months prior to their death. Due to dierences between
protein turnover rates, this type of change might be
recorded in the hair but not in the bone, resulting in a
dierence between the observed isotopic values of these
two tissues. It is also possible that diagenesis of either
the hair or bone could have resulted in alteration of the
isotopic values of either tissue, although this was
not apparent in the usual criteria by which isotopic
analyses of archaeological material are judged. Other
studies comparing isotopic values of dierent archaeological human body tissues also encountered problems,
primarily due to seasonal variation in diet (White,
1993; Aufderheide et al., 1994; White & Schwarcz,
1994), which results in diering isotopic values between
bone collagen and hair keratin or muscle protein, again
due to diering turnover times (Stenhouse & Baxter,
1979; Jones et al., 1981; OConnell & Hedges, 1999a).
Here we attempt to bypass the problems inherent in
studying an archaeological population, by comparing
the isotopic values of bone collagen and hair keratin
from living humans in reasonable health.
We also present an isotopic comparison of hair and
nail keratin from living individuals, to assess whether
both types of keratin have similar isotopic values. The
term keratin refers to a group of about 100 dierent
proteins (Marshal, Orwin & Gillespie, 1991) with some
diversity in their amino acid sequence. The two dierent keratins that comprise hair and nails are both
classified as hard keratins, and have similar amino
acid profiles, varying within narrow limits (Goldsmith,
1991). Proteins with similar amino acid compositions,
such as hair and nail keratin, should be similar in bulk
isotopic composition, unless the same amino acid has a
dierent isotopic value in each protein. This can be
tested by a direct comparison of hair and nail isotopic
values.
The results in this paper aim to provide clear data
on the dierences that might be encountered, at the
level of protein, between the various isotopic measurements made on dierent body parts, and we oer
some explanation for the small dierences in isotope
composition indicated by our results.

Isotopic Comparison of Hair, Nail and Bone 1249

Experimental Analyses
Samples were collected from two groups. Hair and
bone samples were collected from eight individuals
who were undergoing orthopaedic surgery at the
Nueld Orthopaedic Centre, Oxford, U.K. (three men
and five women aged between 58 and 88 years, mean
age 708124 (1 )). Hair and nail samples were
collected from 12 individuals living in Bradford and
Oxford, U.K. (three men and nine women aged
between 21 and 30 years, mean age 23833 (1 )).
All individuals gave informed consent to their participation in the study. The sampling of hair and bone
from individuals undergoing surgery was approved by
the Central Oxford Research Ethics Committee.
At the time of sampling, all subjects gave details of
their diet over the year prior to sampling. None had
significantly changed their diet in the six months
prior to sampling. Diet details taken included listing
the types of protein, carbohydrate, fats and sugars
consumed, as well as the frequency of consumption.
All reagents used in sample preparation were of
analytical grade or above, and all water used was
deionized and distilled.
Bone
Bone samples were taken from individuals undergoing
orthopaedic surgery. Three individuals were having a
total knee replacement, four a total hip replacement,
and one an exploration of the tibia/fibula. Samples of
bone that would otherwise have been discarded were
collected during the operation. Bone samples were
frozen until preparation for analysis.
Collagen was extracted from each bone sample for
isotopic analysis by the standard method of this
laboratory. The samples were rinsed in water, defatted
by soaking in methanol and chloroform (2:1 v/v) for
3 h whilst in an ultrasonic bath, changing the solvent
at least five times, then demineralized in 05 M hydrochloric acid for 23 days, gelatinized in water of pH 3
for 24 h, and the soluble collagen lyophilized. The
dried collagen was then weighed into tin capsules for
continuous flow combustion and isotopic analysis.

