Forensic Science International 221 (2012) 106–112

Contents lists available at SciVerse ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locate/forsciint

Age estimation charts for a modern Australian population

Matthew Blenkin *, Jane Taylor
University of Newcastle, School of Health Sciences, Australia

A R T I C L E I N F O A B S T R A C T

Article history: Calculation of the biological age of an individual has application in many fields of dentistry. It can be used
Received 21 July 2011 to determine the appropriate timing of interventionist treatment for example in orthodontics; to analyse
Received in revised form 20 March 2012 the developmental stage of an individual relative to the general population in the management of genetic
Accepted 14 April 2012
or congenital conditions which disturb growth; and to estimate the age of a living or deceased person for
Available online 6 May 2012
forensic purposes.
Many of the techniques used to estimate age can be quite time consuming to complete. This time
Keywords:
component is a major disadvantage in a forensic context when age estimations in mass disasters are
Age estimation
required as part of the post-mortem examination process. Consequently, forensic practitioners have
Dental development
Disaster victim identification tended to use the simpler but less reliable atlas style techniques of Schour and Massler and Ubelaker in
Forensic odontology these situations.
Forensic science For mass disaster situations, such as the recent Victorian Bushfires, it would be advantageous to have
access to Australian specific data in the convenient Schour and Massler format.
This project reinterpreted the Australian data previously collected by Blenkin and other relevant
studies and applied it to a schematic similar to that of Ubelaker to develop a reliable, convenient and
contemporary reference for use in age estimation.
ß 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction institutionalised children with no differentiation between male

Calculation of the biological age of an individual has application
in many fields of dentistry. It can be used to determine the
appropriate timing of interventionist treatment for example in
orthodontics; to analyse the developmental stage of an individual
relative to the general population in the management of genetic or
congenital conditions which disturb growth; and to estimate the
age of a living or deceased person for forensic purposes.
While accurate, many of the techniques based on the analysis of
dental development do require training and experience to ensure
precision and the process can be quite time consuming to
complete. This time component is a limitation in a forensic
context when age estimations in mass disasters are required as
part of the post-mortem examination process. Consequently,
forensic practitioners have tended to use the simpler but less
reliable atlas style techniques of Schour and Massler [1] and
Ubelaker [2] in these situations.
Valid criticisms of the Schour and Massler charts include the
fact that, although augmented by additional data, they were based
on the work of Logan and Kronfeld [3] and as such suffer from
similar limitations of a small sample size of chronically ill,
and female. In assigning a different diagram to each year of growth,
the range about a mean age can only be 6 months, a range too
small to be credible. Being an ‘atlas type’ system any comparison is
subject to a higher degree of inter-observer disagreement when
compared to systems that rely on objective physical measurements.
However, the wide applicability, despite the sample on which it is
based, may be testament to the relative stability of the pattern and
rate of dental development regardless of the impact of environmental
factors or disease on growth. Finally, the fact that the published
images are of solid teeth, yet are most often compared to radiographs
of the case at hand for assessment can prove problematic.
More recently, Ubelaker [2] modified and improved upon the
chart produced by Schour and Massler. Using data from numerous
population studies he modified the stages of tooth development
and position within the existing diagrams and included new
additional diagrams. The population basis for this work was
American Indian and other ‘nonwhite’ populations. Owing to
issues with identification of sex of the sub-adult samples, no
differentiation in the data was made between males and females.
The current study is aimed at addressing many of these
shortfalls.

2. Materials and methods

* Corresponding author at: Dental Department, HMAS Stirling, PO Box 2188,
Rockingham DC, WA 6958, Australia. Tel.: +61 419 246 124. The charts developed by Ubelaker were examined in detail. Each diagram within
E-mail address: matthew.blenkin@defence.gov.au (M. Blenkin). these charts was analysed using the modified Demirjian method as described by

0379-0738/$ – see front matter ß 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.forsciint.2012.04.013
 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112 107

