You are on page 1of 3


9 brief comms MH NEW 9/9/05 10:31 AM Page 333

Vol 437|15 September 2005

Age written in teeth by nuclear tests
A legacy from above-ground testing provides a precise indicator of the year in which a person was born.

Establishing the age at death of individuals is an a 900 b 1995

important step in their identification and can
be done with high precision up to adolescence 800

Estimated date of birth

by analysis of dentition, but it is more difficult 700
in adults. Here we show that the amount of 600 1975
radiocarbon present in tooth enamel as a result ∆14C (‰) 500
of nuclear bomb testing during 1955–63 is a 400
remarkably accurate indicator of when a per- 1955
son was born. Age is determined to within 1.6
years, whereas the commonly used morphol- 200 Pre-
ogical evaluation of skeletal remains and tooth 100
wear is sensitive to within 5–10 years in adults. 0
1950 1960 1970 1980 1990 2000 Pre- 1955 1965 1975 1985 1995
The amount of carbon-14 isotope (14C) in –100 bomb
Year Actual date of birth
the atmosphere remained relatively stable
until 1955, when above-ground nuclear bomb Figure 1 | Date of birth determined from 14C in teeth. a, Nuclear bomb tests during 1955–63 produced
tests caused it to rise dramatically1,2. Although large amounts of 14C, which have since declined exponentially (blue line). The 14C:C ratio has ranged
the bombs were detonated at only a few loca- from 1.151012 to 2.201012 since 1950. 14C represents the 14C value corrected for radioactive
tions, the additional 14C in the atmosphere decay and 13C fractionation (see supplementary information). To estimate an individual’s date of birth,
rapidly equalized around the globe. Since the the 14C concentration measured in their tooth enamel is plotted on to the curve of atmospheric 14C
against time (blue) to find the year of enamel synthesis (right-pointing arrows), and the known age at
Test Ban Treaty in 1963, atmospheric 14C has enamel formation for individual teeth was then subtracted from the year obtained to give the date of
been dropping exponentially (Fig. 1a). This is birth (left-pointing arrows; dashed vertical lines). Two representative cases are shown (red and green);
not primarily because of radioactive decay (the two teeth were analysed for the case depicted in green. Solid vertical lines, actual dates of birth.
half-life of 14C is 5,730 years) — it is also due to b, Relation between estimated and actual dates of birth. Each point corresponds to one individual,
diffusion from the atmosphere3. Atmospheric except for the ‘pre-bomb’ point, which represents four individuals; coloured points are cases shown in a.
C reacts with oxygen to form carbon dioxide,
which is incorporated into plants by photo- wisdom teeth at 12 years of age. For individu- verified on a larger number, and perhaps on a
synthesis; by eating plants, and animals that als born before 1943 (12 years before the onset wider geographical range, of cases before it can
feed on plants, the 14C concentration in the of nuclear bomb testing), we can therefore be applied to forensic work.
human body closely parallels that in the conclude by this method only that birth Although the nuclear bomb tests were con-
atmosphere at any given time4–6. occurred before that year, albeit with a high ducted several decades ago and the resulting
The enamel of individual teeth is formed at degree of certainty (100% correct in our analy- change in atmospheric 14C is now decreasing
distinct, well characterized times during child- sis (Fig. 1b); n4). In any case of ambiguity as only slowly (Fig. 1a), the method described
hood7,8 and it contains 0.4% carbon. There is to whether birth occurred before or after the here should allow precise age determination
no turnover of enamel after it has been laid peak of nuclear bomb testing, it is necessary to for a long time to come because techniques for
down, so the 14C concentration reflects that in analyse two teeth that were formed at different 14
C measurement are becoming increasingly
the atmosphere at the time of enamel forma- ages: this distinguishes whether the 14C mea- sensitive. In addition, accelerator mass spec-
tion. We measured the 14C content of tooth surements relate to the rising or falling part of trometry for 14C analysis has become more
enamel (for methods, see supplementary the 14C curve (Fig. 1a). accessible and inexpensive, making the pot-
information) and related it to the known con- The sensitivity of our method is mainly ential application of our dating method no
centrations in the atmosphere in different determined by variation between individuals more difficult than other methods now used
years to establish the year of tooth formation. in their age at tooth formation, and the preci- in routine forensic examinations.
This date was then related to the known age sion of the 14C measurement. The degree of Kirsty L. Spalding*, Bruce A. Buchholz‡,
for enamel deposition of individual teeth7 to inter-individual variation is different for dif- Lars-Eric Bergman†, Henrik Druid†,
establish the person’s year of birth (Fig. 1a). ferent teeth, so selection for 14C measurement Jonas Frisén*
We found that this method gave a remark- of teeth with the least variation8 and of several Departments of *Cell and Molecular Biology,
ably precise estimate of age for 22 individuals teeth from the same individual should give a Medical Nobel Institute, and †Forensic Medicine,
(R20.99 from regression shown in Fig. 1b; more accurate date of birth. With regard to Karolinska Institute, 17177 Stockholm, Sweden
for details, see supplementary information). measurement precision, we cannot exclude the e-mail:
The average systematic deviation from the possibility that differences in diet or in local ‡ Center for Accelerator Mass Spectrometry,
correct value was 0.2 years, and the average conditions might contribute some variability Lawrence Livermore National Laboratory,
absolute error for individual measurements in the amount of 14C incorporated into tooth Livermore, California 94551, USA
was 1.61.3 years (s.d.). This indicates that enamel. Although such an effect is not sup-
the precision is substantially higher than that ported by results from comparative analyses of 1. De Vries, H. Science 128, 250–251 (1958).
2. Nydal, R. & Lovseth, K. Nature 206, 1029–1031 (1965).
obtained by other available methods9. different foodstuffs produced in rural and 3. Levin, I. & Kromer, B. Radiocarbon 46, 1261–1272 (2004).
The final formation of enamel is for the industrial areas10, the method will need to be 4. Libby, W. F., Berger, R., Mead, J. F., Alexander, G. V. &