methanol and chloroform (2:1 v/v) for about 1 h each,

to remove any lipid or shampoo residue, then samples
were rinsed twice in water for about 20 min each time.
The hairs were then wrapped lengthways in aluminium
foil, and cut into sections of 15 cm length. The
wrapped samples were then dried under vacuum, and
rolled into balls for isotopic analysis.
Nails
Nail samples were collected from each subject by
clipping the free (distal) edge of a fingernail. Samples
were cleaned following a similar protocol to that used
for hair, but with an additional initial physical cleaning. Samples were shot blasted to remove the surface
layer of nail together with any surface contamination.
Samples were then rinsed twice in methanol and
chloroform (2:1 v/v), first for about 1 h, then for
30 min. Samples were then rinsed in water for 20 min,
then rinsed in water a second time in an ultrasonic bath
for 1 h, then dried overnight at 60C. Samples were
then cut into between one and four sections depending
on sample size, and wrapped in small (1 cm) squares of
aluminium foil, ready for continuous flow combustion
and isotopic analysis.
Analyses
Isotopic analyses were performed using an automated
carbon and nitrogen analyser coupled to a continuousflow isotope ratio monitoring mass spectrometer
(Carlo Erba elemental analyser coupled to a PDZ
Europa Geo 20/20 mass spectrometer). Replicate
measurement errors are less than 02 for
13C and

15N. Analyses showed that tin and aluminium foil
blanks contained no carbon or nitrogen. Results are
reported in units of permil () with
13C relative to
VPDB, and with
15N relative to AIR.
Amino acid profiles of protein samples were
measured using an ABI 420A derivatizer/analyser (PE
Biosystems, Warrington, U.K.) following hydrolysis
for 24 h in 57 N hydrochloric acid at 110C
(Heinrikson & Meredith, 1984).

Results
Hair
Hair samples were taken from subjects in both study
groups in the same way. Samples of 1520 hairs were
cut from the crown of the scalp (where possible), the
scalp end of the sample wrapped in micropore tape to
anchor the strands, and the hair stored in a plastic bag
until it was prepared for analysis.
After removal of endogenous lipids, hair can be
considered as eectively pure keratin. Therefore, after
cleaning, minimal sample preparation was required
before isotopic analysis. The preparation procedure
was as detailed in OConnell & Hedges (1999a). Each
sample was cleaned by two successive immersions in

The results are shown in Table 1 and 2. The isotopic

analyses of all collagen and keratin samples were
judged to be valid according to the following criteria.
The measured atomic C/N ratios for bone collagen had
a mean of 330 and standard deviation of 012, within
the accepted range (2936) for uncontaminated
samples (DeNiro, 1985). The atomic C/N ratios of all
hair keratin samples measured in this study had a mean
of 354 and standard deviation of 007, which is well
within the range observed for modern hair samples,
300380 (OConnell & Hedges, 1999a). Nail keratin
sample atomic C/N ratios were also all within the range
of 300380 (mean 349, standard deviation 003).

1250 T. C. OConnell et al.

Table 1. Isotopic analyses of hair and bone samples
Hair
Subject

Bone

C/N
ratio*

13

C
()

15

N
()

357
362
347
355
350
346
353
351

2158

2099

2121

2104

2144

2128

2078

2062

964
988
898
968
949
800
1001
1015

A
B
C
D
E
F
G
H
Overall mean dierence
Overall standard deviation

(bonehair)

C/N
ratio*

13

C
()

15

N
()

13

 C
()

15N
()

323
320
337
330
327
319
325
356

1941

1962

2040

1975

1999

1980

1897

1975

1041
1086
1004
1035
1019
910
1089
1087

217
137
081
129
145
147
180
088
141
045

078
098
105
067
070
110
088
072
086
017

Each value quoted is the mean of three replicate analyses, except for the hair sample from Subject B, which only
provided enough material for two analyses.
*All hair samples had C/N ratios within the range observed in modern hair, 300380, and all bone collagen
samples had C/N ratios within the range 290360 (see discussion in text).
Table 2. Isotopic analyses of hair and nail samples
Hair
Subject
1
2
3
4
5
6
7
8
9
10
11
12
Overall mean dierence
Overal standard deviation

