CONTINUING EDUCATION ARTICLE

Interarch tooth size relationships of 3 populations:

“Does Bolton’s analysis apply?”

Stephanie S. Smith, DDS, MS,a Peter H. Buschang, PhD,b and Etsuko Watanabe, DDSc
Dallas, Tex

This study evaluates whether Bolton’s interarch ratios extend across populations and genders. The data
were derived from systematically collected preorthodontic casts of 180 patients, including 30 males and 30
females from each of 3 populations (black, Hispanic, and white). Forty-eight mesiodistal contact points were
digitized on each model, and the lengths of the anterior, posterior, and overall arch segments were
calculated. The results showed significant (P < .05) ethnic group differences in all 6 arch segment lengths
and in all 3 interarch ratios. Whites displayed the lowest overall ratio (92.3%), followed by Hispanics
(93.1%), and blacks (93.4%). The group differences were due primarily to the relationships between the
posterior segments. The arch segments of males were significantly larger than females; the overall and
posterior ratios were also significantly larger in males than in females. Multiple regression analyses showed
that individual differences in the overall ratio were most closely associated with the size of the lower second
premolar, followed by the upper lateral incisors, upper second premolars, and the lower central incisors. In
combination, these 4 teeth explained approximately 50% of the variation in the overall ratio between
subjects. We conclude that interarch tooth size relationships are population and gender specific. Bolton
ratios apply to white females only; the ratios should not be indiscriminately applied to white males, blacks,
or Hispanics. (Am J Orthod Dentofacial Orthop 2000;117:169-74)

