Content uploaded by Richard Russell
Author content
All content in this area was uploaded by Richard Russell
Content may be subject to copyright.
1 Introduction
Though a widespread belief in the social sciences holds that standards of beauty are
arbitrary cultural conventions (Etcoff 1999), recent work has shown that our preferences
are at least partly based in biology, with evidence mounting for several biologically
based factors of attractiveness (Fink and Penton-Voak 2002; Rhodes 2006). Yet groom-
ing behaviors such as the application of facial cosmetics have largely not been linked
to these factors, and are still considered products of fashion (but see Ca
¨rdenas and
Harris 2006; Fink and Neave 2005; Law Smith et al 2006 for suggestions of possible
biological underpinnings). Use of cosmetics to beautify the face long predates the iden-
tification of these factors of attractiveness, which indicates that cosmetics were not
developed explicitly to manipulate these factors. However, it is possible that cosmetic
usage has evolved over time to exploit one or more of these factors. The study reported
here supports this perspective, by showing that cosmetics manipulate one of the
biological factors of beauty
ö
sexual dimorphism. Specifically, it is demonstrated that
cosmetics exaggerate a previously unknown sex difference in facial contrast.
Female skin has been observed and measured to be lighter than male skin, provided
the males and females being compared are from the same ethnic group (Darwin 1871;
Edwards and Duntley 1939; Frost 1988, 2005; Jablonski and Chaplin 2000). This sex
difference is found in ethnic groups on all continents. That women have lighter skin
than men is well known to people living in ethnically homogenous settings, for whom this
sex difference is the major source of pigmentation variation. For the most readers
this will be a surprise, as the greatest source of pigmentation variation in the modern
world is between racial groups. The studies that have established this sex difference
have measured body parts not exposed to the sun, to control for the possibility of sex
differences in sun exposure and tanning. Because of this, the pigmentation of the face
has not been measured in detail, and it is not known whether females are lighter across
the entire face, including features such as the eyes and lips.
A sex difference in facial contrast and its exaggeration
by cosmetics
Perception, 20 09, volume 38, pages 1211 ^ 1219
Richard Russellô
Department of Psychology, Harvard University, 33 Kirkland Street, Cambridge, MA 02138, USA;
e-mail: rrussell@gettysburg.edu
Received 15 November 2008, in revised form 6 February 2009; published online 6 August 2009
Abstract. This study demonstrates the existence of a sex difference in facial contrast. By measuring
carefully controlled photographic images, female faces were shown to have greater luminance
contrast between the eyes, lips, and the surrounding skin than did male faces. This sex difference
in facial contrast was found to influence the perception of facial gender. An androgynous face
can be made to appear female by increasing the facial contrast, or to appear male by decreasing
the facial contrast. Application of cosmetics was found to consistently increase facial contrast.
Female faces wearing cosmetics had greater facial contrast than the same faces not wearing
cosmetics. Female facial beauty is known to be closely linked to sex differences, with femininity
considered attractive. These results suggest that cosmetics may function in part by exaggerating
a sexually dimorphic attribute
ö
facial contrast
ö
to make the face appear more feminine and hence
attractive.
doi:10.1068/p6331
ô Contact address (as of 17 August 2009): Department of Psychology, Gettysburg College, Gettysburg,
PA 17325, USA; e-mail: rrussell@gettysburg.edu
In faces of both sexes, the eyes and lips are darker than the surrounding skin,
forming an inverted triangle of dark regions surrounded by lighter regions that is
characteristic of faces (Sinha 2002; Watt 1994). Curiously, increasing or decreasing the
magnitude of this facial contrast has opposite effects on the attractiveness of male
and female faces (Russell 2003). In that study, female faces were rated more attractive
when the eyes and lips were darkened than when they were lightened, while male faces
were rated more attractive with those features lightened than with them darkened.
This is interesting considering the relation between sexual dimorphism and attractiveness.
