ORIGINAL ARTICLE

Assessment of facial golden proportions among

young Japanese women
Yasushi Mizumoto,a Toshio Deguchi, Sr,b and Kelvin W. C. Fongc
Chiba and Nagano, Japan, and Singapore

Introduction: Facial proportions are of interest in orthodontics. The null hypothesis is that there is no differ-
ence in golden proportions of the soft-tissue facial balance between Japanese and white women.
Methods: Facial proportions were assessed by examining photographs of 3 groups of Asian women: group
1, 30 young adult patients with a skeletal Class 1 occlusion; group 2, 30 models; and group 3, 14 popular
actresses. Photographic prints or slides were digitized for image analysis. Group 1 subjects had standardized
photos taken as part of their treatment. Photos of the subjects in groups 2 and 3 were collected from
magazines and other sources and were of varying sizes; therefore, the output image size was not considered.
The range of measurement errors was 0.17% to 1.16%. ANOVA was selected because the data set was nor-
mally distributed with homogeneous variances. Results: The subjects in the 3 groups showed good total facial
proportions. The proportions of the face-height components in group 1 were similar to the golden proportion,
which indicated a longer, lower facial height and shorter nose. Group 2 differed from the golden proportion,
with a short, lower facial height. Group 3 had golden proportions in all 7 measurements. The proportion of
the face width deviated from the golden proportion, indicating a small mouth or wide-set eyes in groups 1
and 2. Conclusions: The null hypothesis was verified in the group 3 actresses in the facial height components.
Some measurements in groups 1 and 2 showed different facial proportions that deviated from the golden pro-
portion (ratio). (Am J Orthod Dentofacial Orthop 2009;136:168-74)

T
he 1992 textbook entitled Esthetics and the Treat- used the term ‘‘golden sectioning.’’ Applying a divider,
ment of Facial Form1 includes interesting and the divine proportion is the length of the longer side in
important references related to the history of fa- 2 linear measurements, at 1.618, and the short side is 1
cial beauty, the evolution of the concepts of facial struc- (Fig 1).
ture, and well-known analyses of the soft-tissue profile. Proffit and Fields6 wrote that the vertical height of
More recent articles discussed angular and linear analy- the midface, from the supraorbital ridges to the base of
ses of the soft-tissue profile, including ethnic differ- the nose, should equal the height of the lower face, and
ences.2,3 The golden proportion has been well known in the lower face, the mouth should be about one third
for hundreds, perhaps thousands, of years, but Ricketts4,5 of the way between the base of the nose and the chin.
might have been the first orthodontist to apply it to the Facial proportions are of interest in orthodontics and
composition of facial hard and soft tissues. He also maxillofacial and plastic surgery.1,7,8 Japanese litera-
ture includes several reports of facial balance in plastic
a
and cosmetic surgery, but there is little information to
Private practice, Chiba, Japan.
b
Professor, Orthodontic Postgraduate Program, Department of Preventive Den- evaluate the proportions of the frontal facial profile.9-11
tistry, Faculty of Dentistry, National University of Singapore and Hospital, Sin- Since Ricketts reported the divine proportion in terms of
gapore; professor and head, Unit of Orthodontics, Department of Hard Tissue facial proportions, only 1 original data set has been
Research, Graduate School of Oral Medicine, Matsumoto Dental University,
Shiojiri, Nagano-ken, Japan. obtained for Japanese subjects, and it was to establish
c
Associate professor and director, Orthodontic Postgraduate Program, Depart- a Japanese norm for surgical treatment goals for patients
ment of Preventive Dentistry, Faculty of Dentistry, National University of Sin- with facial asymmetry and mandibular protrusion.8
gapore, Singapore.
The authors report no commercial, proprietary, or financial interest in the prod- The purposes of this study were to measure soft-tis-
ucts or companies described in this article. sue facial proportions in 3 groups of Japanese women
Reprint requests to: Toshio Deguchi, Postgraduate Orthodontic Program, and to compare these values to the golden proportion.
Department of Preventive Dentistry, National University of Singapore and
Hospital, Level 3, Dentistry Block, 5 Lower Kent Ridge Rd, Singapore 11974; Ricketts’ golden ratio was used as 1 representative
e-mail, pndtd@nus.edu.sg. facial balance analysis in this study. In Japan, the silver
Submitted, January 2007; revised and accepted, July 2007. ratio is well known as the proportion for traditional tem-
0889-5406/$36.00
Copyright Ó 2009 by the American Association of Orthodontists. ple construction, as well as for portraits and statues; we
doi:10.1016/j.ajodo.2007.07.029 also applied the silver ratio in this study.12,13
168
American Journal of Orthodontics and Dentofacial Orthopedics Mizumoto, Deguchi, and Fong 169
Volume 136, Number 2

Fig 1. The golden ratio is divided.