Nail
13

15

(hairnail)
13

15

13

N**

C/N
ratio*

C
()

N
()

N**

C/N
ratio*

C
()

N
()

 C
()

15N
()

3
3
3
3
3
3
3
3
2
3
3
3

362
356
355
354
352
367
339
363
345
348
359
363

2156

2070

2123

2197

2093

2158

2025

2127

2107

2192

2105

2123

850
837
861
849
881
895
842
953
892
640
885
944

4
1
4
1
4
3
4
1
3
4
3
3

347
357
346
349
348
351
350
347
345
350
351
348

2154

2160

2120

2148

2089

2170

2085

2125

2144

2259

2149

2127

938
910
939
928
925
941
892
1019
975
729
940
970

002
090

004

049

004
012
060

002
038
067
044
004
021
039

087

073

078

079

045

046

049

065

083

089

054

026

065
020

*All hair samples had C/N ratios well within the range observed in modern hair, 300380, and all nail samples had
C/N ratios within the range 340367 (see discussion in text).
**Number of sub-samples analysed.

Both hair and nail keratin have a theoretical calculated

atomic C/N ratio of 340. Bone collagen isotopic values
given are the mean of three replicate analyses, with
standard deviations of less than 01 in
13C and less
than 02 in
15N for each sample. Hair keratin
isotopic values given are the mean of analyses of three
contiguous samples (]15 cm each) taken from along
the same section of hair, with standard deviations of
less than 03 in
13C and less than 04 in
15N for
each sample. Nail keratin isotopic values given are the
mean of the analyses of between one and four samples
taken from across the same section of nail, with
standard deviations of less than 02 in
13C and less
than 03 in
15N for each sample. Within samples
from a single individual, there is little isotopic variability; there is slightly more variation within hair and

nail samples than within bone collagen, which is possibly to be expected, since replicate bone collagen
analyses were aliquots taken from a homogenous
sample, whereas nail and hair samples were analyses of
samples taken from adjacent locations on the sample.
Bone and hair
When comparing the isotopic values of bone collagen
and hair keratin from the same individual, there was a
mean enrichment of 141 in bone collagen
13C
values relative to hair keratin
13C values (standard
deviation of 045), and a mean enrichment of 086
in
15N (standard deviation of 017) (Table 1). The
mean dierences in both carbon and nitrogen isotopic
values are significant (Wilcoxon matched pairs signed

Isotopic Comparison of Hair, Nail and Bone 1251

12.0

12.0

B B

H
H
H

9.0

8.0

7.0
22.0

15

10.0

15

N ()

B
HH
H
H

21.0

20.0
13
C ()

19.0

18.0

19.0

20.0

13

Hair keratin C ()

10.0

9.0

8.0

Figure 1. Carbon and nitrogen isotopic values of hair and bone

sample pairings. H=hair keratin; B=bone collagen.

21.0

22.0

23.0
22.0

)
)

11.0

B
Hair keratin N ()

11.0

H = 1.095 B 1.844 (
B = 0.865 H + 2.142 (

21.0
20.0
19.0
13
Bone collagen C ()

18.0

Figure 2. Comparison of carbon isotopic values of hair keratin and

bone collagen.

ranks test: for
13C, Z=
2524, P=0012, N=8; for

15N, Z=
2527, P=0012, N=8). Figure 1 shows a
plot of
13C against
15N for all bone and hair sample
pairings. Figure 2 shows a comparison of
13C for hair
and bone for each sample, and Figure 3 shows a similar
comparison of
15N. The nitrogen isotopic values of
the two proteins seem to be better correlated than the
carbon isotopic values. The regression lines for
15N
(of bone on hair and hair on bone) are plotted in
Figure 3, and the regression coecients and their
standard errors are given in Table 3, and are consistent
with a simple fixed oset between the two proteins