©2005 Nature Publishing Group

Ross, J. F. Science 146, 1170–1172 (1964). 10. Otlet, R. L., Walker, A. J., Fulker, M. J. & Collins, C. HO RO 0.6
5. Harkness, D. D. Nature 240, 302–303 (1972). J. Envir. Radioact. 34, 91–101 (1997). 50 HO N

6. Spalding, K. L., Bhardwaj, R. D., Buchholz, B. A., Druid, H. & OH OH

Fluorescence intensity
Frisén, J. Cell 122, 133–143 (2005). Supplementary information accompanies this 40

7. Nolla, C. M. J. Dent. Child. 27, 254–266 (1960). communication on Nature’s website.
O 0.4
8. Bolanos, M. V., Manrique, M. C., Bolanos, M. J. & Competing financial interests: declared (see online version 30 Dopaxanthin Betanin
Briones, M. T. Foren. Sci. Int. 110, 97–106 (2000). of the communication). (R=glucose)
9. Ritz-Timme, S. et al. Int. J. Legal Med. 113, 129–136 (2000). doi:10.1038/437333a 20 0.2

0 0.0
BOTANY 300 400 500 600 700
Wavelength (nm)

Floral fluorescence effect Figure 1 | Spectra of dopaxanthin and betanin.

Dopaxanthin is used as a model betaxanthin
because of its structural (insets) and biochemical
The way flowers appear to insects is crucial for spectrum of dopaxanthin and the absorbance similarity to betacyanins. When excited by blue
light, betaxanthins emit green fluorescence4.
pollination1–3. Here we describe an internal spectrum of betanin indicates that the light
Fluorescence spectra (blue line, excitation
light-filtering effect in the flowers of Mirabilis emitted by the fluorophore is strongly re- spectrum; green line, emission spectrum) for
jalapa, in which the visible fluorescence emit- absorbed (Fig. 1). Addition of increasing con- natural dopaxanthin (6.0 M, in water) are shown;
ted by one pigment, a yellow betaxanthin, is centrations of betanin to the dopaxanthin violet line, absorbance spectrum of pure betanin
absorbed by another, a violet betacyanin, to solution reduced the intensity of its fluores- (8.4 M, in water). Note the overlap of the
create a contrasting fluorescent pattern on the cence, until only 30% of the initial fluorescence emission and absorbance spectra of the pigments.
flower’s petals. This finding opens up new pos- was detectable at a ratio of 8.5:1. (For details
sibilities for pollinator perception as fluores- and methods, see supplementary information.) The fluorescence micrograph shows that fluo-
cence has not previously been considered as a This internal light-filtering effect between rescence is inhibited in areas where betaxan-
potential signal in flowers. the two types of betalain plant pigment causes thins coexist with betanin (Fig. 2d) — the dark
We investigated the spectra and distribution a fading of visible fluorescence on parts of area corresponds to the orange area in Fig. 2c.
of the pigments in the multicoloured, strik- the flower where both types are present; areas Fluorescence can be an important signal in
ingly patterned flowers of M. jalapa (Nyctagi- containing only betaxanthins appear yellow mate choice for budgerigars6 and possibly in
naceae), which open only in the late afternoon. under white light because of a combination of mantis shrimp7, and it may be that in flowers it
This and related plants, such as Bougainvillea, fluorescence and reflectance of non-absorbed attracts pollinators. The patterns arising from