MATERIAL AND METHODS

The study sample comprised photographs of 3 groups
of Japanese women. Group 1 consisted of 30 treated Fig 2. Method for taking a frontal photo of the patients.
orthodontic patients, group 2 consisted of 30 magazine
models, and group 3 included 14 well-known actresses.
To correct anterior crowding, 24 group 1 patients had Three measurements for the proportions of total face
2 or 4 first premolars extracted; the other 6 were treated heights were made: 7 for the face height components
without extractions. All 30 had good treatment results and 3 for face widths. All landmarks were measured
with acceptable mesofacial, skeletal Class I patterns, once per day for 5 days by an author (Y.M.). The range
and good profiles. The mean age at posttreatment (or of measurement error was 0.17% (TR-ME distance,
retention) when records were gathered was 17.0 years group 1) to 1.16% (LN-CH, group 1). ANOVA was
(SD, 3.3 years). Photographs (slides) were taken in a used because the data were normally distributed with
photographic studio at the Matsumoto Dental Hospital, homogeneous variances.
Shiojiri, Japan, by a professional photographer at a fixed The following landmarks were used (Fig 3): trichion
distance between the camera and the subject (Fig 2). The (TR), the point at the top of the forehead at the junction
patients were photographed with the lips in repose. (hairline) of the face and skull fascia; TS, the width of
The ages of the subjects in groups 2 and 3 could not the head at the temporal soft tissue above the ears at
be obtained, but we tried to find models and actresses of the level of the supraorbital ridges (eyebrow); LC, the
similar ages to allow comparisons with group 1. Anterior point at the lateral canthus of the eyes; LN, the point
facial photographs of subjects in groups 2 and 3 were ob- at the lateral rim of the nose; cheilion (CH), the point
tained from magazines. Although these photos were not at the corner of the mouth; ME, soft-tissue menton.
standardized, each was suitable for measuring the pro- Facial height was measured on the anterior face (Fig
portions of the anterior face and was selected based on 3): 1, TR-ME; 2, LC-ME; 3, TR-LN; 4, TR-LC; 5, LN-
the following criteria: mouth closed, lips in repose, min- ME; 6, LC-CH; 7, CH-ME; 8, LC-LN; and 9, LN-CH.
imal cosmetics, landmarks visible, and minimal devia- Facial width on the anterior face was measured (Fig
tion of the anterior face from normal head position. 4): 10, TS-TS; 11, LC-LC; 12, LN-LN; and 13, CH-CH.
Slides and photographs were digitized for analysis in The facial balance in height and width was studied
an image processing program. Slides were scanned with from the measurements as follows: proportion of total
a film scanner (Super Coolscan 4000ED, Nikon, Tokyo, face height (Fig 3) with TR-LC:TR-ME; LN-ME:TR-
Japan) and converted to TIFF images with 2.99 Mb and ME; and LC-CH:TR-ME.
240 pixels per inch resolution by using Adobe Photoshop The proportion of face height components was
(version 7.0.1, Adobe, San Jose, Calif). Photographs studied (Fig 3) with TR-ME:LC-ME; TR-LC:LC-ME;
were scanned in a flatbed scanner (GT-6500 Epson, LN-ME:TR-LN; LC-LN:LN-ME; CH-ME:LC-CH;
Shiojiri, Japan) and saved as TIFF files with a resolution LN-CH:LC-LN; and LN-CH:CH-ME.
of 100 dpi. Because the photographs in groups 2 and 3 The proportion of face width components was
were of different sizes, the output image size was not con- measured (Fig 4): LNr-l:CHr-l; LCr-l:CHr-l; and
sidered; sizes were used that did not damage the measure- TSr-l:LCr-l.
ment points. Measurements were taken by using image Mean measurements were converted to percentages,
processing software (NIH Image, version 1.62, National assuming that the divine proportion was 100%.
Institutes of Health, Bethesda, Md). In setting the soft- In Japan, another ratio of balanced proportion is of-
ware, the measurements of the lengths were calculated. ten used: the silver ratio.12,13 The silver ratio is 1:1.414
170 Mizumoto, Deguchi, and Fong American Journal of Orthodontics and Dentofacial Orthopedics
August 2009

