Frontal (A) and lateral (B) views
of the average African American face. Surface landmarks are denoted. Explanation
of the abbreviations is given in the first footnote
to Table 1.
Two-section facial profile. The
average African American woman has a special facial height (endocanthion-gnathion
[en-gn]) of less than the special head height (vertex-endocanthion [v-en]),
although the proportions are roughly equal.
Three-section facial profile.
Average proportions are illustrated for African American women. The forehead
height II (trichion-nasion [tr-n]) is roughly equal to the lower face height
(subnasale-gnathion [sn-gn]). Nose length (nasion-subnasale [n-sn]) is significantly
shorter than both measures.
Four-section facial profile. The
average proportions illustrate that lower facial height (subnasale-gnathion
[sn-gn]) is greater than special upper face height (glabella-subnasale [g-sn]),
which is greater than forehead height I (trichion-glabella [tr-g]). v-tr indicates
height of calva (vertex-trichion).
Nasoaural proportion. The average
African American woman has an ear length (superaurale to subaurale [sa-sba])
of greater than the nose length (nasion-subnasale [n-sn]). The ratio is approximately
Orbitonasal proportion. The average
relationship for African American women is depicted. The ratio of interocular
distance (endocanthion-endocanthion [en-en]) to alar width (alare-alare [al-al])
is approximately 4:5.
Nasofacial proportion. The relationship
depicted is the average for African American women. Twenty-eight percent of
the facial width (zygion-zygion [zy-zy]) is represented by the nose width
Nasoaural inclination. The nasal
bridge inclination is more horizontal than the ear inclination.
Porter JP, Olson KL. Anthropometric Facial Analysis of the African American Woman. Arch Facial Plast Surg. 2001;3(3):191-197. doi:
From the Bobby R. Alford Department of Otorhinolaryngology and Communicative
Sciences, Baylor College of Medicine, Houston, Tex.
Copyright 2001 American Medical Association. All Rights Reserved.
Applicable FARS/DFARS Restrictions Apply to Government Use.2001
Objective To assess the differences in facial proportions between African American
and Caucasian women. Differences within the African American population are
Design Anthropometric survey.
Participants Volunteer sample of African American women (N = 108), aged 18 through
30 years, with African American parents and no previous facial surgery or
Intervention Photographs and 16 standard anthropometric measurements were taken in
concordance with the 9 neoclassical canons. Results were compared with the
North American white standard and the neoclassical canons, and an intragroup
evaluation was performed. One-way analysis of variance, 99.7% confidence intervals,
and t tests were used to test differences for significance.
Main Outcome Measures Anthropometric measures.
Results Compared with white women, the following measurements were found to
be significantly different (P<.003) in African
American women: special head height was shorter; forehead height II was longer;
nose length was shorter; lower face height was longer; height of the calva
was shorter; forehead height I was longer; and ear length was shorter. In
addition, most horizontal measures were wider, ie, eye-fissure width, nasal
width, mouth width, and facial width. The nose and ear have greater angles
of inclination. Of the 9 neoclassical canons, the orbital proportion was found
to include the most proportional subjects (30.6%), followed by the nasoaural
proportion (13.0%) and the nasofacial proportion (9.3%). Subcategorization
based on nasal dorsal height yielded the most significantly different measures.
Conclusions African American female facial anthropometric measures, especially those
of the horizontal dimension, differ significantly from those of young white
subjects. The average African American woman does not fit the neoclassical
standard of facial proportion.
PROPORTIONAL evaluation of the face is used by surgeons during the planning
stages of facial plastic surgery. These normal proportions are used to critique
the face during consultation for rejuvenative or cosmetic changes. Facial
analysis and proportions are well described for North American white subjects.1-2 Proportional evaluation of the face
stems from the neoclassical canons of facial proportion, developed by artists
and anatomists of the 17th and 18th centuries. These canons describe the aesthetic
proportional relationships of the face and are the foundation on which modern
facial analysis is based.
Often, the African American patient is compared with the white patient
when the face is analyzed, despite inherent differences in physical appearance.