and widely used method for detecting interarch tooth

Specific dimensional relationships must
exist between the maxillary and mandibular teeth to
ensure proper interdigitation, overbite, and overjet.
Because patients with interarch tooth size discrepan-
cies require either removal (eg, interdental stripping)
or addition (eg, composite build-ups/porcelain
veneers) of tooth structure to open or close spaces
in the opposite arch, it is important to determine
the amount and location of a tooth size discrepancy
before starting treatment.
Orthodontists have used several methods to detect
interarch tooth size discrepancies in patients presenting
for orthodontic treatment. Most methods, including
Kesling’s diagnostic setup,1 Howes’ ratio of canine
fossa width to total maxillary tooth width,2 and Neff’s
anterior coefficient3,4 are not commonly used. The
Bolton analysis,5-7 based on the ratios between the
mesiodistal tooth diameter sums of the mandibular and
the maxillary dentitions, remains the most recognized
From the Department of Orthodontics, Baylor College of Dentistry, Dallas, Tex.
aGraduate Resident.
bAssociate Professor, Director of Orthodontic Research.
cPostdoctoral Fellow, Department of Orthodontics, Nagasaki University School
of Dentistry.
Reprint requests to: Peter H. Buschange, PhD, Department of Orthodontics,
Baylor College of Dentistry, 3302 Gaston Avenue, Dallas, Texas 75243
Copyright © 2000 by the American Association of Orthodontists.
0889-5406/2000/$12.00 + 0 8/1/98115
size discrepancies.8-10
Bolton developed his overall and anterior ratios
based on 55 patients with excellent Class I occlusions.
Although Bolton’s analysis has proven extremely use-
ful in the clinical setting to guide the orthodontist in
cases with extreme tooth size discrepancies, it is not
without limitations. First, Bolton’s estimates of varia-
tion were underestimated because his sample was
derived from perfect Class I occlusions.9-11 Second,
and perhaps more important, the population and gender
composition of Bolton’s sample were not specified,
which implies potential selection bias.
There is good evidence that populations differ with
respect to interarch tooth size relationships because dif-
ferences in tooth sizes are not systematic.12-14 For
example, blacks have larger maxillary canines, premo-
lars, and first molars than whites even though there are
no differences for the maxillary central or lateral
incisors.15 Moreover, tooth size differences between
males and females are not systematic across all teeth.12-
22 Because the population and gender differences in
maxillary tooth size are not the same as the differences
in mandibular tooth size, different interarch relation-
ships might be expected.
Few studies have directly examined ethnic or gen-
der differences in interarch tooth size relationships.
Richardson and Malhotra19 found significant differ-
169
170 Smith, Buschang, and Watanabe
Fig 1. Maxillary and mandibular contact points.
ences between blacks and whites using an interarch
ratio that included the second molars. Their overall
ratio for blacks was 94%, which was very different
from Bolton’s value of 91%. In contrast, their anteri-
or ratio was similar to Bolton’s value, suggesting that
blacks differ markedly in the posterior arch segment
relationship. Lavelle20 concluded that blacks have
larger overall and anterior ratios than whites and
Asians, although the actual differences were not test-
ed and the arch segments responsible for the differ-
ences were not evaluated. He also showed that the
overall and anterior ratios were consistently larger in
males than in females, regardless of race. Based on
the substantial indirect evidence and the limited
direct evidence available, studies are needed to assess
the applicability of Bolton’s ratios across populations
and genders.
The purpose of this study was to extend Bolton’s
work by accomplishing the following:
1. Comparing the maxillary and mandibular tooth size ratios
of 3 untreated populations: blacks, Hispanics, and whites
2. Comparing the tooth size ratios of males and females
3. Identifying the individual teeth that most affect the inter-
arch relationship and that may be the best indicators of a
tooth size discrepancy.
American Journal of Orthodontics and Dentofacial Orthopedics
February 2000
MATERIAL AND METHODS
Preorthodontic records were systematically collect-
ed from consecutive cases treated between 1988 and
1997. The final sample included 180 persons, with 30
males and 30 females from each of 3 populations
(black, Hispanic, and white). Subjects ranged between
12 and 38 years of age. To minimize error variance, the
following selection criteria were used:
1. A fully erupted permanent dentition from first molar to
first molar
2. Ethnicity verified by photographs and patient histories
3. Good quality study models
Rejection criteria included:
1. Gross restorations, buildups, crowns, onlays, Class II
amalgams or composite restorations that affect the tooth’s
mesiodistal diameter.
2. Congenital defects or deformed teeth.
3. Obvious interproximal or occlusal wear of teeth.
Model Analysis
A total of 48 mesiodistal contact points (2 points
per tooth for each arch) were marked on the maxillary
and mandibular models (Fig 1). Contact points were
defined at the greatest distance between approximate
surfaces, as observed or estimated when the teeth were
rotated or poorly aligned.13 The 3-dimensional coordi-
nates of each contact point were digitized and the
mesiodistal tooth diameters were calculated using with
following formula,
Mesiodistal tooth size = √ (Mtx - Dtx)2 + (Mty – Dty)2 + (Mtz – Dtz)2
where M and D refer to the mesial and distal contacts
of tooth t and x,y,z refer to the transverse, vertical, and
anterioposterior coordinates of each contact. Replicate
analyses of 16 randomly selected patients showed that
method errors24 of the mesiodistal diameters for the
individual teeth ranged between 0.17 mm and 0.48
mm. The molars showed the largest errors; the premo-
lars and canines showed the smallest errors. None of
the measures displayed significant systematic errors.
Method Error = √ (Σ2 /2n)
The individual tooth diameters were added to derive the
anterior (canine to the canine), posterior (first molar to
first premolar), and overall (first molar to first molar)
arch segments. The segments were used to define the
following ratios:
1. Overall ratio: overall mandibular arch segment divided by
the overall maxillary arch segment.
2. Anterior ratio: anterior mandibular arch segment divided
by the anterior maxillary arch segment.
American Journal of Orthodontics and Dentofacial Orthopedics Smith, Buschang, and Watanabe 171
Volume 117, Number 2