Blenkin [4]. In essence, each diagram was treated as a ‘case’ that needed an age 3. Results
estimate. Each tooth of the right mandibular quadrant, excepting the third molar
(41–47) was assessed and rated according to the Demirjian system. These ratings
were converted to numerical scores and estimates of age, including range at the 95% The completed charts are depicted in Figs. 1–4.
confidence interval, were made using Blenkin’s standards. For those diagrams By way of example, a male child with erupted and occluded first
representing ages below the utility of the Blenkin method, data from Fanning were molars and maxillary central incisors, lateral incisors erupted but
used to estimate the age of the cases depicted [5]. For those diagrams representing not yet in occlusion, and deciduous molars and canines still present
ages older than the utility of the Blenkin method, data collected on the development
of mandibular third molars by the authors and Dr Bassed et al. were used [6].
would be assessed as being 8.8  1.8 years, or of the age range of 7–
The confidence intervals of each of the depicted stages of development varies 10.6 years. A further example: 69% of females aged 22 years will have
according to the source of the data from which the age estimates were derived. The reached the stage of apex closure of the permanent third molars.
first four diagrams up to the age of 1.4 years were assessed using the work of The validation study resulted in the correct estimation of age
Fanning [5] and as such represent the stage of development attained between the
range of the case at hand in 84% of the female cases and in 79% of
10th and 90th percentiles. The next six diagrams up to the age of 6.3 years (6.7 years
for females) were assessed using Blenkin’s data, the 95% confidence intervals of the male cases. While 13–14% of cases had their the age ranges
which had been calculated using the standard error of estimate for each age group underestimated, overestimates of age range occurred in only 7% of
[7]. The next five diagrams up to the age of 13.8 years (14 years for males) were cases (3% in females). Importantly, none of the incorrect age range
assessed using Blenkin and Evans’ data, the 95% confidence intervals of which had estimates were greater than one diagram removed form the correct
been calculated using the residual error from the ANOVA output in order to provide
a simplified and uniform confidence interval from across the sample age groups [4].
one. Table 2 summarises these results.
The detail regarding the distribution of the sample by age can be found in Blenkin’s The analysis of inter-observer error resulted in a kappa statistic
paper, with the overall n = 3261 (m = 1637, f = 1624). The diagram depicted as 15 of 0.74 for female cases and 0.70 for males. The analysis of intra-
years (14.5 yrs for males) was assessed using the same data collected by Blenkin observer error resulted in a kappa statistic of 0.8 for both sexes.
above, but where the least squares method was used to regress only the
These statistics represent levels of ‘substantial reliability’ accord-
development data on third molars for both sexes individually (n = 1526 with
f = 816, m = 710), the 95% confidence interval have been calculated using the ing to Landis and Koch [8].
residual error from the ANOVA output. The last two diagrams were assessed using
data collected by Bassed et al. [6], the percentiles represent the percentage of 4. Discussion
individuals of the given ages who had attained that specific stage of development.
Again, the age distribution of the sample population can be found in Bassad’s paper
Up until the landmark study by Logan and Kronfeld in 1933 it
with the overall n = 973.
Each pictogram was then assigned an ‘Australian’ age, and diagrams which was generally held that the calcification of all the permanent teeth
represented stages with inadequate differentiation were removed. began at the same time [3]. This assumption was based upon tables
A validation study was undertaken using a sample of 204 OPGs (m = 100, f = 104). produced 40 years earlier by Legros and Magitot [9].
The distribution of the sample by chronological age is given in Table 1.
Today we know that the development of the dentition
The cases were randomly selected from cases at the Victorian Institute of
Forensic Medicine, the Oral Health Centre of Western Australia and Private commences at differing stages in the growth of the child
Orthodontic practice in Australia. All cases had been previously de-identified prior dependent on the teeth in question. It is recognised that there is
to use in this study. Each case was blindly assessed and given an age estimate by a degree of variation in the actual timing due to the inherent
two assessors who discussed and agreed upon the estimate based on the charts biological variability of human beings, hence the use of ranges
provided here. Each estimate was then compared to the chronological age for each
rather than discrete timings. It is also widely acknowledged that
respective case in order to establish the validity of the method.
A study of inter observer error was undertaken. One assessor was the author, the the teeth develop through a succession of stages that are common
other, a general practitioner dentist with some forensic experience. The results of to all humans and that all teeth, both deciduous and permanent,
their individual age assessment for each of the 204 cases used in the validation normally pass through all these stages of development. These
study, were compared and analysed using the kappa statistic. Intra observer error
stages are predictable and observable and it is the observation of
was assessed by way of the author reassessing 50% of the sample (n = 102) after an
intervening period of 8 weeks.
these stages that many methods of age estimation are based upon.
The study by Logan and Kronfeld in 1933 was a cross sectional
study using histological sectioning and radiographic examination of
0–15 year olds. The study was based on the examination of 25 fresh
Table 1 post-mortem specimens, however, 19 of these were under 2 years-
Age distribution of the validation sample.
of-age. From the data the authors published a detailed account of the
Age Male Female Total positioning and relative development of the human dentition.
0* 19 18 37 In 1941, Schour and Massler published an important study that
1 6 4 10 summarised the development of the human dentition in an ‘atlas
2 4 3 7 style’ chart [1]. It was based on the data collected by Logan and
3 2 1 3
Kronfeld and other histological studies of the developing dentition
4 1 1 2
5 2 1 3
undertaken by Schour around the same time. The summary of this
6 3 1 4 information into a pictorial or ‘atlas type’ chart provided the
7 – 3 3 profession with a tool that would not only be used in every day
8 3 4 7 practice, but would also prove useful in estimating the age of an
9 5 2 7
individual in a forensic context. The method involves comparison of
10 6 6 12
11 5 6 11 a radiograph of, preferably, the entire maxillary and mandibular
12 12 12 24 dentitions against diagrams depicting the stage of the development
13 12 19 31 of the dentition representative of each year of age. By matching the
14 6 11 17
radiographic image to a specific diagram the estimated age of the
15 5 5 10
16 7 5 12
child is that listed with the associated diagram.
17 1 – 1 The advantages of using this system to estimate the age of an
18 – 1 1 individual are that it is non-destructive, being able to conduct the
19 1 – 1 assessment on radiographs of the jaws; it is simple and does not
22 – 1 1
require any specialised training to recognise specific stages of
Total 100 104 204 development, as required in many of the more recently developed
*
Each age group is represented by the midpoint of the range. For example age = 1 systems; and it does not require the use of specialised equipment
represents the range 0.51–1.5 years of age. beyond an X-ray apparatus.
108 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112

