Brook and Scheers 2006.1


33

*Correspondence to:  A. H. Brook, School of Dental 
Studies, Edwards Building, University of Liverpool, 
Pembroke Place, Liverpool L69 3GN, UK
Email:  a.h.brook@liverpool.ac.uk

Tooth root morphology can provide valuable 
additional evidence to crown morphology in studies 
of prehistoric, historic and modern populations.  
The determination of root morphology may be 
multifactorial, as is crown morphology, with both 
genetic and environmental factors involved (Winter 
and Brook, 1989).  Variations of root morphology 
include the number of roots, as with accessory roots 
or fused roots, their shape, as in taurodontism, or 
their size.  Ethnic differences in root morphology 
have been recognized (Dixon and Stewart, 1976).

The present study aimed to derive data for 
root anomalies in a homogeneous Romano-British 
population and to investigate associations with other 
dental anomalies in this group.  Also the study aimed 
to develop further the methodology of measurement 
and the reproducibility of diagnosis of root anomalies 
in archeological material, enhancing comparisons 
with other ancient and modern populations.

MATERIALS AND METHODS

The skulls investigated were from a cemetery of 
the Roman town of Durnovaria, close to the site of the 
modern Poundbury, Dorset, UK.  The cemetery dates 
from the 3rd to 5th century AD and is of a Christian 
character.  The population was of native British origin 
throughout this period (Farwell and Molleson, 1993). 

Variations of Tooth Root Morphology in a Romano-
British Population

Alan H. Brook* and Marlene Scheers

School of Dental Studies, University of Liverpool, United Kingdom

ABSTRACT:    Tooth morphology can provide valuable 
evidence in studies of prehistoric, historic and modern 
populations.  The aims of this study were to derive data 
for root anomalies in a Romano-British population, to 
investigate associations between anomalies, and to 
compare findings with other populations to provide 
evidence concerning etiology.  An additional aim was 
to develop further the methodology and reproducibility 
in such studies.  From the Christian cemetery of 3rd-5th 
century AD in Poundbury, UK, 385 skulls were suitable 
for examination.  Radiographic technique was stan-
dardized with custom-made skull supports and criteria 
established for each anomaly.  There was a high level 
of reproducibility for the diagnosis of each anomaly.  

The prevalence of the anomalies in individuals was:  
three-rooted mandibular first molars 1.8%, fused roots 
14.0%, cuneiform roots 16.9%, taurodontism 26.9%, and 
invaginated teeth 1.1%.  There were highly significant 
(P < 0.001) associations between fused and cuneiform 
roots, and both were significantly associated with third 
molar hypodontia (P < 0.002; P < 0.05).  These reduc-
tions in root morphology were commonly bilateral and 
more frequent in females, as is hypodontia.  The find-
ings of this study are compatible with a multifactorial 
etiology of these anomalies, showing continuous varia-
tion in root morphology.  The gradients of anomalies 
observed are also compatible with the concept of mor-
phogenetic fields.  Dental Anthropology 2006;19(2):33-38.

The excavated skulls are housed at the British Museum 
(Natural History), London, UK. The total collection 
from this burial site consists of 1,100 crania, but a 
large proportion of these are very fragmented and 
unsuitable for this study.  The criterion for inclusion 
in the present investigation was a jaw that had at 
least one permanent molar and one permanent incisor 
present.  Juvenile skulls with a dental age of less than 
9 years were excluded. The resultant sample was 385 
skulls suitable for examination of root morphology.

Age and sex determinations were made by the 
staff of the British Museum based on the long bones, 
pelvic girdles and skulls. Of the sample, 40.0% 
(154) were estimated male, 38.7% (149) female, and 
for 21.3% (82) no determination could be made.

Radiographs were taken of all teeth using an 
industrial apparatus and Kodak ultraspeed dental 
occlusal films. A pilot study established the optimum 
voltage, current and exposure time as well as the 
standardized positioning of the x-ray tube, skull and 
films.  Customized wooden blocks were developed 
for positioning the skulls.  A total of 6 films per skull 



34 A.H. BROOK AND M. SCHEERS

provided full coverage of the teeth (Fig. 1).  To calibrate 
measurements, all films were taken including a 20 mm 
length of orthodontic wire.  The films were developed 
and viewed under standardized conditions.