3. Posterior ratio: posterior mandibular arch segment divided Table I. Population differences (mm) in overall, anteri-
by the posterior maxillary arch segment. or, and posterior segment lengths
Populations
Statistical Analysis
Segment F-ratio Significance White Black Hispanic
The skewness and kurtosis statistics showed that the Maxilla
variables were normally distributed. Two-way analyses Overall 13.51 <0.001 94.4 98.4 98.0
of variance (ANOVA) were used to evaluate population Anterior 7.14 0.001 46.7 48.4 48.3
and gender differences. There were no significant (P > Posterior 16.63 <0.001 47.7 50.0 49.6
.05) interaction effects. Tukey’s honestly significant Mandible
Overall 22.67 <0.001 87.2 92.0 91.1
difference test was used post hoc to evaluate population Anterior 10.65 <0.001 37.2 38.4 38.9
differences. A regression through the origin (no-inter- Posterior 31.82 <0.001 50.0 53.6 52.2
cept model) was used to evaluate the relationship of the
overall maxillary segment (independent variable) and
the overall mandibular segment (dependent variable). Table II. Genderdifferences (mm) in overall, anterior,
Stepwise multiple regression analyses were used to and posterior segment lengths
sequentially identify the individual teeth that were most Gender
closely associated with the overall interarch ratio.
Segment F-ratio Significance Males Females
RESULTS Maxilla
Tables I and II show consistent population (P < Overall 7.53 0.007 97.9 96.0
.001) and gender (P < .019) differences in the lengths Anterior 5.60 0.019 48.3 47.3
Posterior 6.74 0.010 49.6 48.7
of all 3 arch segments (overall, anterior, posterior ) for Mandible
both the maxilla and mandible. The overall arch seg- Overall 15.89 <0.001 91.3 88.3
ments of whites were 3.6 mm to 4.8 mm smaller than Anterior 11.45 0.001 38.7 37.6
the arch segments of Hispanics or blacks, with the pos- Posterior 14.44 <0.001 52.7 51.2
terior segments accounting for most of the size differ-
ences. The arch segments of Hispanics and blacks tend-
ed to be similar and, with the exception of the mandibu- Table III. Population
differences (%) in overall, anterior,
lar posterior segment, the differences were not statisti- and posterior segment ratios ([mandibular segment/-
maxillary segment]*100)
cally significant. Male arch segments were also consis-
tently larger than female segments. The overall maxil- Populations
lary arch segment was 1.9 mm greater for males than Segment F-ratio Significance White Black Hispanic
for females; the mandibular segment was 3 mm larger
Overall 4.53 0.012 92.3 93.4 93.1
in males. The gender differences were distributed
Anterior 3.24 0.041 79.6 79.3 80.5
equally between the anterior and posterior segments. Posterior 9.65 <0.001 104.8 107.2 105.3
There were also significant population and gender
differences in the ratios between the upper and lower arch
segments (Table III). Post hoc tests showed that the over-
all ratio was significantly larger in blacks than in whites showed that maxillary segments explained 74% to 81%
only; blacks and Hispanics showed no significant differ- of the variation in overall mandibular segments. The
ences, nor did whites and Hispanics. The anterior ratio slope of whites was significantly smaller (P = .003) than
was significantly larger in Hispanics than blacks. Whites the slope of blacks only. Hispanic slopes did not differ
showed no differences in the anterior ratio from either significantly from either whites or blacks. Males also dis-
blacks or Hispanics. The posterior ratio demonstrated the played significantly (P = .019) larger slopes than females.
greatest population differences. Blacks had relatively Finally, multiple regression analysis was used to iden-
larger posterior mandibular teeth than either whites or tify the individual teeth that explained individual differ-
Hispanics, whose ratios were similar. Significant gender ences in the overall interarch ratio (Table VI). The tooth
differences (Table IV) were observed for the overall and most closely related with the overall ratio was the lower
posterior ratio, but not for the anterior ratio. second premolar. It accounted for approximately 10% of
A regression analysis was performed to further evalu- the variation; when the premolar was larger than average,
ate the relationships between the overall segment lengths the overall ratio tended to be larger than average.
(Table V). Correlations ranging between 0.86 and 0.94 Controlling for the size of the lower second premolars,
172 Smith, Buschang, and Watanabe American Journal of Orthodontics and Dentofacial Orthopedics
February 2000