Fig. 1. Age estimation guide for a modern Australian population for females from the prenatal period to the sixth year. The primary teeth are the darker ones shaded in the
illustration.
Modified from Ref. [2].

The limitations of using an atlas-based system are that they
are based on the premise that ‘one size fits all’. The clinician is
required to make a qualitative assessment of the case at hand
and match it to the nearest ‘best fit’ diagram. In practice rarely
will the case at hand match one diagram exactly and will quite
often have components of at least two of the depicted stages.
Thus a high risk of error exists. It is for this reason that these
methods should only be used as a screening tool after which time
a more detailed quantitative analysis and age estimate be
undertaken.
4.1. Variation in dental development
It has been estimated that the contribution of genetic control
over the rate of dental development is as much as 90% [10–12]. The
timing of the emergence of deciduous incisors has also been shown
 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112 109

Fig. 2. Age estimation guide for a modern Australian population for females from the seventh year to adulthood. The primary teeth are the darker ones shaded in the illustration.
Modified from Ref. [2].

to be under strong genetic control [13]. Even within a population of
similar genetic heritage there will be a range of rates of
development due to natural biological variation, which can be
considered to be variation due to chance, hence the reason for
giving age estimations as a range and not a single point value.
Studies of the pattern of dental mineralisation show that the
early stages of tooth development are almost the same for both
males and females and the sexual dimorphism in developmental
rates occurs at around the crown completion stage and continues
to increase during the root development stage, when girls reach
the majority of developmental stages ahead of boys [14–17],
highlighting the need for separate charts for each sex.
Regional variation in dental development is well documented.
Numerous studies have shown dental maturation rates that are
consistent for a given racial group, but vary significantly between
groups [18–28]. Some researchers have posited that these
differences may reflect differences in methodology or be due to
secular trends and therefore geographic-specific standards may not
be necessary [29–32]. Recent studies by Thevissen et al. focussed on
the development of third molars, have shown that while regional
differences do exist and are statistically significant, the actual
differences are small and not consistent over the age ranges
considered [33,34]. In these studies it was suggested that, as a result
of their findings, it may be appropriate to use population specific
datasets for age estimation of cases from other populations.
Increasing global mobility has led to the intermixing of gene
pools resulting in less distinct racial traits. Most studies involving
methods that focus on dental development have aimed to select
racially homogeneous samples in the hope of limiting the
variability associated with rate of dental maturation. This
approach may not be reasonable, however, given that one such
study revealed that significant variation can even exist within
110 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112