The anomalies diagnosed were three-rooted 
mandibular first permanent molars, fused (pyramidal) 
and cuneiform roots, taurodontism and crown and root 
invaginations.

The radiographic criteria for the five anomalies 
diagnosed in the study were:  (1) three-rooted 
mandibular first molar:  evidence of a third root; 
(2) fused molar root: a pyramidal root form with 
no evidence of an interradicular bony septum or 
periodontal ligament but with separation of root canals; 
(3) cuneiform molar root:  a root form with a central 
root canal whose shape followed the root outline; (4) 
taurodontism:  criteria of Holt and Brook (1979; Fig. 2); 
and (5) crown invaginations:  criteria of Hallett (1953); 
types 2, 3 and 4 were scored following Grahnen et al. 
(1959) and Brook (1974).

For each anomaly, prevalence for skulls, prevalence 
for teeth, sex distribution, and symmetry were 
investigated.

To test the reproducibility of the diagnosis, the 
radiographs of 20% of the sample were read on a 
second, separate occasion.  The reproducibility findings 
are in Table 1.

RESULTS

The findings for the prevalence for skulls and sex 
distribution of the root anomalies studied are shown in 
Table 2, which also indicates the number of individuals 
suitable for scoring each anomaly.

Three-rooted mandibular first molars—those with 
an accessory root—had a prevalence of 1.8% of skulls 
and a tooth prevalence of 1.5%.  In half of the individuals 
the anomaly was bilateral and equal numbers of males 

and females were affected.
For reduction in root number the prevalence for 

fused roots was 14.0% of skulls and for cuneiform 
roots was 16.9% of skulls.  The tooth prevalence for 
reduced root number was 2.7% for fused roots and 
3.4% for cuneiform roots, some individuals possessing 
both anomalies.  Maxillary molars were affected more 
frequently than mandibular molars.  Third molars 
were more often affected than second molars, with no 
example being found in first permanent molars.  The 
male to female ratios of 1:2.2 for fused roots and 1:2 for 
cuneiform roots were statistically significant (P < 0.02 
and P < 0.01, respectively).  In 30% of affected skulls 
these anomalies were bilateral, and occasionally a 
fused root was seen on one side of the dental arch with 
a cuneiform root on the contralateral tooth.

Taurodontism was found in 26.9% of skulls, with 
a tooth prevalence in lower molars of 11.7%.  Third 
molars were the teeth most often affected and first 
molars the least.  Taurodontism was bilateral in 47% 
of affected skulls.  The male to female ratio was 1:0.67 
and statistically significant (P < 0.05).  The differences 
between mean values of a and of a:b ratios (Fig. 2) 
in those teeth showing taurodontism compared to 

Fig. 1. Full coverage of the dentition using six 
radiographs.

Baseline = axis between mesial and distal points of 
amelo-cemental junction.

 a = distance from baseline to highest point on 
pulp chamber floor.

 b = distance from baseline to apex of distal 
root.

Fig. 2. Measurement used for taurodontism (after 
Holt and Brook, 1979).



35TOOTH ROOT MORPHOLOGY IN ROMANO-BRITONS

those without were highly significant (P < 0.001).  
For measurement b, the differences were statistically 
significant for second molars (P < 0.01) and third 
molars (P < 0.001).

The prevalence of invaginated teeth was 1.1% 
of skulls.  All invaginations in this sample occurred 
in maxillary lateral incisors.  There was no evidence 
of periapical bone loss in relation to the crown 
invaginations that were all of the mild Hallett (1953) 
Class 2 category.  No example of root invaginations 
was seen in this study.

The statistically significant associations between 
anomalies in this study were between fused and 
cuneiform molar roots (P < 0.001) and between 
taurodontism and fused molar roots (P < 0.05).  Using 
results from a study of anomalies of tooth number 
and size in this population (Brook and John, 1995) 
statistically significant associations were found 
between congenital absence of third molars and fused 
molar roots (P < 0.02), cuneiform molar roots (P < 0.05), 
and taurodont lower molars (P < 0.001).

DISCUSSION

The high level of reproducibility of measurements 
and diagnosis suggests that the present data are 
reliable within the constraints of the sample (Table 1).  
Comparisons with other historic and modern sample 
findings are therefore worthwhile.