Fig 3. Landmarks and measurements for anterior facial Fig 4. Landmarks and measurements for facial width
height analysis: A, TR; B, LC; C, LN; D, CH; E, ME; M, analysis: Fr-l 5 TS r-l; Gr-l 5 LC r-l; Hr-l 5 LN r-l; and
base point line. Three measurements for proportion of Ir-l 5 CH r-l. Three measurements of facial width compo-
total anterior height: 1, TR-LC:TR-ME; 2, TR-ME:LN- nents: 1, LN r-l:CHr-l; 2, CHr-l:LCr-l; 3, LCr-l:TSr-l.
ME; 3, TR-ME:LC-CH. Seven measurements of face
height components: 1, TR-ME:LC-ME; 2, TR-LC:LC-
ME; 3, LN-ME:TR-LN; 4, LC-LN:LN-ME; 5, CH-ME:LC- of LN-CH:LC-LN (9:8) and LC-LN:LN-ME (8:5) devi-
CH; 6, LN-CH:CH-ME; 7, LN-CH:LC-LN. ated from the divine proportion.
In group 2 (models), LN-CH:LC-LN, LC-LN: LN-
ME, and TR-ME:LC-ME were 97% to 102%, similar to
(O2). We also used this measure as a balanced ratio for the divine proportion. The value of LN-CH:CH-ME was
the Japanese subjects. small (1.5, 93%) but close to the divine proportion. In con-
trast, TR-LC:LC-ME, LN-ME:TR-LN, and CH-ME:LC-
RESULTS CH deviated more: 1.55 (96%), 1.69 (104%), and 1.72
(106%), respectively. All 7 measurements in group 3
ANOVA was applied for statistical analysis of the
(actresses) were similar to the divine proportions.
measurements in the 3 groups (Table). The means of
In this study of face width. LC-LC:TS-TS (10:11) in
the measurements were also converted to percentages.
group 1 was similar to the divine proportion, but the
Three measurements—TR-LC:LN-ME:LC-CH — other measurements in groups 1 and 2 were different
showed similar proportions (means, 2.59, 2.52 and from the divine proportion.
2.69, respectively) to the total face height (TR-ME); The results of the Table were plotted on a graph
therefore, the proportion of these 3 measures was close (y-axis, ratio; x-axis, 7 measurements) to illustrate the
to 1:1:1. The mean of LN-ME:TR-ME was larger in
differences in the groups (Figs 5 and 6). The silver ratio
group 2 (2.81) than in group 1 (2.52); this indicates
(1.4:1) was plotted on a graph to assess how well all
a shorter length of LN-ME in group 2.
measurements fit this ratio.
The Table shows the proportion of the facial height
components. Group 1 patients were similar to the divine
proportion (100%) in the values of TR-ME:LC-ME DISCUSSION
(1:2) and TR-LC:LC-ME (4:2). In addition, LN- Peck and Peck14 reviewed data related to historical
ME:TR-LN (5:3), CH-ME:LC-CH (7:6), and LN- art collections and discussed the scientific evaluation
CH:CH-ME (9:7) had ratios of 94% to 97%. The values of dental occlusions and faces with the golden ratio.
American Journal of Orthodontics and Dentofacial Orthopedics Mizumoto, Deguchi, and Fong 171
Volume 136, Number 2

Table. Facial analysis with golden proportion at total facial proportion, height, and width dimensions
Group 1 (n 5 30) Group 2 (n 5 30) Group 3 (n 5 14) P value*