7.0

8.0

9.0
10.0
11.0
15
Bone collagen N ()

12.0

Figure 3. Comparison of nitrogen isotopic values of hair keratin and

bone collagen. Regression lines and equations shown for hair (H) on
bone (B) and vice versa.

isotopic values, since the gradient (m) is close to 1, and

the standard error on the gradient (m) is small.
The possibility that the bone samples taken did not
yield collagen of good quality (that can be compared
with collagen analysed in other studies) was considered, since the bone samples were removed for a
medical reason. However this does not seem to be the
case since amino acid profiles of the bone collagen
samples analysed (Figure 4) are very similar to those
reported in the literature (Eastoe, 1955). The magnitude of the isotopic dierences between hair and bone
is also not related to the age of the subject, or to the
type of bone sample taken.
These findings are in broad agreement with analyses
of hair and bone from archaeological individuals
(N=23), where bone collagen
13C was enriched by
05 in
13C (1 =06) and 10 in
15N
(1 =11) relative to hair keratin (OConnell &
Hedges, 1999b). The variability in the dierences is
substantially less in the modern samples, which might
be expected for reasons outlined earlier: we knew that
all individuals sampled were in relatively good health,
and had not changed their diets in the last weeks or
months, which could lead to a dierent isotopic signal
in the hair and bonethis could not be done for the
archaeological populations.
These results are also broadly similar but not identical to results from animal feeding experiments: the
carbon isotopic results suggest that the dierence
between collagen and keratin in adult humans is less
than that observed in animals (+141 as opposed to
+2 to +4) (DeNiro & Epstein, 1978; Tieszen &
Fagre, 1993); the nitrogen isotopic results indicate a
consistent enrichment of bone relative to hair of
approximately +09, whereas DeNiro & Epstein

1252 T. C. OConnell et al.

Table 3. Regression coecients for
15N of hair and bone, and hair and nail
[y=mx+C]
Bone=x, Hair=y
Hair=x, Bone=y
Nail=x, Hair=y
Hair=x, Nail=y

m

C

R2 value

1095
0862
1091
0862

0106
0083
0087
0069

1844
+2142

1489
+1839

1093
0792
0809
0595

0947
0947
0940
0940

Produced using the statistics package SPSS.

m and C are the standard errors on m and C respectively.

0.35
Bone collagen (literature)
Hair keratin (literature)
A bone
A hair
G bone
G hair

Percentage of amino acid in sample

0.30

0.25

0.20

0.15

0.10

0.05

0.00

Asp

Glu

OHpro

Ser

Gly

His

Arg

Thr

Ala

Pro

Tyr

Val

Met

Cys

Ile

Leu

Phe

Lys

Figure 4. Amino acid profiles of bone collagen and hair keratin. Samples taken from subjects A & G. Literature values taken from Eastoe
(1955) and Goldsmith (1991). Analysis performed as described in the text.

(1981) found that collagen and keratin
15N were very

similar.
Hair and nail
Comparisons of hair keratin and nail keratin from the
same individual show that hair keratin is slightly more
enriched in
13C than nail keratin by 021 (standard
deviation of 039) but that nail keratin is on average
065 more enriched in
15N than hair (standard
deviation of 020) (Table 2). The mean dierence
(hairnail) of 021 in
13C is not significant, but the
mean dierence (hairnail) of
065 in
15N is
significant (Wilcoxon matched pairs signed ranks test:
for
13C, Z=
1721, P=0085, N=12; for
15N,
Z=
3070, P=0002, N=12). Figure 5 gives a plot of

13C against
15N for all hair and nail sample pairings,
and Figures 6 and 7 show a comparison of
13C and