Table IV. Gender

differences (%) in overall, anterior, and Whites showed the smallest overall ratio, followed by
posterior segment lengths ([mandibular segment/- Hispanics and blacks. The differences between blacks
maxillary segment]*100) and whites were highly significant and indicate that rel-
Gender ative to the maxillary teeth, the mandibular teeth of
Segment F-ratio Significance Males Females whites were smaller than the mandibular teeth of
blacks. The difference between whites and blacks
Overall 6.21 0.014 93.3 92.6 closely approximated the difference reported by
Anterior 2.18 0.142 80.1 79.5
Posterior 3.94 0.049 106.2 105.3
Lavelle.20 Lavelle also showed that blacks had the
highest overall ratio, whites had the smallest, with
Mongoloids (who have been anthropologically related
Table V. Least-squares regression of overall maxillary to Hispanics) in between. We were unable to statisti-
arch length (independent variable) with overall cally confirm the small differences in the overall ratio
mandibular arch length (dependent variable) between whites and Hispanics (0.7%) or between
Males Females Hispanics and blacks (0.3%). The difference between
Population R Slope SE R Slope SE
whites and blacks in the overall ratio was due primari-
ly to the size differences of the posterior teeth. The pos-
Black 0.87 0.938 0.004 0.94 0.931 0.003 terior maxillary segment was 2.3 mm larger in blacks
Hispanic 0.90 0.932 0.003 0.90 0.927 0.004 than in whites, whereas the posterior mandibular seg-
White 0.89 0.927 0.004 0.86 0.918 0.004
ment was 3.6 mm larger in blacks. Merz et al15 report-
ed similar results, with blacks having much larger pos-
terior teeth compared to whites. Our results showed
Table VI. Stepwise multiple regression showing teeth
most closely associated with individual differences in only small differences between blacks and whites in the
the overall ratio. anterior ratio, as previously reported by Merz. The
anterior ratio actually showed that whites have relative-
Regression step Tooth R Probability SE
ly larger anterior mandibular teeth than blacks. In con-
1 L5 0.29 <0.001 2.04 trast, Lavelle20 showed that blacks had larger anterior
2 U2 0.57 <0.001 1.76 ratios than whites. Even though the anterior ratios were
3 U5 0.63 <0.001 1.66 similar, both the maxillary and mandibular arch seg-
4 L1 0.73 <0.001 1.48
5 L6 0.77 <0.001 1.38
ments of blacks were larger than whites, as previously
6 U6 0.84 <0.001 1.18 demonstrated.15,25 The lack of differences in the over-
7 U1 0.86 <0.001 1.10 all ratio between Hispanics and blacks is deceiving
8 L2 0.89 <0.001 0.99 because there were significant differences in both their
9 L4 0.92 <0.001 0.88 anterior and posterior ratios. There was no difference in
10 U4 0.95 <0.001 0.69
11 L3 0.96 <0.001 0.60
the overall ratio because the larger anterior ratio of
12 U3 1.00 <0.001 0.13 Hispanics was counterbalanced by the larger posterior
ratio of blacks. The maxillary teeth of Hispanics were
only slightly smaller than the maxillary teeth of blacks.
The anterior mandibular teeth of Hispanics were larger
the upper lateral incisors were the second most important than the anterior mandibular teeth of blacks, whereas,
teeth explaining variation in the overall arch ratio, fol- the posterior mandibular teeth of blacks were larger.
lowed by the upper second premolars and the lower cen- Although none of the ratios showed significant differ-
tral incisors, respectively. In combination, these 4 teeth ences between Hispanics and whites, both maxillary
produced a multiple correlation of 0.73, explaining and mandibular tooth sizes were consistently larger in
approximately 50% of the variation between subjects. Hispanics, with significant differences for the posterior
The first molars explained an additional 20% of the vari- mandibular segment only.
ation. The canines and first premolars were the last teeth There were also significant differences between
to enter the regression, indicating that they were the least males and females. The difference in the overall ratio
likely to explain individual differences in the overall ratio. was due to both the anterior and posterior arch seg-
ment relationships, even though only the posterior
DISCUSSION ratio showed a significant difference. Posterior ratio
The 3 populations exhibited significantly different differences could be explained by the relatively larger
relationships between the lower and upper teeth. mandibular arch segments of males. Lavelle20 also
American Journal of Orthodontics and Dentofacial Orthopedics Smith, Buschang, and Watanabe 173
Volume 117, Number 2