Mean SD Mean SD Mean SD Groups 2 vs 3 Groups 1 vs 3 Groups 1 vs 2

Facial proportion
TR-LC:TR-ME 2.594 0.141 2.470 0.125 2.578 0.134 — — —
TR-ME:LN-ME 2.521 0.110 2.810 0.151 2.667 0.139 — — —
TR-ME:LC-CH 2.694 0.142 2.569 0.109 2.675 0.153 — — —
Vertical height
TR-ME:LC-ME 1.631 0.056 1.647 0.044 1.643 0.030 0.9629 0.7206 0.4003
100.8% 101.8% 101.6%
TR-LC:LC-ME 1.595 0.141 1.552 0.115 1.560 0.093 0.9863 0.6619 0.4064
98.6% 95.9% 96.4%
LN-ME:TR-LN 1.522 0.109 1.687 0.120 1.647 0.137 0.6101 0.0064 S \0.0001 S
94.0% 104.3% 101.8%
LC-LN:LN-ME 1.844 0.180 1.598 0.184 1.648 0.226 0.6944 0.0101 S \0.0001 S
113.9% 98.7% 101.8%
CH-ME:LC-CH 1.535 0.136 1.715 0.178 1.599 0.185 0.0974 0.4841 0.0003 S
94.8% 106.0% 98.8%
LN-CH:CH-ME 1.572 0.145 1.505 0.174 1.635 0.188 0.095 0.6301 0.3038
97.1% 93.0% 101.1%
LN-CH:LC-LN 1.399 0.131 1.573 0.184 1.619 0.220 0.8505 0.0018 S 0.0008 S
86.4% 97.2% 100.1%
Horizontal width
LNr-l:CHr-l 1.183 0.098 1.305 0.099 1.407 0.103 0.009 S \0.0001 S \0.0001 S
73.1% 80.7% 87.0%
CHr-l:LCr-l 2.084 0.191 1.945 0.148 1.911 0.209 0.8418 0.0145 S 0.0136 S
128.8% 120.2% 118.1%
LCr-l:TSr-l 1.563 0.064 1.469 0.074 1.481 0.061 0.847 0.0016 S \0.0001 S
96.6% 90.8% 91.5%

*Statistical analysis by 1-factor ANOVA (post hoc: Scheffé) significance level (S) P \ 0.05.

They pointed out a problem with historical subjects used Japanese skeletal pattern.18,19 One characteristic of
for the divine proportion and also reported that the gen- skeletal Class II and Class III patterns in Japanese20,21
eral public prefers a fuller, more protrusive profile than and Chinese22 subjects could be a backward rotated
customary cephalometric standards indicate.15 The mandible with a steeper mandibular plane angle.
esthetic female facial profile is not static and showed When assessing facial profile attractiveness in Chi-
a trend in the 1990s toward fuller, more anteriorly posi- nese23 and Japanese24 subjects, dental professionals
tioned lips in fashion magazines.16 Considering ethnic- prefer a flatter profile or a retruded mandible. The hypo-
ity, Barnes et al17 reported that the faces of North divergent lower face height in group 2 tended to have
American black men showed a strong statistical ten- a small mandible in the retruded position.
dency toward the golden ratio. Studying 30 young Asian women and 30 young Asian
The greatest weakness of our study was the inclu- men with normal skeletal morphology, Kawakami et al9
sion of nonstandardized photographs for group 2 (model reported that the values of CH-ME:LC-CH, AL (same
magazines) and group 3 (monthly magazines) subjects. as LN in this study)-CH:CH-ME, and AL(LN)-CH:LC-
Nevertheless, total facial profile proportions and face AL(LN) showed deviations from the golden ratio, and
height and width components were studied in 3 groups the deviations were greater in men. The deviations indi-
of Japanese women: well-treated patients with accept- cated a shorter AL(LN)-CH and a longer CH-ME than
able mesofacial skeletal Class I patterns, models influ- those of white people (the golden ratio).
enced by the fashion media, and popular actresses in Five of the 7 measurements of face height in group 1
Japan. Three measurements of the total facial propor- were close to the divine proportion, whereas LC-LN/LN-
tion were close to 1:1:1 ratio, which indicated good pro- ME (8:5) and LN-CH/LC-LN (9.8) deviated from the
portions in the 3 groups. In comparison with groups 1 divine proportion (Table). Those deviations could be
and 3, group 2 models had shorter lower anterior facial related to a longer CH-ME and a shorter LN-CH; these
heights (CH-ME and LN-ME); this hypodivergent are the same as the results of Kawakami et al.9 In group
lower face pattern is not generally observed in the 2, 4 of 7 measurements were close to the golden
172 Mizumoto, Deguchi, and Fong American Journal of Orthodontics and Dentofacial Orthopedics
August 2009

Fig 5. Graph of face height components in group 1 (patients), group 2 (models), and group 3 (ac-
tresses). Note all 7 measurements with golden ratio in group 3. Y-axis is the ratio, and x-axis is the
7 measurements.