It has been shown by a variety of methods that sex differences in facial appearance
influence attractiveness judgments. The relation is particularly strong in female faces,
with more feminine females considered more attractive (eg Bruce et al 1994; Cunningham
1986; Jones and Hill 1993; O'Toole et al 1998; Perrett et al 1998; reviewed by Rhodes 2006).
This background motivated the hypothesis that there is a sex difference in facial
contrast, with female faces having greater contrast than male faces. Further, if this sex
difference exists, people may consider faces with greater contrast as more feminine,
even though they are not aware of the sex difference. Experiment 1 was designed
to determine whether facial contrast is sexually dimorphic, and experiment 2 was
designed to see whether facial contrast plays a role in the perception of facial gender.
Because typical application of cosmetics involves darkening the eyes and lips, experi-
ment 3 was designed to determine whether female faces have greater facial contrast
with cosmetics than without.
2 Experiment 1
2.1 Method
To test the hypothesis that there is a sex difference in facial contrast (the luminance
difference between the eyes, lips, and the surrounding facial skin), photographs were
taken of 118 clean-shaven and cosmetics-free MIT students. This included 51 East Asians
(25 female, aged 18^ 22 years, mean age 19.4 years; 26 male, aged 16 ^ 23 years, mean
age 19.8 years) and 67 Caucasians (31 female, aged 18 ^ 27 years, mean age 20.5 years;
36 male, aged 17^ 28 years, mean age 20.7 years). To avoid systematic differences in
illumination, all photographs were taken under standard lighting conditions that have
been described in two other studies that used a subset of the images collected for the
present study (Russell et al 2006, 2007). The heights of the camera and lamps were
fixed, and the chair used by the photographic subjects was adjusted such that their
heads were all at the same height. The faces were illuminated by two studio lamps
with large diffusing heads, in a small room with white walls (for greater ambient
illumination). These two lamps, the camera and tripod, and the chair on which the
photographic subjects sat, were kept in locations that were fixed with respect to each
other and the room. The lights were centered at 08elevation (level with the head),
to eliminate cast shadows and to minimize variation from shading (Liu et al 1999).
As demonstrated in figure 1, gray-scale versions of each image were individually
hand-labeled to define regions corresponding to the eyes (including the skin between
the epicanthal fold and the eye, and the skin immediately below the eye), the lips,
annuli surrounding the eyes (with the approximate width of the eyes but not including
the eyebrow), and an annulus surrounding the lips (with the approximation width
of the mouth). Luminance values of all pixels within the eyes were averaged, as were
all the pixels in the lips, the annuli surrounding the eyes, and the annulus surround-
ing the lips, yielding mean luminance values for each of the four regions (eyes, eye annuli,
lips, lip annulus). The mean luminance values for the eyes and lips were averaged to
produce the mean feature luminance. Similarly, the mean luminance values for the eye
annuli and lip annulus were averaged to produce the mean skin luminance. Skin and
feature luminance, both being the averages of 8-bit pixel values, could range from 0
1212 R Russell
(black) to 255 (white). Facial contrast was calculated as C
F
(feature luminance ÿskin
luminance)=(feature luminance skin luminance).This is an adapted version of Michelson
contrast, which varies from 0 to 1, with higher values indicating greater contrast, and 0
indicating no contrast.
2.2 Results
Female faces had greater facial contrast than male faces in both the East Asian
(mean facial contrast: females 0.160; males 0:150) and the Caucasian (mean facial
contrast: females 0.123; males 0.113) samples, and the East Asian faces (with dark
eyes) had greater facial contrast than the Caucasian faces (with lighter eyes). A 2 (sex)
62 (race) analysis of variance (
ANOVA
) of facial contrast found significant main
effects of sex (F
1 114
9:69,p0:002,Z
2
p
0:08) and race (F
1 114
134:25,p50:001,
Z
2
p
0:54). The interaction between sex and race was not significant (F
1 114
0:00,
p0:997,Z
2
p
0:00), indicating that the sex difference in facial contrast did not differ
between East Asian and Caucasian faces. Supporting this notion, the effect sizes for
the sex difference in facial contrast were very similar for East Asian (d0:55) and
Caucasian (d0:60 ) faces.