15N respectively for hair and nail for each sample.
Figure 7 also shows the regression lines for
15N (of

nail on hair and hair on nail). The regression coefficients and their standard errors (Table 3) are again
consistent with a simple fixed oset between the two
isotopic values, since the gradient (m) is close to 1, and
the standard error (m) is small.
The possibility of the isotopic inequality being a
result of lack of contemporaneity of growth is possible
but unlikely. Exact time correlation was not possible,
due to variables in the growth rate of nails and hair,
dependent on sex, age, length of nail, which finger,
growth phases of the hairs, season of year etc. However
we consider that the samples taken covered the same
period of growth for both tissues. Nail samples were
taken from the free edge (distal end) of the nail,
meaning that they were produced between four and six
months before sampling, given the typical length of a
finger nail and that 1 cm of nail represents three
months growth (rate of 010012 mm/day, Zaias,
1990). Hair samples were taken next to the scalp, but
were all at least 5 cm long, and 1 cm of hair represents

Isotopic Comparison of Hair, Nail and Bone 1253

11.0

11.0

15

9.0

10.0
Nail keratin N ()

N
N
NH H
N N
NN
N
N
HN
H
HH
H
H H
H

15

N ()

10.0

8.0
N

9.0

8.0

7.0

7.0

H = 1.091 N 1.489 (
N = 0.862 H + 1.839 (

H
6.0
23.0

22.0

21.0
13
C ()

20.0

19.0

Figure 5. Carbon and nitrogen isotopic values of hair and nail

sample pairings. H=hair keratin; N=nail keratin.

20.0

13

Nail keratin C ()

7.0

8.0
9.0
10.0
15
Hair keratin N ()

11.0

Figure 7. Comparison of nitrogen isotopic values of hair and nail

keratin. Regression lines and equations shown for hair (H) on nail
(N) and vice versa.

subject 12 (a dierence of 65 residues versus one of

four residues), and the hairnail
13C and
15N spacing
is correspondingly greater (05 versus 00 for
13C,
and
05 versus
03 for
15N). Of the four
individuals whose hair was measured for amino acid
composition, the variations in relative proportions of
amino acids are especially marked for cysteine (A,
164% of total amino acids; H, 148% of total amino
acids; 11, 181% of total amino acids; 12, 119% of total
amino acids).

19.0

21.0

Conclusions

22.0

23.0

6.0

)
)

22.0
21.0
20.0
13
Hair keratin C ()

19.0

Figure 6. Comparison of carbon isotopic values of hair and nail

keratin.

about 1 months growth (rate of 035 mm/day, Saitoh

et al., 1969).
Variations in amino acid composition might be
responsible for some of the dierence in isotopic
values. Amino acid profiles of the hair and nail samples
from two of the subjects were similar (Figure 8), and
close to that reported in the literature for hair and nail
keratin (Goldsmith, 1991), however there was some
variability between individuals. Considering the isotopic analyses of hair and nail carried out for subjects
11 and 12 in which amino acid compositions were also
measured: the dierence in cysteine content between
subject 11s hair and nail is much greater than for

These results establish the extent to which the major

proteins in bone (collagen) and in hair and in nail
(keratins) dier in isotopic composition in humans.
Bone collagen was significantly enriched relative to
hair keratin from the same individual in both
13C and

15N, by 141 and 086 respectively. Nail keratin
was significantly enriched in
15N relative to hair
keratin from the same individual, by 065, but there
was no significant dierence between hair and nail
keratins in
13C. Such data are needed for investigations, both archaeological and modern, of dietbody
isotopic relationships, using dierent body tissues as
proxy samples.
In addition to quantifying the isotopic dierences
between the tissues, we may also consider the causes of
the small dierences we observe. One obvious factor is
that the tissues may represent dierent periods of
dietary intake. In elderly subjects, the bone collagen
could integrate diet over the last 30 years, while the
hair will respond to the last six months. However, hair
and nail samples are sampling diet over the same
period, and we note that the dierences in isotopic