Fig 6. Graph of facial width components in 3 groups. X-axis is the 3 measurements.

proportion, and LN-ME:TR-LN, CH-ME:LC-CH, and similar to the golden proportion than did the group 2
LN-CH:CH-ME (5:4, 7:6, and 9:7, respectively) deviated models. These results indicate that the divine proportion
from the golden proportion. All measurements that devi- might not be the limiting factor in determining facial
ated from the golden proportion were related to lower balance. Recently, fashion trends have favored small
anterior facial height (CH-ME and LN-ME) in group 2 faces, especially small jaws. As a result, the group 2
(Table, Fig 5). Group 2 subjects had shorter anterior lower models might have had relatively small jaws. The group
facial height. All measurements in groups 1 and 2 were 3 actresses showed golden proportions in all 7 measure-
significantly different (P \ 0.05). ments of facial height (Table, Fig 5); this indicates that
The models in group 2 generally had more balanced the golden proportion provides valuable information on
faces compared with average young women. However, balanced facial proportions of the most popular ac-
the group 1 patients had more measurements that were tresses in Japan. Moreover, those results indicate that
American Journal of Orthodontics and Dentofacial Orthopedics Mizumoto, Deguchi, and Fong 173
Volume 136, Number 2

the golden proportion might be useful when applied to CONCLUSIONS

balanced facial proportions in young Japanese women. The null hypothesis that there is no difference in
Proffit and White7 reported a ratio of the widths of golden proportions of the soft-tissue facial balance
the orbits, eyes, nose, mouth, and mandible in frontal between Japanese and white women was accepted
facial symmetry. Some findings are that the width of 100% for the group 3 actresses in facial height compo-
the base of the nose should be approximately the same nents. The golden and silver ratios indicated that young
as the distance between the inner canthi, whereas the Japanese women have well-balanced facial height pro-
width of the mouth should be approximately the dis- portions, except for a few measurements. Lower facial
tance between the irises. widths deviated from the golden proportion.
In this study of face width, LC r-l:TS r-l (11:10) in
group 1 was similar to the golden proportion, indicating We thank Toru Kageyama, Graduate Orthodontics,
that the upper anterior face width might be well bal- Matsumoto Dental University, for assisting with the sta-
anced. The value of CH r-l:LCr-l (13:11) in groups 1 tistical analyses.
and 2 deviated more (129% and 120%, respectively) REFERENCES
and could indicate small lip width with normal eye
1. McNamara JA. Esthetics and the treatment of facial form.
width. The LN r-l:CHr-l (12:13) ratios were smaller Proceedings of the Annual Symposium on Craniofacial Growth.
(73% and 81%, respectively). This smaller ratio might Volume 28. Craniofacial Growth Series. Ann Arbor: Center for
predict the combination of small lip width and average Human Growth and Development; University of Michigan;
nose width. Three measurements of face width differed 1992. p. 1-39, 77-146.
2. Fernan-Riveiro P, Suarez-Quintanilla D, Smyth-Chamosa E,
significantly in all 3 groups (Table, Fig 6).
Suarez-Cunqueiro M. Linear photogrammetric analysis of the
Yanagi12,13 introduced the silver ratio. He reported soft tissue facial profile. Am J Orthod Dentofacial Orthop 2002;
that a historical Japanese painting (Fuji Sanjuurokkei 122:59-66.
painted by Katsushika Hokusai) and the Horyu-ji tem- 3. Fernandez-Riveiro P, Smyth-Chamosa E, Suarez-Quintanilla D,
ple (in Nara city, the oldest capital) built in 607 AD Suarez-Cunqueiro M. Angular photogrammetric analysis of the
were designed and constructed with both golden soft tissue facial profile. Eur J Orthod 2003;25:393-9.
4. Ricketts RM. The golden divider. J Clin Orthod 1981;15:752-9.
(1:1.618) and silver (1:O2 5 1.414) ratios. The golden 5. Ricketts RM. Divine proportion and Fibonacci series. Am J
ratio was introduced to China and Japan via the Silk Orthod 1982;81:351-70.
Road. Those ratios were devices, sashigane, that are still 6. Proffit WR, Fields HW. Contemporary orthodontics. 3rd ed.
used today (eg, in home building and camera lenses). If St Louis: Mosby; 1997. p. 157-63.
either the golden or the silver ratio indicates a well-bal- 7. Proffit WR, White RP Jr. Surgical-orthodontic treatment.
St Louis: Mosby-Year Book; 1991. p. 105-28.
anced proportion, most measurements, except LC- 8. Ferrario VF, Sforza C, Poggio CE, Tartaglia G. Facial morphom-
LN;LN-ME (8:5) in group 1 and LN-ME:TRI-LN etry of television actresses compared with normal women. J Oral
(5:3) and CH-ME:LC-CH (7:6) in group 2, indicate Maxillofac Surg 1995;53:1008-14.
that facial height proportions are well balanced in young 9. Kawakami S, Tsukada S, Hayashi H, Takada Y, Koubayashi S.
Golden proportion for maxillofacial surgery in Orientals. Ann
Japanese women (Fig 5).
Plast Surg 1989;23:417-25.
Ferrario et al8 studied the 3-dimensional coordinates 10. Nanba K. Changes in the Japanese face over history, with refer-
of standardized soft-tissue facial landmarks in televi- ence to the lower third of the face. Jpn J Plast Reconstr Surg
sion actresses and normal women and found significant 1992;35:167-78, 235-47.
differences. The actresses had larger foreheads, larger 11. Shibamoto H, Hata Y, Yano K, Matsuka K, Ito O. A geometric
middle facial thirds relative to the total face, wider study of the facial features of Japanese people. J Jpn Plast
Reconstr Surg 1992;12:429-42.
dimensions and deeper faces, smaller noses, and less 12. Yanagi A. Division of gold between the pyramids and Le Corbus-
facial convexity. ier. 28th ed. Tokyo: Bijitsusyuppan; 1965. p. 26-30, 62-9.
To understand the ethnic differences in the soft-tis- 13. Yanagi A. Sequel to division of gold between the Horyu Temple
sue profile, the same terminology and landmarks with and Ukioe print: a Japanese comparison. 12th ed. Tokyo: Bijitsu-
standardized photographs in young people should be syuppan; 1967. p. 27-40.
14. Peck S, Peck L. Selected aspects of the art and science of facial
used from a clinical viewpoint in orthodontics. esthetics. Semin Orthod 1995;1:105-26.
Clifton Olds said that ‘‘The history of the human 15. Peck H, Peck S. A concept of facial esthetics. Angle Orthod 1970;
face in art is a fascinating one, and one that is intricately 40:284-317.
interwoven with the history of science, mathematics, 16. Auger TA, Turley PK. The female soft tissue profile as presented
medicine, and philosophy. It is also intimately involved in fashion magazines during the 1900s: a photographic analysis.
Int J Adult Orthod Orthognath Surg 1999;14:7-18.
with questions of public and private morality, and on 17. Barnes EM, Russell DM, Kudlick EM. A soft-tissue study of
that level the history of human beauty remains to be North American blacks utilizing the golden proportions [thesis].
written.’’1 Washington DC: Howard University; 1985.
174 Mizumoto, Deguchi, and Fong American Journal of Orthodontics and Dentofacial Orthopedics
August 2009