Celosia, Gomphrena and Portulaca, contain radiation (Fig. 2a). The effect can be demon- the internal light-filtering effect between beta-
pigments known as betalains. These comprise strated in a system designed to visualize green lain pigments described here could encourage
the yellow, fluorescent betaxanthins4 and vio- fluorescence, which filters the incident light to bees1 and bats8, which have visual receptors
let betacyanins, of which betanin (betanidin- blue and causes betaxanthins in the flower to that are sensitive to green light and can detect
O--glucoside) is the most common. fluoresce by emitting green light (Fig. 2b). bright targets better than dim ones9. Variation
We extracted and purified the pigments of Detailed images of different zones of petal in light emission by flowers at visible wave-
M. jalapa flowers and analysed them by high- coloration were obtained by using light and flu- lengths also modifies their colour, which
performance liquid chromatography, as previ- orescence microscopy. A brightfield image would enhance their visibility to pollinators10.
ously described5. The analysis confirmed that under white light shows some cells containing Fernando Gandía-Herrero,
the pigmentation pattern on the flowers was only betaxanthins (Fig. 2c, yellow), others with Francisco García-Carmona, Josefa Escribano
due to a mixture of betaxanthins and betanins. betacyanins (Fig. 2c, deep-red spots), and some Departamento de Bioquímica y Biología
Measurement of the fluorescence-emission with both pigments together (Fig. 2c, orange). Molecular A, Unidad Docente de Biología,
Facultad de Veterinaria, Universidad de Murcia,
Figure 2 | Visible 30100 Espinardo, Murcia, Spain
a c fluorescence in e-mail:
Mirabilis jalapa petals.
a, b, Flower with areas 1. Gumbert, A. Behav. Ecol. Sociobiol. 48, 36–43 (2000).
of red or yellow 2. Giurfa, M., Eichmann, B. & Menzel, R. Nature 382,
coloration under white 458–461 (1996).
3. Heiling, A. M., Herberstein, M. E. & Chittka, L. Nature 421,
light (a); only the 334 (2003).
yellow areas emit green 4. Gandía-Herrero, F., García-Carmona, F. & Escribano, J.
fluorescence when J. Chromatogr. A 1078, 83–89 (2005).
excited by blue light (b) 5. Gandía-Herrero, F., Escribano, J. & García-Carmona, F.
(scale bar, 1.5 cm). Plant Physiol. 138, 421–432 (2005).
c, d, Light micrographs 6. Arnold, K. E., Owens, I. P. F. & Marshall, N. J. Science 295,
92 (2002).
of a section of a single 7. Mazel, C. H., Cronin, T. W., Caldwell, R. L. & Marshall, N. J.
red-and-yellow petal, Science 303, 51 (2004).
b d showing brightfield (c) 8. Winter, Y., Lopez, J. & von Helversen, O. Nature 425,
and fluorescent (d; 612–614 (2003).
excitation wavelength, 9. De Ibarra, N. H., Vorobyev, M., Brandt, R. & Giurfa, M.
450–490 nm) images J. Exp. Biol. 203, 3289–3298 (2000).
10. Vorobyev, M., Marshall, J., Osorio, D., De Ibarra, N. H. &
(scale bar, 500 m). Menzel, R. Color Res. Appl. 26 (suppl.), 214–217 (2001).
Green fluorescence is
due to betaxanthins; Supplementary information accompanies this
dark areas correspond communication on Nature’s website.
to orange areas in c, Competing financial interests: declared none.
where light emitted doi:10.1038/437334a
from the fluorescent
pigment is absorbed BRIEF COMMUNICATIONS ARISING online
by betanin. ➧ see Nature contents.

©2005 Nature Publishing Group