reported relatively larger overall ratios in males com-

pared to females in white, black, and Mongoloid pop-
ulations. The tooth size data reported by Moorrees et
al12,13 also imply gender differences in the overall
ratio, which agrees with our findings. Importantly,
other studies16,19,21 with similar ratios for males and
females suggest that gender differences in the overall
ratio may be population specific.
On the basis of the foregoing, it is clear that the
Bolton’s ratios are not applicable for both genders or
across all populations. Bolton believed that his overall Fig 2. Absolute differences between overall maxillary
ratio represented an “ideal” that could be applied equal- and mandibular segment lengths.
ly to all occlusions. Occlusions that deviated from this
proposed ideal ratio possessed tooth size discrepancies
in 1 arch or the other.5-7 According to our results, as
well as those of Lavelle,20 black patients should have
higher overall ratios than white patients. Similarly, it
may be ideal for male patients to have higher overall
ratios than female patients.
Even though our results were based on patients with
malocclusions, there is no evidence that malocclusion
affects the Bolton overall ratio. Crosby and Alexander9
showed no differences in the overall and anterior ratios
between subjects with various Angle Class I, II
(Divisions 1 and 2), and Class II surgical malocclu- Fig 3. Differences between Bolton’s estimate of
sions. More recently, Freeman et al10 reported that the mandibular segment length and estimates derived from
mean overall and anterior ratios of 157 orthodontic the regression slopes.
patients were nearly identical to those of Bolton.
Because the type of malocclusion does not predispose range and closely approximates Bolton’s 8.1 mm aver-
a patient to a tooth size discrepancy problem, the dif- age. However, the average white male in our sample pos-
ferences between our results and Bolton’s cannot be sesses a tooth size discrepancy of 1.5 mm! Blacks and
explained by the orthodontic samples used. Hispanics showed interarch differences that fell well out
Bolton’s overall ideal ratio was significantly small- of Bolton’s suggested range. If one were depending on
er than all of our estimates, except those for white the Bolton’s analysis to predict tooth size discrepancies
females. This suggests that Bolton’s original sample for anyone other than white females, false interarch tooth
may have been composed primarily of white females size discrepancies could be diagnosed or true discrepan-
and implies that the Bolton ratio is only applicable to cies could be masked. Separate standards for different
white females. Although Bolton’s published results did populations and genders are clearly warranted.
not specify the number of males or females used or The regression slopes (Fig 3) provided independent
their ethnicity, it is well known that most orthodontic support that interarch relationships were significantly
patients during the 1950s were white females.26 Black different for whites and blacks and that Bolton’s ratios
and Hispanic males exhibited overall ratios that were are applicable only to white females. For any given
2% to 2.5% higher than Bolton’s average and fell out- maxillary segment length, blacks have larger mandibu-
side his 1 standard deviation confidence interval. lar segment lengths than whites; the differences
Although black and Hispanic females fell within the 1 increased from approximately 1 mm for smaller maxil-
standard deviation confidence interval, their ratios were lary lengths to over 1.5 mm for larger maxillary seg-
still significantly different from Bolton’s average. ments. The results were not able to detect the smaller
Based on Bolton’s analysis, arches should “ideally” Hispanic differences, which may be related with sam-
show a 7.5 to 9.5 mm difference (Fig 2), with the overall ple size and lack of discriminatory power. With the
maxillary segment being larger than the overall exception of white females, the slopes for all other sub-
mandibular segment. The white females in our study samples were significantly larger than the relationship
showed a 7.7 mm difference, which falls within the ideal defined by Bolton’s ratio. Black and Hispanic males
174 Smith, Buschang, and Watanabe
showed the greatest discrepancies; individuals with
large maxillary arch segments could differ by more
than 2 mm. The slopes clearly suggest that the indis-
criminate application of any single ratio could bias esti-
mates of interarch tooth-size relationships across popu-
lations, genders, and tooth sizes.
Finally, the results demonstrated that specific teeth
could be relied on to explain individual differences in the
overall interarch ratio. The mandibular second premolars
explained most of the observed difference in the interar-
ch discrepancy, followed by the maxillary lateral
incisors, maxillary second premolars, and lower central
incisors. Discrepancies in the sizes of these 4 teeth
accounted for approximately 50% of the interarch dis-