Fig. 3. Age estimation guide for a modern Australian population for males from the prenatal period to the sixth year. The primary teeth are the darker ones shaded in the
illustration.
Modified from Ref. [2].

what was thought to be a ‘fairly homogeneous’ population [23].
Therefore, if significant variations in the rate of dental develop-
ment occur between (and within) different populations, and since
a significant proportion of the Australian population is comprised
of subgroups born in many other countries, the variation found
within an Australian sample may well be much larger than the
variation reported in countries with less genetically mixed
populations. Moreover, the greater the variation, the wider the
confidence intervals associated with predictive models.
This multicultural nature of Australian society means that
unless the genetic or racial origin of the victim is known or
suspected, and that appropriate conversion tables for that specific
group can be used, reliance will have to be placed on standards
that, as discussed above, are probably associated with broader
error margins around any given age estimate. As such, while the
age ranges for the depicted diagrams are greater than those for the
original charts on which they are based, they are now far more
applicable to the contemporary Australian population.
 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112 111

Fig. 4. Age estimation guide for a modern Australian population for males from the seventh year to adulthood. The primary teeth are the darker ones shaded in the illustration.
Modified from Ref. [2].

Consistent with other body systems, variation in dental
development increases with increasing age. The early develop-
mental stages of early forming teeth are notably less variable than
the later developmental stages of the later forming teeth [10,16],
thus with increasing age, the range of variation for age estimation
must also rise.
Wide variation also exists in the timing of tooth emergence or
eruption. Emergence times are affected by a diverse range of
factors including infection, pathology, trauma, crowding, extrac-
tion, agenesis and the presence of supernumerary teeth
[5,15,35,36]. Further, a number of published studies do not
address sexual dimorphism in emergence times and the data may

Table 2
Distribution of cases by attribution to diagram correctly estimating chronological age.

Age range diagram 1 earlier than correct Correct Age range diagram 1 later than correct Total