1254 T. C. OConnell et al.

Percentage of amino acid in sample

0.20
Nail keratin (literature)
Hair keratin (literature)
11 nail
11 hair
12 nail
12 hair

0.15

0.10

0.05

0.00

Asp

Glu

OHpro

Ser

Gly

His

Arg

Thr

Ala

Pro

Tyr

Val

Met

Cys

Ile

Leu

Phe

Lys

Figure 8. Amino acid profiles of hair and nail keratin. Samples taken from subjects 11 and 12. Literature values taken from Goldsmith (1991).
Analysis performed as described in the text.

values for hairbone and nailhair have the same

variance in both
13C and
15N, which seems unlikely
if the hair and nail pairings are representing similar diet
whereas the bonehair pairings are not.
A second cause is the dierent amino acid compositions of collagen and keratin. There are limited data as
to the relative dierences in both
13C and
15N
between dierent amino acids within a single protein.
Neither the precise underlying causes nor the extent to
which the pattern of isotopic dierences varies are
established, and insucient isotopic values for individual amino acids exist to make a confident evaluation of
the magnitude of the isotopic dierence between proteins on the basis of amino acid abundance dierence.
Nevertheless, the unusually high glycine content of
collagen (319% of total amino acids: Eastoe, 1955),
and the observed enrichment of glycine in
13C relative
to other amino acids can be expected to account for an
enrichment of at least 2 in collagen
13C compared
to a protein with an average amino acid composition
(e.g. myosin IIa: Weiss, Schiano & Leinwand, 1999)
made from isotopically similar amino acids (data taken
from Hare et al., 1991). Such an isotopic dierence
between collagen and muscle protein is seen in the
few animal feeding experiments published (DeNiro &
Epstein, 1978; Tieszen & Fagre, 1993). Keratins are
unusually rich in cysteine and also in serine (combined
total of 281% of amino acids for hair and 219% of
amino acids for nail), both of which are metabolically
close to glycine, and have similar enrichment in 13C
relative to other amino acids (Hare et al., 1991).
Therefore we would expect collagen to be only slightly

(12) enriched in
13C with respect to keratin,

which is what is observed in this study.
Amino acid composition dierences between hair
and bone provide a quantitatively adequate explanation for the observed carbon isotopic dierences,
and also explain the observation of very little dierence in carbon isotope values between hair and nail.
However, nitrogen isotope dierences appear to be too
large to explain in this way. Firstly, there is less
variation in general between nitrogen isotope values of
dierent amino acids (Hare et al., 1991). The most
significant source of variation is threonine, which is
usually strongly depleted relative to an average protein
15
N content (Gaebler, Vitti & Vukmirovich, 1966;
Hare et al., 1991), and indeed keratins contain 67%
threonine while collagen contains about 18%. But we
estimate that this eect accounts for up to only 025
dierence between bone and hair or nail, and predicts
equality between hair and nail. The values observed are
+086 (bonehair) and
065 (hairnail), so that
additional explanations are called for. While it may
be dicult to refute an appeal to the amino acid
compositional dierence (or turnover times) between
bone and hair, such an explanation cannot be used
to account for hairnail isotopic dierences that are
almost as large. The eect of metabolism on isotopic
composition is greater for nitrogen isotopes than for
carbon and therefore it may not be unreasonable to
expect local dierences between similar metabolic
pathways in the production of keratins. For example,
the intracellular turnover of amino acids may be different in the hair follicle and the nail matrix, and this

Isotopic Comparison of Hair, Nail and Bone 1255

may be connected with transaminations that may be

conducive to isotopic fractionation. This is of course
speculative, but indicates that constancy of amino acid
composition and therefore isotopic values between
tissues, even for similar proteins, cannot be taken for
granted.

Acknowledgements
The authors would like to thank all volunteers who
provided samples. We are grateful to Anita Reed of the
Nueld Orthopaedic Centre, Oxford, for her assistance in sample collection, to Ken Neal, RLAHA,
University of Oxford, for help with sample preparation, and to Tony Willis, MRC Immunochemistry
Unit, University of Oxford, for the amino acid
analyses. TCOCs fellowship is supported by the
Wellcome Trusts BioArchaeology Fund. This project
was approved by the Central Oxford Research Ethics
Committee, reference number 97.281.