In other studies of three-rooted mandibular first 
molars either radiographs (Souza-Freitus et al., 1971) 
or extracted teeth (Curzon, 1973) have been used.  The 
radiographs in the present study were of good quality 
(Fig. 1) having been carefully standardized in the pilot 

study.  For Caucasians the findings for three-rooted 
mandibular first molars have usually been of the order 
of 1% of individuals affected, while the frequency 
in Mongoloid peoples is much higher (Scott and 
Alexandersen, 1992).  The Romano-British figure of 
1.8% conforms to modern Caucasian results reviewed 
by Alexandersen (1963).

The criterion used for the diagnosis of cuneiform 
roots is shown by Holt (1976) to be highly reproducible, 
especially in lower molars.  In this study, the one 
reversal of diagnosis (Table 1) was a maxillary molar 
that on the second occasion was diagnosed as having 
a fused root.  The criterion for fused molar roots was 
adapted from Brabant and Kovacs (1961) and provided 
an acceptable degree of consistency (Table 1).

Similar to Brabant and Kovacs (1961), the highest 
frequency of cuneiform roots was found in maxillary 
third molars.  For mandibular second molars the 
Romano-British prevalence was 3.2% of skulls, 
comparable to the findings of Pedersen (1949) in East-
Greenland skull material and of Holt (1976) based on 
Brook’s (1974) large population sample of modern 
British Caucasian schoolchildren.  Reduction in root 
number affecting first permanent molars would seem 
to be rare as no example was found in these Romano-
British skulls or by Pedersen (1949), Holt (1976) or 
Molnar and Horvath (1995).

The tendency for a bilateral occurrence of anomalies 
of reduced root number is found in other studies also, 
with the same trend for fused and cuneiform roots to 
occur in antimeric teeth where each anomaly was not 
bilaterally symmetrical (Holt, 1976; Ross and Evanchik, 
1982; Tamse and Kaffe, 1981; Molnar and Horvath, 

  No. diagnosed on  No. diagnosed on  No. diagnosed on
 Feature first occasion second occasion on both occasions

 Three-rooted
 mandibular first molars 1 1 1

 Fused molar roots 11 13 11

 Cuneiform molar roots 16 15 15

 Invaginations  2 1 | 1 2 1 1 1

B. Using ratios

  No. of teeth No. within 0.5 of
 Feature measured ratio on first reading

 Taurodont mandibular molars 272 212

TABLE 1. Reproducibility of measurements and diagnosis

A. Clinical diagnosis



36 A.H. BROOK AND M. SCHEERS

 
 

N
o.

 o
f 

 
N

o.
 o

f 
 

N
o.

 o
f 

 
 

 
af

fe
ct

ed
 

 
af

fe
ct

ed
 

 
u

n
kn

ow
n

 
 

Se
x 

ra
ti

o 
To

ta
l

 
A

n
om

al
y 

m
al

es
 

P
re

v
al

en
ce

 
fe

m
al

es
 

P
re

v
al

en
ce

 
se

x 
P

re
v

al
en

ce
 

M
al

e:
Fe

m
al

e 
ca

se
s 

P
re

v
al

en
ce

T
h

re
e-

ro
ot

ed
m

an
d

ib
u

la
r 

fi
rs

t 
m

ol
ar

 
3 

2.
3%

 
3 

2.
4%

 
0 

0%
 

1:
1 

6 
1.

8%
 

 
(n

=
13

2)
 

 
(n

=
12

7)
 

 
(n

=
71

) 
 

 
(n

=
33

0)

Fu
se

d
 r

oo
t 

(u
p

p
er

an
d

 lo
w

er
 m

ol
ar

 t
ee

th
) 

13
 

8.
5%

 
29

 
19

.5
%

 
12

 
14

.6
%

 
1:

2.
2 

54
 

14
.0

%
 

 
(n

=
15

3)
 

 
(n

=
14

9)
 

 
(n

=
82

) 
 

 
(n

=
38

4)

C
u

n
ei

fo
rm

 r
oo

t 
(u

p
p

er
an

d
 lo

w
er

 m
ol

ar
 t

ee
th

) 
19

 
12

.4
%

 
38

 
25

.5
%

 
10

 
12

.2
%

 
1:

2 
65

 
16

.9
%

 
 

(n
=

15
3)

 
 

(n
=

14
9)

 
 

(n
=

82
) 

 
 

(n
=

38
4)

Ta
u

ro
d

on
ti

sm
(l

ow
er

 m
ol

ar
s)

 
48

 
34

.8
%

 
31

 
22

.3
%

 
15

 
20

.5
%

 
1:

0.
67

 
94

 
26

.9
%

 
 

(n
=

13
8)

 
 

(n
=

13
9)

 
 

(n
=

73
) 

 
 

(n
=

35
0)

R
oo

t 
in

v
ag

in
at

io
n

 2
|

2 
0 

0.
0%

 
0 

0.
0%

 
0 

0.
0%

 
0 

0 
0.

0%
 

 
(n

=
11

1)
 

 
(n

=
10

3)
 

 
(n

=
65

) 
 

 
(n

=
27

9)

C
ro

w
n

 in
v

ag
in

at
io

n
 2

|
2 

2 
1.

8%
 

1 
1.

0%
 

0 
0%

 
1:

0.
5 

3 
1.

1%
 

 
(n

=
11

1)
 

 
(n

=
10

3)
 

 
(n

=
65

) 
 

 
(n

=
27

9)

T
A

B
L

E
  2

. P
re

va
le

n
ce

s 
in

 s
ku

ll
s 

an
d 

se
x 

di
st

ri
bu

ti
on

s 
of

 r
oo

t 
an

om
al

ie
s



37

1995).  There is also agreement that these anomalies 
are more common in females than males (Brabant and 
Kovacs, 1961; Holt, 1976; Ross and Evanchik, 1981; 
Table 1).

For taurodontism, comparisons are limited by the 
use of different criteria in different studies.  Using 
the same methodology as the present study, Holt and 
Brook (1979) found that 6.3% of 1,115 modern British 
Caucasian schoolchildren had a taurodont mandibular 
first permanent molar compared to 1.8% of the 
Romano-British skulls.  The prevalence and degree 
of taurodontism is often greater in second and third 
molars (Molnar and Horvath, 1995); in the present 
Romano-British sample the prevalence was 26.9% of 
skulls. Shifmann and Chanannel (1978) report that 
taurodontism occurred bilaterally in most cases while in 
this study and in Holt and Brook (1979) approximately 
equal numbers of tooth pairs were affected bilaterally 
and unilaterally.  The sex ratio in this study of male 1:
female 0.67 is similar to that of Holt and Brook (1979).

The prevalence of invaginated teeth in these 
Romano-Britons at 1.6% is lower than that for modern 
British, 4.1% (Brook, 1974) and modern Swedish 3.0% 
(Grahnen et al., 1959) samples.

Building on the comments of previous authors who 
remarked on a tendency for fused and cuneiform molar 
roots to be found together, this study provides evidence 
of the highly statistically significant association between 
these two anomalies.  Similarly, the association of root 
number reduction of molars with congenital absence 
of third molars noted by Keene (1966) was found to be 
statistically significant in these Romano-Britons.

Stoy (1960), Stenvik et al. (1972) and Holt and Brook 
(1979) describe the association of taurodontism and 
hypodontia.  In this study the association between 
taurodontism and hypodontia of third molars was 
shown to be highly significant statistically.  This finding 
is compatible with a morphogenetic field effect with a 
varying influence anteroposteriorly.

The strong association between fused and cuneiform 
molar roots could indicate that the cuneiform root 
is one extreme of a continuous variation showing 
different degrees of confluence of roots and their 
canals.  Complete root separation would represent the 
opposite extreme.

In conclusion, for root anomalies in this Romano-
British population, the prevalence for skulls and for 
teeth, the sex distribution and bilateral symmetry 
has been established. The radiographic technique 
developed and the criteria used have high degrees of 
reproducibility.  The statistically significant associations 
demonstrated in the Romano-Britons showed the 
relationship between fused roots and cuneiform roots 
as reductions in root number and shape and also their 
relationship with congenital absence of teeth.  The 
gradients of anomalies observed were compatible with 
the concept of morphogenetic fields.  The findings were 

also compatible with multifactorial etiology, showing 
continuous variation in root size and shape.