Figure 2 shows skin luminance plotted against eye and lip (feature) luminance for
each face. The sex difference in contrast can be appreciated by noting that the regres-
sion line for female faces lies further from the line of equal luminance (along which
the skin and features are equally dark) than does the regression line for male faces.
, ,
,
190
150
110
Skin luminance
Asian females
Asian males
Caucasian females
Caucasian males
r0:88
r0:89
equal luminance
80 120 160
Feature luminance
Figure 2. Skin luminance plotted against
feature luminance. Larger values indi-
cate brighter regions. The regression
line for female faces lies further from
the line of equal luminance (faces with
features that are no darker or lighter
than the skin would lie along this line)
than does the regression line for male
faces, indicating that female faces have
greater facial contrast than male faces.
This sex difference does not differ bet-
ween East Asian and Caucasian faces.
Figure 1. Illustration of feature labelling. Solid lines
demonstrate how the boundaries of the eyes and lips were
defined. Dashed lines indicate how the boundaries of the
annuli surrounding those features were defined.
Sex difference in facial contrast 1213
Consistent with previous reports, female skin was lighter than male skin [mean skin
luminance: females 157 (SEM 1:5); males 147 (SEM 1:3)]. Though female eyes
and lips were lighter than male eyes and lips [mean feature luminance: females 119
(SEM 1:6); males 114 (SEM 1:4)], the difference was much smaller than the
difference between male skin and female skin. Thus, the sex difference in facial
contrast is a result of the sex difference in feature luminance being much smaller than
the sex difference in skin luminance.
To rule out the possibility that the sex difference in contrast is simply a function
of the sex difference in skin luminance, a 2 (sex)62 (race) analysis of covariance
(
ANCOVA
) was carried out with facial contrast as the dependent variable and skin
luminance as a control variable. Including skin luminance as a control variable did
not affect the pattern of results, with significant main effects of sex (F
1113
10:75,
p0:001,Z
2
p
0:09) and race (F
1 113
106:49,p50:001,Z
2
p
0:49), but no significant
interaction between sex and race (F
1 113
0:03,p0:854,Z
2
p
0:00). This indicates
that the sex difference in facial contrast exists even when overall skin lightness is controlled.
We can also consider the contrast around the eyes or mouth alone, with eye
contrast calculated as C
E
(eye luminance ÿeye skin luminance)=(eye luminance
eye skin luminance) and mouth contrast calculated as C
M
(mouth luminance ÿ
mouth skin luminance)=(mouth luminance mouth skin luminance). A 2 (sex)62(race)
ANOVA
of eye contrast showed significant main effects of sex (F
1114
6:71,
p0:011,Z
2
p
0:06) and race (F
1 114
155:51,p50:001,Z
2
p
0:58). Female faces
had greater eye contrast than male faces in both the East Asian (mean eye contrast:
females 0.206; males 0.189) and the Caucasian (mean eye contrast: females 0.139;
males 0.131) samples. The East Asian faces (with consistently dark eyes) had much
greater eye contrast than the Caucasian faces (with lighter eyes). The interaction between
sex and race was not significant (F
1114
0:84,p0:362,Z
2
p
0:01). A 2 (sex)62(race)
ANOVA
of mouth contrast showed a significant main effect of race (F
1 114
10:29,
p0:002,Z
2
p
0:08) and a trend toward a main effect of sex (F
1114
3:29,p0:072,
Z
2
p
0:03). Female faces had greater mouth contrast than male faces in the Caucasian
sample (mean mouth contrast: females 0.109; males 0.098) but there was virtually
no sex difference in the East Asian sample (mean mouth contrast: females 0.115;
males 0.114). However, the interaction between sex and race was not significant
(F
1 114
1:58,p0:211,Z
2
p
0:01). Of the two features, the sex difference in contrast
was larger for the eyes than the mouth, particularly for East Asian faces.