Furthermore, others have combined the African American patient with various
groups denoted by "non-Caucasian" or "ethnic" when facial analysis is discussed.3-4 As facial cosmetic surgery is becoming
more common among people from various ethnic backgrounds, the concept of a
single aesthetic standard of beauty is inadequate.
Several studies have evaluated anthropometric differences between racial
groups. Nasal analysis has been examined in both the African American and
Facial analysis by means of photogrammetry and anthropometry has also been
evaluated in Asian populations.8-9
Jeffries et al10 evaluated the African American
face using photogrammetric analysis, an indirect means of measuring the structures
of the face. Although anthropometric analysis of the African American and
black Caribbean faces was first reported by Farkas et al,11
proportional analysis, applications to facial analysis, and subgroups within
the population were not noted.
This study assesses the differences in facial proportion between young
African American and North American Caucasian women. Average anthropometric
data for young African American women are presented. In addition, the results
are compared with the neoclassical canons of facial proportions. Subcategorization
of the sample population of African American women is discussed.
One hundred eight African American female volunteers participated in
this study, which was approved by our institutional review board. Subjects
included in the study were required to be 18 through 30 years of age to minimize
the effects of aging on the facial proportions. Other inclusion criteria consisted
of both parents of African American heritage, no previous plastic or reconstructive
surgery of the face, no major trauma to the face, body mass index (calculated
as weight in kilograms divided by the square of height in meters) of no greater
than 27, and no history of craniofacial syndromes.
After obtaining informed consent, demographic data were obtained, including
age, weight, height, place of birth, and parental heritage. Standard photographs
of the face were obtained, including the frontal, right and left lateral,
right and left oblique, and base views. Photographs were analyzed for face
shape, classification as described by Ofodile et al5
(hereafter referred to as the Ofodile classification), nostril shape, and
All measurements were obtained by the same investigator (J.P.P.). Surface
landmarks were noted on the face before taking standard anthropometric measurements
Sixteen standard anthropometric measurements were obtained, including special
head height (vertex-endocanthion), special face height (endocanthion-gnathion),
forehead height II (trichion-nasion), nose length (nasion-subnasale), lower
face height (subnasale-gnathion), height of calva (vertex-trichion), forehead
height I (trichion-glabella), special upper face height (glabella-subnasale),
ear length (superaurale to subaurale), interocular distance (endocanthion-endocanthion),
nose width (alare-alare), eye-fissure width (exocanthion-endocanthion), mouth
width (cheilion-cheilion), facial width (zygion-zygion), nasal bridge inclination,
and ear inclination. Linear measurements are reported in millimeters, and
inclinations are expressed in degrees.
The 9 neoclassical canons of facial proportion were examined, including
vertically (canons I-IV) and horizontally (canons V-XIII) oriented measures
and angles of inclination (canon IX). The formulas for the 9 canons are as
follows: canon I, 2-section facial profile (vertex-endocanthion = endocanthion-gnathion);
canon II, 3-section facial profile (trichion-nasion = nasion-subnasale = subnasale-gnathion);
canon III, 4-section facial profile (vertex-trichion = trichion-glabella =
glabella-subnasale= subnasale-gnathion); canon IV, nasoaural proportion (superaurale-subaurale
= nasion-subnasale); canon V, orbitonasal proportion (endocanthion-endocantion
= alare-alare); canon VI, orbital proportion (exocanthion-endocanthion= endocanthion-endocanthion);
canon VII, naso-oral proportion ([alare-alare] × 1.5 = cheilion-cheilion);
canon VIII, nasofacial proportion ([zygion-zygion] × 0.25 = alare-alare);
and canon IX, nasoaural inclination (nasal bridge inclination = ear inclination).
In addition, the results were compared with the North American Caucasian
standard,1 and an intragroup evaluation was
performed. Data were entered on spreadsheets and analyzed using commercially
available software (SPSS version 8.0; SPSS Inc, Chicago, Ill). We used 99.7%
confidence intervals to assess differences between the results of our sample
population and those of others; overall chance of type I error is P<.05. Differences between subgroup means of our sample population
were assessed using t tests and 1-way analysis of
variance. Proportional relationships were said to exist if the difference
was not greater than 1 mm or 2°.