crepancies. Maxillary lateral incisors have previously
been shown to be quite variable in size and are frequent-
ly the reason for a tooth size discrepancy between arch-
es. Doris18 showed that the maxillary lateral incisors, as
well as maxillary second premolars, mandibular second
premolars, and mandibular canines were the most vari-
able and displayed the most pronounced group differ-
ences. Interestingly, the canines were less closely related
with interarch differences than the molars, which have
been shown to be least variable.18,25
CONCLUSIONS
Mesiodistal and labial-lingual tooth dimensions, as
well as overbite must all be considered for a complete
assessment of interarch occlusal relationships. With
respect to mesiodistal dimensions, 4 general conclu-
sions can be drawn.
1. Significant differences in the overall, anterior, and poste-
rior interach ratios between whites, blacks, and Hispanics
suggest that population specific standards are necessary
for clinical assessments.
2. Significant gender differences were shown for both the
overall and posterior ratios. The overall Bolton ratio only
applies to white females; new standards are warranted for
white males.
3. The relationships between the sizes of the mandibular and
maxillary teeth are dependent on population, gender, and
arch segment lengths. The larger the maxillary arch seg-
American Journal of Orthodontics and Dentofacial Orthopedics
February 2000
ment length, the greater the discrepancy between Bolton’s
ratios and the actual ratios.
4. The mandibular second premolars, the maxillary lateral
incisors, the maxillary second premolars, the mandibular
central incisors, and the first molars explain most of the
variation in the interarch discrepancy. These teeth should
be examined and dealt with first when an interarch tooth
size discrepancy is suspected.
REFERENCES
1. Kesling HD. The philosophy of the tooth positioning appliance. Am J Orthod
1945;31:297-340.
2. Howes AE. Case analysis and treatment planning based upon the relationship of the
tooth material to its supporting bone. Am J Orthod 1947;33:499-533.
3. Neff CW. Tailored occlusion with anterior coefficient. Am J Orthod 1949;35:309-13.
4. Neff CW. The size relationship between the maxillary and mandibular anterior seg-
ments of the dental arch. Angle Orthod 1957;27:138-47.
5. Bolton WA. Thesis for Masters Degree, University of Washington. Th 7180; 1952:
1-40.
6. Bolton WA. Disharmonies in tooth size and its relation to the analysis and treatment
of malocclusions. Angle Orthod 1958;28:113-30.
7. Bolton WA. The clinical application of a tooth size analysis. Am J Orthod
1962;48:504-29.
8. White LW. The clinical use of occlusograms. J Clin Orthod 1982;16:92-103.
9. Crosby DA, Alexander CG. The occurrence of tooth size discrepancies among differ-
ent malocclusion groups. Am J Orthod Dentofacial Orthop 1989;95:457-61.
10. Freeman JE, Maskeroni AJ, Lorton L. Frequency of Bolton tooth-size discrepancies
among patients. Am J Orthod Dentofacial Orthop 1996;110:24-7.
11. Lundstrom AA. Intermaxillary tooth width ratio analysis. Eur J Orthod 1981;3:285-7.
12. Moorrees CFA, Thomsen SO, Jensen E, Yen PK. Mesiodistal crown diameter of the
deciduous and permanent teeth in individuals. J Dent Res 1957;36:39-47.
13. Moorrees CFA. The Aleut dentition. Cambridge, Mass; Harvard University Press; 1957.
14. Bailit HL. Dental variation among populations. Dental Clinics of North America
1975;19:125-39.
15. Merz ML, Isaacson RJ, Germane N, Rubenstein LK. Tooth diameters and arch perime-
ters in a black and white population. Am J Orthod Dentofacial Orthop 1991;100:53-8.
16. Garn SM, Lewis AB, Swindler DR. Genetic control of sexual dimorphism in tooth
size. J Dent Res 1967;46:963-72.
17. Bishara SE, Jakobsen JR, Abdallah EM, Garcia AF. Comparisons of mesiodistal and
buccolingual crown dimensions of the permanent teeth in three populations from Egypt,
Mexico, and the United States. Am J Orthod Dentofacial Orthop 1989;96:416-22.
18. Doris JM, Bernard DW, Kuftinec MM. A biometric study of tooth size and dental
crowding. Am J Orthod 1981;79:326-36.
19. Richardson ER, Malhotra SK. Mesiodistal crown dimension of the permanent denti-
tion of American Negroes. Am J Orthod 1975;68:157-64.
20. Lavelle CLB. Maxillary and mandibular tooth size in different racial groups and in dif-
ferent occlusal categories. Am J Orthod 1972;61:29-37.
21. Buschang PH, Demirjian A, Cadotte L. Permanent mesiodistal tooth size of French-
Canadians. J Can Dent Assoc 1988;54:441-4.
22. Gillien RJ, Schwartz RS, Hilton TJ, Evans DB. An analysis of selected normative tooth
proportions. Int J Prosthod 1994;7:410-7.
23. Seipel CM. Variation of tooth position. Svensk Tandl.-Tidskrift 1946;39:Suppl.
24. Dahlberg G. Statistical methods for medical and biological students, London: George,
Allen, and Unwin; 1940.
25. Keene HJ. Mesiodistal crown diameters of permanent teeth in male American
Negroes. Am J Orthod 1979;76:95-9.
26. Proffit WR. Forty year review of extraction frequencies at the University Orthodontic
Clinic. Angle Orthod 1994;64:407-14.