It is relevant that there is a sex difference not only in overall luminance of the
face, but also in luminance contrast, because contrast is a robust signal for visual
perception, and the property to which most neurons in the early visual stream respond.
Though luminance contrast is a robust visual cue, the effect size of the sex difference
in facial contrast (d0:55 for East Asians and d0:60 for Caucasians) is much
smaller than effect sizes for well-known sexual dimorphisms such as height (d1:4;
Ogden et al 2004) and waist-to-hip ratio (d1:7; Dobbelsteyn et al 2001). The smaller
effect size is likely the reason why people are not aware of the sex difference in facial
contrast, while they are aware of the sex differences in height and waist-to-hip ratio.
Though people are not conscious of the sex difference in facial contrast, they may
nevertheless use it as a cue in determining the sex of a face or making judgments of
facial masculinity or femininity.
3 Experiment 2
A demonstration of the utility of facial contrast for determining the sex of a face can
be seen in figure 3. Both images were created by manipulating the same original image
of a perceptually androgynous face made by morphing together male and female average
faces (see figure A1 in the appendix). To make both images (with Adobe Photoshop),
,
,
,
,
,
,
,
,
,
1214 R Russell
the eyes and lips were left unchanged, but the rest of the image was darkened to
produce the left image or lightened to produce the right image. Because the eyes and
lips were unchanged while the rest of the face was made darker or lighter, facial
contrast was decreased or increased. Though a subtle manipulation, it has a powerful
effect
ö
making the image on the left with decreased contrast appear male and the image
on the right with increased contrast appear female.
In the Illusion of Sex shown in figure 3, the face with lower contrast also has darker
skin. Because males have darker skin, it is possible that the difference in skin darkness
drives the illusion rather than the difference in facial contrast. To test this possibility,
we can create a pair of faces in which facial contrast is manipulated by darkening or
lightening the eyes and lips but keeping the rest of the face unchanged. These faces
appear in figure 4. Again, the face on the left with decreased contrast appears male
and the face on the right with increased contrast appears female. Because the two faces
have the same skin tone, this provides evidence that facial contrast alone can be used as
a cue for determining the sex of a face. Indeed, facial contrast may in fact be more
important than overall lightness
ö
darkening or lightening the entire face has no effect
on perceived gender (see figure A2 in the appendix). Collectively, these versions of the
Figure 3. The Illusion of Sex. The face on the left appears male, while the face on the right
appears female. Both images were produced by making slight alterations to the same original
image. The eyes and lips were unaltered, and hence equally dark in both images. The remainder
of the image was darkened to produce the left image, and lightened to produce the right image.
The eyes and lips may appear darker in the right image than in the left image, but are not
ö
it
is an example of simultaneous contrast.
Figure 4.Variant of the Illusion of Sex. The face on the left appears male, while the face on the right
appears female. Both images were produced by making slight alterations to the same original
image. The eyes and lips were lightened to produce the left image, and darkened to produce
the right image. The rest of the face was unaltered, and hence equally dark in both images.
That decreasing or increasing facial contrast is sufficient to make a face appear male or female
indicates that facial contrast plays a role in the perception of facial gender.
Sex difference in facial contrast 1215
Illusion of Sex indicate that, while people are not consciously aware of the sex difference
in facial contrast, they nevertheless use it as a cue in perceiving the sex of a face.
Though facial contrast is known to have different effects on male and female
attractiveness (Russell 2003), its relation to perceived masculinity and femininity has
not been explained. Because facial contrast is used as a cue for perceiving the sex of a
face, it would not be surprising if it were also used as a cue for making judgments
of masculinity and femininity. To test the hypothesis that facial contrast is related
to judgments of masculinity and femininity, twenty-nine subjects (fifteen female) gave
Likert-scale ratings of masculinity (for male faces) or femininity (for female faces) to
full-color (RGB) versions of 117 of the 118 images described above. The image excluded
from the set was an Asian female personally known by several subjects. Cronbach's a
indicated high reliability of the judgments of masculinity (a0:88) and femininity
(a0:95). Facial contrast was positively correlated with rated femininity of female
faces (r0:31,p0:022), but negatively correlated with rated masculinity of male faces
(rÿ0:46,p50:001).