One hundred twelve African American women were enrolled in the study;
of these, 4 were excluded because of failure to meet the inclusion criteria.
Two women were excluded because of parental heritage; 1 woman, because of
failure to meet the height and weight requirements; and 1 woman, because of
auricular deformity. African American women (n = 108) enrolled in the study
had an average age of 25.0 years (range, 18-30 years). The average height
and weight were 164.6 cm and 62.7 kg, respectively. Most subjects (68.5%)
were born in our geographic region (Texas, Louisiana, Arkansas, and Mississippi).
The results of the anthropometric measurements are summarized in Table 1. A significant difference existed
between the African American and the North American Caucasian norms in 13
of the 16 measures taken. We noted that the African American nasal width was
wider and the nasal bridge was less inclined (P<.003).
No one subject fit all of the average measures or proportions.
All 9 formulas for the neoclassical canons of the face were evaluated
for the subjects tested, the results of which are presented in Table 2. Very few subjects fit the neoclassical canons. Compliance
with the 2-section facial profile (canon I) was noted in very few subjects.
Most subjects had a special face height of less than the special head height
(Figure 2). None of the subjects
met the criteria for the 3-section facial profile (canon II). The average
relationship of canon II is illustrated in Figure 3. All of the subjects had a lower facial height and a forehead
height II of greater than the nose length. Most subjects had a forehead height
II of greater than the lower facial height. Lower facial height is greater
than the forehead height II in approximately one third of subjects. None of
the subjects fulfilled the proportional criteria for the 4-section facial
profile (canon III). Lower facial height was always greater than the height
of the calva (Figure 4). In addition,
the height of the calva was smaller than all other canon measures (forehead
height I, special upper facial height, and lower facial height) in most subjects.
In approximately a quarter of the subjects, the forehead height I was greater
than the special upper facial height. The forehead height I was shorter than
the lower facial height in most subjects. The special upper facial height
was smaller than the lower facial height in three quarters of the subjects.
Nasoaural proportion, as established by neoclassical canon IV, was found in
very few subjects. Most subjects had an ear length that was greater than their
nasal length (Figure 5). Assessment
of orbitonasal proportion (canon V) showed that alar width was rarely equal
to or was less than the interocular distance. Most had an alar width of greater
than the interocular distance, as depicted in Figure 6. The most common finding in evaluation of canon VI was
an eye-fissure length of greater than the interocular distance. The differences
between these subgroups were a few millimeters. Naso-oral proportion (canon
VII) was present in very few subjects. In most subjects, 1.5 times the alar
width was greater than the mouth width. Evaluation of nasofacial proportion
(canon VIII) showed that the alar width was wider than one quarter of the
facial width in most subjects (Figure 7).
Finally, the nasal bridge inclination (canon IX) was greater than the ear
inclination in almost all subjects (Figure
A variety of classification schemes were assigned to each subject to
determine whether a classification scheme could be used to subcategorize subjects.
The Ofodile classification scheme5 was applied
to our subjects, who were assigned to 1 of the following 3 categories: African,
Afro-Caucasian, and Afro-Indian. The anthropometric means of each group were
compared using 1-way analysis of variance. Three of the 16 measures were statistically
different between the African and Afro-Caucasian subgroups (P<.05). In addition, the African and Afro-Indian subgroups varied
based on 3 of the 16 measures (P<.05). However,
no anthropometric measures were different between Afro-Caucasian and Afro-Indian
Subjects were also subjectively divided according to overall facial
shape. Analysis revealed 2 measures, interocular distance and nasal bridge
inclination, to be the only differences when comparing facial shape. Analysis
of the nostril shape based on the classification scheme proposed by Farkas
et al12 showed types III, IV, and V to be the
most common, found in 28.7%, 25.0%, and 25.0% of subjects, respectively. Comparison
of nostril type V with types II and III revealed nasal width to be different
(P<.05). Predictably, only nasal width was significantly
different when subjects were separated on the basis of the breadth of the
nasal tip (P = .001).