F 14 (13%) 87 (84%) 3 (3%) 104

M 14 (14%) 79 (79%) 7 (7%) 100

Total 28 (13.7%) 166 (81.4%) 10 (4.9%) 204

112 M. Blenkin, J. Taylor / Forensic Science International 221 (2012) 106–112
be sourced from small samples sizes for which the reliability and
accuracy of information is unknown. The appropriate sample size
is very much dependant on the research goal, however a review of
the literature would indicate that a minimum of 30 cases of each
sex for each age group would generally be considered adequate.
The charts developed here, and representing radiographic view of
development, depict specific stages of emergence through alveolar
bone for specific age ranges. It is reasonable to assume that
variation from that depicted is possible, being one of the significant
limitations of this type of technique.
While the data with which these charts have been reanalysed
have been individually validated in previous studies, it is
recommended that these charts as presented here be the subject
of further, larger validation studies as a single consolidated method
of age estimation.
The charts presented here have not been modified from the
originals, only analysed using relevant datasets. As such, the age
ranges represented by each diagram are not neatly aligned to
whole year groups. A further project is currently underway
reinterpreting the data from the original samples in order to
redraw new diagrams that depict consistently spaced whole year
age groups. A similar work was recently published by AlQahtani
et al. [37]. Their work was based on cases taken from a range of
sources and assessed for stages of development using the system
devised by Morreess, Fanning and Hunt [15]. The main advantages
of this work over the previous work of Shour and Massler and
Ubelakker is the availability of diagrams for each year of
development. As with any categorical system of assessment, the
larger number of diagrams (within reason) results in more accurate
estimates of age. Such an undertaking requires considerable time
and resources. In this current paper the authors wished to
demonstrate a simple method by which existing charts may be
modified to reflect age estimates more suited to a specific
population than that on which the charts were originally based.
5. Conclusion
The age estimation charts developed here provide a useful tool to
estimate age at time of death for cases occurring within the
Australian population. The datasets on which the charts a based are
relevant, contemporaneous, based upon significant sample sizes and
sex specific. Notwithstanding the inherent limitations of the atlas
style, these charts are recommended for use as an initial age
assessment, or ‘screening’ tool especially in the mortuary in a mass
disaster situation. A more precise and detailed age estimation
analysis using a technique such as that based on the work of
Demirjian, but using population specific datasets, should be
undertaken upon completion of the post-mortem, when the time
and other pressures of working in the mortuary have been removed.
Acknowledgements
The collaboration of Dr R. Bassed and Dr J. Graham of the
Victorian Institute of Forensic Medicine was much appreciated. As
were the efforts of Dr S. Chiam of the ACT Department of Health.
The authors also wish to acknowledge Dr D. Ubelaker of the
Smithsonian Institution for his assistance in giving permission to
use his diagrams as a basis for this work. This research was, in part,
made possible by the award of an Australian Dental Research
Foundation grant.
References
[1] I. Schour, M. Massler, Studies in tooth development: the growth pattern of human
teeth, Part II, J. Am. Dent. Assoc. 27 (1940) 1918–1931.
[2] D. Ubelaker, Human Skeletal Remains, 2nd ed., Taraxacum, Washington, DC,
1989 , p. 63.
[3] W.H.G. Logan, R. Kronfeld, Development of the human jaws and surrounding
structures from birth to the age of fifteen years, J. Am. Dent. Assoc. 20 (1933)
379–427.
[4] M. Blenkin, W. Evans, Age estimation from the teeth using a modified Demirjian
system, J. Forensic Sci. 55 (2010) 1504–1508.
[5] E. Fanning, A longitudinal study of tooth formation and root resorption, N. Z. Dent.
J. 57 (1961) 202–217.
[6] R.B. Bassed, C. Briggs, O. Drummer, Age estimation using CT imaging of the third
molar tooth, the medial clavicular epiphysis, and the spheno-occipital synchon-
drosis: a multifactorial approach, Forensic Sci. Int. 212 (2011) 273e1–273e5.
[7] M. Blenkin, Forensic Odontology and Age estimation: An Introduction to Concepts
and Methods, VDM Verlag, Saarbrucken, 2009, p. 171.