References
Ambrose, S. H. (1993). Isotopic analysis of palaeodiets: methodological and interpretive considerations. In (M. K. Sandford, Ed.)
Investigations of Ancient Human Tissue. Langhorne, PA: Gordon
& Breach Science Publishers, pp. 59130.
Ambrose, S. H. & Norr, L. (1993). Experimental evidence for the
relationship of the carbon isotope ratios of whole diet and dietary
protein to those of bone collagen and carbonate. In (J. B. Lambert
& G. Grupe, Eds) Prehistoric Human BoneArchaeology at the
Molecular Level. Berlin: Springer Verlag, pp. 137.
Aufderheide, A. C., Kelley, M. A., Rivera, M., Gray, L., Tieszen, L.
L., Iversen, E., Krouse, H. R. & Carevic, A. (1994). Contributions
of chemical dietary reconstruction to the assessment of adaptation
by ancient highland immigrants (Alto Ramirez) to coastal conditions at Pisagua, North Chile. Journal of Archaeological Science
21, 515524.
DeNiro, M. J. (1985). Postmortem preservation and alteration of in
vivo bone collagen isotope ratios in relation to palaeodietary
reconstruction. Nature 317, 806809.
DeNiro, M. J. & Epstein, S. (1978). Influence of diet on the
distribution of carbon isotopes in animals. Geochimica et
Cosmochimica Acta 42, 495506.
DeNiro, M. J. & Epstein, S. (1981). Influence of diet on the
distribution of nitrogen isotopes in animals. Geochimica et
Cosmochimica Acta 45, 341351.
Eastoe, J. E. (1955). The amino acid composition of mammalian
collagen and gelatin. Biochemical Journal 61, 589600.
Gaebler, O. H., Vitti, T. G. & Vukmirovich, R. (1966). Isotope
eects in metabolism of 14N and 15N from unlabeled dietary
proteins. Canadian Journal of Biochemistry 44, 12491257.
Goldsmith, L. A. (1991). Physiology, Biochemistry and Molecular
Biology of the Skin. Oxford: Oxford University Press.
Hare, P. E., Fogel, M. L., Staord, T. W. Jr, Mitchell, A. D. &
Hoering, T. C. (1991). The isotopic composition of carbon and
nitrogen in individual amino acids isolated from modern and fossil
proteins. Journal of Archaeological Science 18, 277292.
Heinrikson, R. L. & Meredith, S. C. (1984). Amino-acid analysis
by reverse-phase high-performance liquid-chromatography
Precolumn derivatization with phenylisothiocyanate. Analytical
Biochemistry 136, 6574.
Jones, R. J., Ludlow, M. M., Troughton, J. H. & Blunt, C. G. (1981).
Changes in natural carbon isotope ratio of the hair from steers fed
diets of C4, C3 and C4 species in sequence. Search 12, 8587.

Katzenberg, M. A. & Krouse, H. R. (1989). Application of stable

isotope variation in human tissue to problems in identification.
Canadian Society of Forensic Science Journal 22, 719.
Macko, S. A., Fogel (Estep), M. L., Hare, P. E. & Hoering, T. C.
(1987). Isotopic fractionation of nitrogen and carbon in the
synthesis of amino acids by microorganisms. Chemical Geology
(Isotope Geoscience Section) 65, 7992.
Marshall, R. C., Orwin, D. F. G. & Gillespie, J. M. (1991). Structure
and Biochemistry of Mammalian Hard Keratin. Electron
Microscope Review 4, 4783.
Minagawa, M. (1992). Reconstruction of human diet from
13C and