18. Miyajima K, McNamara JA Jr, Kimura T, Murata S, Iizuka T. Cra- 21. Ishii N, Deguchi T, Hunt NP. Craniofacial difference between Jap-
niofacial structure of Japanese and European-American adults anese and British Caucasian females with a skeletal Class III mal-
with normal occlusions and well-balanced faces. Am J Orthod occlusion. Eur J Orthod 2002;24:493-9.
Dentofaical Orthop 1996;110:431-8. 22. Lau JW, Hãgg U. Cephalometric morphology of Chinese with
19. Deguchi T, Mimura H, Togari A. Comparison of body height and Class II division 1 malocclusion. Br Dent J 1999;27:188-90.
mandibular length between Caucasian and Japanese children. 23. Soh J, Chew MT, Wong HB. Professional assessment of facial
Aust Orthod J 1993;13:23-8. profile attractiveness. Am J Orthod Dentofacial Orthop 2005;
20. Ishii N, Deguchi T, Hunt NP. Morphological differences in the 128:201-5.
craniofacial structure between Japanese and Caucasian girls 24. Ioi H, Nakata S, Nakashima A, Counts AL. Anteroposterior lip
with Class II division malocclusions. Eur J Orthod 2002;24: positions of the most-favored Japanese facial profiles. Am J
61-7. Orthod Dentofacial Orthop 2005;128:206-11.