LITERATURE CITED

Alexandersen V. 1963.Double-rooted human lower 
canine teeth. In: Brothwell DR, editor. Dental 
anthropology. New York: Pergamon Press, p 235-
244.

Brabant H, Kovacs I. 1961. A contribution to the study 
of taurodontism in modern races and of its possible 
relation to the pyramidal roots of molars. Bull 
Group Rech Scient Stomot 4:232-286.

Brook AH. 1974. Dental anomalies of number, form 
and size: their prevalence in British schoolchildren. 
J Int Assoc Dent Child 5:37-53.

Brook AH, John CC. 1995. Dental anomalies of number 
and size in a Romano-British population. In: 
Radlanski RJ, Renz H, editors. Proceedings of 10th 
International Symposium on Dental Morphology. 
Berlin: C and M Brunne GbR, p 177-180.

Curzon MEJ. 1973. Three-rooted mandibular permanent 
molars in English Caucasians. J Dent Res 52:181.

Dixon GH, Stewart RE. 1976. General aspects of 
anomalous tooth development. In: Stewart R E and 
Prescott G H, editors. Oral Facial Genetics. St Louis: 
C V Mosby, p 133-135.

Grahnen H, Lindahl B, Omnell KA. 1959. Dens 
invaginatus: I. a clinical roentgenological and 
genetical study of permanent upper lateral incisors. 
Odont Revy 10:115-137.

Hallett GEM. 1953. The incidence, nature and clinical 
significance of palatal invaginations in the maxillary 
incisor teeth. Proc R Soc Med 46:491-499.

Holt RD. 1976. The prevalence of root anomalies in 
children. MSc Thesis, University of London.

Holt RD, Brook AH. 1979. Taurodontism: a criterion 
for diagnosis and its prevalence in mandibular 
first permanent molars in a sample of 1,115 British 
schoolchildren. J Int Assoc Dent Child 10:41-47.

Keene HJ. 1966. A morphological and biometric study 
of taurodontism in a contemporary population. Am 
J Phys Anthropol 25:208-209.

Molnar E, Horvath G. 1995. Developmental anomalies 
of the teeth in historic skeletal samples. In: Radlanski 
RJ, Renz H, editors. Proceedings of the 10th 
International Symposium on Dental Morphology. 
Berlin: C and M Brunne GbR, p 377-385.

Oehlers FA. 1958. The radicular type of dens invaginatus. 
Oral Surg Oral Med Oral Pathol 11:251-260.

Pedersen PO. 1949. The East Greenland Eskimo 
dentition: numerical variations and anatomy. 
Copenhagen: C A Retzels Forlag.

Ross IF, Evanchik PA. 1981. Root fusion in molars: 
incidence and sex linkage. J Periodontol 52:663-
667.

Scott GR, Alexandersen V. 1992. Dental morphological 
variation among Medieval Greenlanders, 

TOOTH ROOT MORPHOLOGY IN ROMANO-BRITONS



38

Icelanders and Norwegians. In: Smith P, Tchernov 
E, editors. Structure, function and evolution of 
teeth. London: Freund Publishing House, p 467-490.

Shifmann A, Chanannel I. 1978. Prevalence of 
taurodontism found in a radiographic dental 
examination of 1,200 young adult Israeli patients. 
Community Dent Oral Epidemiol 6:200-203.

Souza-Freitus J A, Lopez E S, Casati-Alvares 
L. 1971. Anatomic variations of lower first 
permanent molar roots in two ethnic groups. 
Oral Surg Oral Med Oral Pathol 31:274-278.

Stenvik A, Zachrisson BU, Svatun B. 1972. 
Taurodontism and concomitant hypodontia in 
siblings. Oral Surg Oral Med Oral Pathol 33:841-845.

Stoy PJ. 1960. Taurodontism associated with other 
dental anomalies. Dent Pract Dent Rec 10:202-205.

Tamse A, Kaffe I. 1981. Radiographic survey 
of the prevalence of conical lower 
second molar. Int Endod J 14:188-190.

Winter GB, Brook AH. 1989. Tooth abnormalities. 
In: Rowe HR, Alexander AG, Johns RB, 
editors. A comprehensive guide to clinical 
dentistry. London: Class Publishing, p 55-103.

A.H. BROOK AND M. SCHEERS