15N in contemporary Japanese hair: a stochastic method for
estimating multi-source contribution by double isotopic tracers.
Applied Geochemistry 7, 145158.
Minson, D. J., Ludlow, M. M. & Troughton, J. H. (1975).
Dierences in natural carbon isotope ratios of milk and hair from
cattle grazing tropical and temperate pastures. Nature 256, 602.
Nakagawa, A., Kitagawa, A., Asami, M., Nakamura, K., Schoeller,
D. A., Slater, R., Minagawa, M. & Kaplan, I. R. (1985).
Evaluation of isotope ratio (IR) mass spectrometry for the study
of drug metabolism. Biomedical Mass Spectrometry 12, 502506.
Nakamura, K., Schoeller, D. A., Winkler, F. J. & Schmidt, H.-L.
(1982). Geographical variations in the carbon isotope content of
the diet and hair of contemporary man. Biomedical Mass
Spectrometry 9, 390394.
OConnell, T. C. & Hedges, R. E. M. (1999a). Investigations into the
ect of diet on modern human hair isotopic values. American
Journal of Physical Anthropology 108, 409425.
OConnell, T. C. & Hedges, R. E. M. (1999b). Isotopic comparison of
hair and bone: archaeological analyses. Journal of Archaeological
Science 26, 661665.
Saitoh, M., Uzuka, M., Sakamoto, M. & Kobori, T. (1969). Rate of
hair growth. In (W. Montagna & R. L. Dobson, Eds) Hair
Growth. Oxford: Pergamon Press, pp. 183201.
Schwarcz, H. P. & Schoeninger, M. J. (1991). Stable isotopic
analyses in human nutritional ecology. Yearbook of Physical
Anthropology 34, 283321.
Sealy, J. C., van der Merwe, N. J., Lee-Thorp, J. A. & Lanham,
J. L. (1987). Nitrogen isotopic ecology in Southern Africa: implications for environmental and dietary tracing. Geochimica et
Cosmochimica Acta 51, 27072717.
Stenhouse, M. J. & Baxter, M. S. (1979). The uptake of bomb 14C in
humans. In (R. Berger & H. E. Suess, Eds) Radiocarbon Dating.
California: University of California Press, pp. 324341.
Tieszen, L. L., Boutton, T. W., Tesdahl, K. G. & Slade, N. A. (1983).
Fractionation and turnover of stable isotopes in animal tissues:
implications for
13C analysis of diet. Oecologia 57, 3237.
Tieszen, L. L. & Fagre, T. (1993). Eect of diet quality and
composition on the isotopic composition of respiratory CO2, bone
collagen, bioapatite, and soft tissues. In (J. B. Lambert & G.
Grupe, Eds) Prehistoric Human BoneArchaeology at the
Molecular Level. Berlin: Springer-Verlag, pp. 121155.
Webb, Y., Minson, D. J. & Dye, E. A. (1980). A dietary factor
influencing 13C content of human hair. Search 11, 200201.
Weiss, A., Schiano, S. & Leinwand, L. A. (1999). Comparative
sequence analysis of the complete human sarcomeric myosin heavy
chain family: implications for functional diversity. Journal of
Molecular Biology 290, 6175.
White, C. D. (1993). Isotopic determination of seasonality in diet
and death from Nubian mummy hair. Journal of Archaeological
Science 20, 657666.
White, C. D. & Schwarcz, H. P. (1994). Temporal trends in stable
isotopes for Nubian mummy tissues. American Journal of Physical
Anthropology 93, 165187.
Yoshinaga, J., Minagawa, M., Suzuki, T., Ohtsuka, R., Kawabe, T.,
Inaoka, T. & Akimichi, T. (1996). Stable carbon and nitrogen
isotopic composition of diet and hair of Gidran-speaking Papuans.
American Journal of Physical Anthropology 100, 2334.
Zaias, N. (1990). The Nail in Health and Disease. Norwalk,
Connecticut: Appleton & Lange.