After subjective analysis of the photographs, we developed a classification
scheme based on assessment of the nasal starting point, subjective analysis
of dorsal height, and the shape of the dorsal profile (convex, concave, or
straight). A high dorsum is defined as one with a nasal starting point at,
or cephalad to, the level of the endocanthion with a convex or straight dorsum
that projects from the face. A low dorsum has a nasal starting point below
the level of the endocanthion, with a dorsum that is straight or concave,
and a low dorsum has very little projection from the face. Analysis of the
difference between these 2 groups based on the anthropometric measures obtained
revealed the following 5 measures that were significantly different: special
face height, nose length, forehead height I, special upper face height, and
nasal bridge inclination (Table 3).
Analysis of the face is a preliminary and important step in the approach
to the patient who presents for facial plastic and reconstructive surgery.
The extensive variability in the human face ensures its individuality. Although
the compilation of facial features of African American women is diverse, our
study shows that the average young African American female face differs significantly
from the North American Caucasian female face. Thirteen of the 16 measures
were significantly different from those of the North American Caucasian face.
Although the probability of type I error is elevated with repeated measures,
the large number of significantly different proportions with an α of
.003 suggests that significant differences between the populations exist.
Although the special head height was shorter for African American women, the
forehead heights I and II were longer and the height of the calva was shorter.
Nose length was shorter for African American women, as was ear length. Alar
width and eye-fissure width were greater in African American women; however,
facial width was not significantly different between groups. The mouth width
was greater and ear length was shorter. Although some of these measures have
a difference of only 2 mm, the overall composition of the values yields a
distinctly different appearance.
The neoclassical canons were originally formulated by scholars and artists
of the Renaissance and were based on classical Greek canons to define the
relationships between various areas of the head and face as a guide for artists.13-14 The influence of the neoclassical
canons, which dominated in the 17th and 18th centuries, diminished by the
late 19th century. They remain as the foundation on which modern facial analysis
is based.15 Neoclassical canons I through IV
relate to vertically oriented proportions, whereas canons V through VIII deal
with horizontal measures. Canon IX relates proportions of the angles of inclination.
Comparison of our data with the neoclassical canons reveals that very few
African American subjects fit the established proportions. Likewise, young
North American Caucasian women rarely fit them.2
The proportions of our subjects are illustrated in Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, and Figure 8. The illustrations are drawn to scale and are based on
the proportions of the average African American woman.
Facial analysis of the African American woman has evolved through several
studies. Ofodile and Bokhari6 conducted an
exploratory study that included physical examination, photographs, and anthropometric
measurements of the African American nose in 80 men and 121 women ranging
in age from 18 to 87 years. They observed that the African American nose could
be divided into 3 distinct groups denoted African, Afro-Caucasian, and Afro-Indian.
This classification was designed to bring order to the wide variations of
morphologic features and anatomy seen in the African American nose. In that
study, 53% of the African noses had a concave dorsum compared with 10% of
the Afro-Caucasian and 8% of the Afro-Indian groups. The most common nostril
shapes, based on the Farkas classification, were types IV (20%), V (27%),
and VI (25%). Anthropometric measurements showed variations according to nasal
type, with African noses being the shortest and widest; Afro-Caucasian, the
narrowest; and Afro-Indian, the longest. A shortcoming of that study was failure
to limit the age range of the subjects. Advanced age causes significant changes
in the appearance of the nose, including nasal elongation, tip ptosis, and
loss of tip support.16-17 Perhaps
the Afro-Indian nose represented an aged Afro-Caucasian subject.
Attempts at subcategorizing our subjects based on the Ofodile classification5 proved to be difficult, especially when differentiating
the Afro-Indian and Afro-Caucasian subtypes. A comparison of the African and
Afro-Caucasian subgroups and the African and Afro-Indian subgroups in our
study showed that 3 of the anthropometric measures were statistically different.