[8] J.R. Landis, G.G. Koch, The measurement of observer agreement for categorical
data, Biometrics 33 (1977) 159–174.
[9] C.H. Legros, E. Magitot, Chronologie des follicles dentaires chez l’homme, in:
Congres de Lyon, 1893.
[10] S.M. Garn, A.B. Lewis, D.L. Polacheck, Sibling similarities in dental development, J.
Dent. Res. 39 (1960) 170–175.
[11] B. Pelsmaekers, R. Loos, C. Carels, C. Derom, R. Vlietinck, The genetic contribution
to dental maturation, J. Dent. Res. 76 (1997) 1337–1340.
[12] D.R. Merwin, E.F. Harris, Sibling similarities in the tempo of human tooth
mineralization, Arch. Oral Biol. 43 (1998) 205–210.
[13] T.E. Hughes, M.R. Bockmann, K. Seow, et al., Strong genetic control of emergence
of human primary incisors, J. Dent. Res. 86 (2007) 1160–1165.
[14] I. Gleiser, E. Hunt, The permanent first molar: its calcification, eruption and decay,
Am. J. Phys. Anthropol. 13 (1955) 253–281.
[15] C.F.A. Moorrees, E.A. Fanning, E.E. Hunt, Age variation of formation stages for ten
permanent teeth, J. Dent. Res. 42 (1963) 1490–1502.
[16] A. Demirjian, G.-Y. Levesque, Sexual differences in dental development and
prediction of emergence, J. Dent. Res. 59 (1980) 1110–1122.
[17] N. Chaillet, M. Nyström, M. Kataja, A. Demirjian, Dental maturity curves in finnish
children. Demirjians method revisited and polynomial functions for age estima-
tion, J. Forensic Sci. 49 (2004) 1324–1331.
[18] E. Fanning, C. Moorees, A comparison of mandibular molar formation in Austra-
lian Aborigines and Caucasoids, Arch. Oral Biol. 14 (1969) 999–1006.
[19] H.T. Loevy, Maturation of permanent teeth in black and Latino children, Acta
Odontol. Pediatr. 4 (1983) 59–62.
[20] D. Owsley, R. Jantz, Formation of the permanent dentition in the Arikira Indians:
timing differences that affect dental age assessments, Am. J. Phys. Anthropol. 61
(1983) 467–471.
[21] U. Hägg, L. Matsson, Dental maturity as an indicator of chronological age: the
accuracy and precision of three methods, Eur. J. Orthod. 7 (1985) 25–34.
[22] M. Nyström, Dental maturity in Finnish children, estimated from the develop-
ment of seven permanent mandibular teeth, Acta Odontol. Scand. 44 (1986)
193–198.
[23] M. Nyström, Comparisons of dental maturity between the rural community of
Kuhmo in northeastern Finland and the city of Helsinki, Community Dent. Oral
Epidemiol. 16 (1988) 215–217.
[24] E. Harris, J. McKee, Tooth mineralization standards for blacks and whites from the
middle southern United States, J. Forensic Sci. 35 (1990) 859–872.
[25] V. Staaf, H. Mörnstad, U. Welander, Age estimation based on tooth development: a
test of reliability and validity, Scand. J. Dent. Res. 99 (1991) 281–286.
[26] P.J. Davis, U. Hägg, The accuracy and precision of the Demirjian system when used
for age determination in Chinese children, Swed. Dent. J. 18 (1994) 113–116.
[27] H. Mörnstad, The validity of four methods for age determination by teeth in
Swedish children: a multicentre study, Swed. Dent. J. 19 (1995) 121–130.
[28] R.L. Tompkins, Human population variability in relative dental development, Am.
J. Phys. Anthropol. 99 (1996) 79–102.
[29] H.M. Liversidge, T. Speechly, Growth of permanent mandibular teeth of British
children aged 4 to 9 years, Ann. Hum. Biol. 28 (2001) 256–262.
[30] J. Braga, Y. Heuze, O. Chabedel, et al., Non-adult dental age assessment: corre-
spondence analysis and linear regression versus Bayesian predictions, Int. J. Legal
Med. 119 (2005) 260–274.
[31] H.M. Liversidge, N. Chaillet, H. Mörnstad, et al., Timing of Demirjian’s tooth
formation stages, Ann. Hum. Biol. 33 (2006) 454–470.
[32] M. Maber, H.M. Liversidge, M.P. Hector, Accuracy of age estimation of radio-
graphic methods using developing teeth, Forensic Sci. Int. 159S (2006)
S68–S73.
[33] P.W. Thevissen, S. Fieuws, G. Willems, Human third molar development:
comparison of 9 country specific populations, Forensic Sci. Int. 201 (2010)
102–105.
[34] P.W. Thevissen, A. Alqerban, J. Asaumi, F. Kahveci, J. Kaur, Y.K. Kim, P. Pittayapat,
M. Van Vlierberghe, Y. Zhang, S. Fieuws, G. Willems, Human dental age estimation
using third molar development stages: accuracy of age predictions not using
country specific information, Forensic Sci. Int. 201 (2010) 106–111.
[35] U. Hägg, J. Taranger, Dental development, dental age and tooth counts, Angle
Orthod. 55 (1985) 93–107.
[36] L. Suri, E. Gagari, H. Vastardis, Delayed tooth eruption: pathogenesis, diagnosis
and treatment. A literature review, Am. J. Orthod. Dentofacial Orthop. 126 (2004)
432–445.
[37] S.J. AlQahtani, H. Liversidge, M.P. Hector, Atlas of Tooth Development and
Eruption, University of London, 2009.