However, a comparison between the Afro-Indian and the Afro-Caucasian subgroups
generated no differentiation of the measures. Given the many similarities
of Afro-Indian and Afro-Caucasian groups, the potential to combine them exists.
Both groups appear to be similar to each other and different from the African
group. This finding provides impetus toward developing another means of classification.
In a later study, Ofodile and Bokhari6
measured 7 variables and calculated 6 area proportion indices related to the
nose. Comparisons were made with anthropometric measurements by Farkas1 for the North American adult population. Comparison
confirmed that the African American nose was wider than the white nose, with
a mean value of 40 mm for African American women compared with 34.7 mm for
white women. Our study supports this finding, with alar width being greater
in African American women compared with the North American white norm (P<.003).
As previously mentioned, photogrammetric analysis is less accurate than
anthropometric analysis.18 Nonetheless, Jeffries
et al10 photogrammetrically examined 200 African
American subjects (100 male and 100 female) aged 18 through 35 years. Computer
analysis of the photographs was performed, and the results were compared with
those of Farkas.1 They determined that African
American and white subjects had similar vertical facial proportions; however,
the horizontal proportions varied significantly. The African American nose
was shorter than the white nose. The horizontal dimensions (interocular distance,
nose width, mouth width, and facial width) showed many differences between
races, including 97% of the study group exhibiting a nose that was wider than
the interocular distance compared with 40.8% of white subjects who had a nose
that was the same size as the interocular distance. Jeffries and colleagues10 conclude that these findings are in agreement with
previously published data. However, they note that there are inherent problems
with measurements taken in this indirect manner. Our study found similar relationships.
Obviously, the African American population is diverse in appearance
and background. As Ofodile et al5 and Ofodile
and Bokhari6 have suggested, subpopulations
tend to emerge. Our subjects were assigned to subcategories based on facial
shape, nostril type, and width of the nasal tip. We could not differentiate
subjects based on division into these subgroups.
We propose a new classification scheme based on profile evaluation of
the nose: high vs low nasal dorsum. This categorization system is straightforward
and allows subjects to be separated easily into 2 groups. Comparison of the
anthropometric measures taken in these 2 groups yielded the following 5 measures
that were significantly different: special face height, nose length, forehead
height, special upper face height, and nasal bridge inclination. Further evaluation
is under way to determine if there are other characteristics that may more
reliably categorize the African American face.
When our results are compared with an anthropometric analysis of 50
young African American women,11 we find 11
of the 16 measures to be different. In addition, 14 of the 16 measures differ
on comparison of the Farkas1 African American
and North American Caucasian populations (P<.05).
Perhaps regional variation may account for these differences seen in our subjects
and the subjects examined by Farkas in the Northeast portion of the United
States. In addition, our study did not include persons of Caribbean descent.
However, similar relationships were found for both groups, ie, the African
American forehead height I for both populations was greater than the North
American white forehead height I.
Anthropometric analysis of the African American female face suggests
differences compared with the North American white face. Thus, a single aesthetic
ideal is inadequate. In addition, the African American woman does not fit
the neoclassical canons of facial proportion. It appears that the horizontal
measures have the most difference, in particular the nasal measure. The use
of previous classification schemes in our subject population did not reveal
significant differences. Basic groundwork has been laid for analyzing the
African American face. Use of the classification scheme high vs low nasal
dorsum may help to subcategorize the African American face into subgroups.
Additional studies are needed to further define a method of analysis.
Accepted for publication March 28, 2001.
This study was supported by a grant from the American Academy of Facial
Plastic and Reconstructive Surgery Foundation, Alexandria, Va.
Presented at the Fall Meeting of the American Academy of Facial Plastic
and Reconstructive Surgery, Washington, DC, September 21, 2000.
Corresponding author and reprints: Jennifer Parker Porter, MD, Bobby
R. Alford Department of Otorhinolaryngology and Communicative Sciences, Baylor
College of Medicine, One Baylor Plaza, SM 1727, Houston, TX 77030.