These correlations between rated femininity and masculinity and facial contrast
could be a byproduct of skin luminance (which is also sexually dimorphic). With skin
luminance as a control variable, facial contrast was still positively correlated with
rated femininity of female faces (r0:36,p0:008) and negatively correlated
with rated masculinity of male faces (rÿ0:39,p0:002). Even with skin luminance
controlled, greater facial contrast was rated more feminine in female faces but less
masculine in male faces. Another possible confound could come from analyzing the
Caucasian and Asian faces together. To take both skin luminance and ethnicity into
account, the results were analyzed as a partial correlation with skin luminance and
ethnicity as control variables. With skin luminance and ethnicity controlled, facial
contrast was correlated positively with rated femininity of female faces (r0:24,
p0:078), but negatively with rated masculinity of male faces (rÿ0:05,p0:700).
Controlling for ethnicity and skin luminance yields the same pattern of results
(a positive correlation between contrast and femininity but a negative correlation
between contrast and masculinity), but the correlations were weaker and not statisti-
cally significant (particularly for male faces). To summarize, in a set of real faces, a
significant relationship was found between facial contrast and rated masculinity and
femininity, but it was modulated by ethnicity. Overall, the ratings were weakly consistent
with greater facial contrast being considered more feminine and less masculine.
4 Experiment 3
Typical cosmetics usage, by darkening the eyes and lips while little changing the lumi-
nance of the rest of the face, should increase facial contrast. To confirm that this
is the case, 12 Caucasian females (18 ^21 years, mean 19.6 years) were photographed.
These individuals were photographed twice, once with and once without cosmetics.
For the photographs with cosmetics, the women were instructed to ``apply cosmetics as
you would when going out at night'', and applied the cosmetics themselves at home.
Facial contrast was measured with gray-scale images of the faces by the same
methods as in experiment 1. Mean facial contrast was 0.16 (SD 0:04) in the faces
wearing cosmetics, and 0.12 (SD 0:01) in the faces without cosmetics, a significant
difference as assessed by a paired-samples, two-tailed t-test (t
12
5:6,p50:001,
d1:70). The increase of facial contrast was also consistent
ö
all 12 faces had greater
contrast with cosmetics than without. This large increase in facial contrast achieved with
cosmetics more clearly differentiates male and female faces. The effect size of the sex
difference in facial contrast comparing the 36 male Caucasian faces and the 12 female
Caucasian faces wearing cosmetics (d1:85) compares favorably to the effect sizes of
the sex differences in height or waist-to-hip ratio. Application of cosmetics increased both
1216 R Russell
eye contrast [with cosmetics, mean eye contrast 0.17 (SD 0:03); without cosmetics,
mean eye contrast 0:13 (SD 0:03); paired-samples, two-tailed t
12
3:2,p0:009,
d1:15] and mouth contrast [with cosmetics, mean mouth contrast 0.16 (SD 0:04);
without cosmetics, mean mouth contrast 0.10 (SD 0:02); paired-samples, two-tailed
t
12
5:1,p50:001,d1:70]. These results confirm that cosmetics are used in a way
that accentuates a sexually dimorphic feature (facial contrast).
5 Discussion
Investigation of a large set of faces showed that females have greater facial contrast
than do males. This sex difference in facial contrast was found in both East Asian and
Caucasian faces. Female or male faces with greater facial contrast were rated as more
feminine or less masculine than faces with less contrast, though the relation was very
weak for male faces. Decreasing or increasing the facial contrast in an androgynous
face (figures 3 and 4) is sufficient to make it appear male or female. These findings
indicate that, while people are not consciously aware of this sex difference, their per-
ceptual systems nevertheless make use of it. Because femininity and attractiveness are
strongly related (Bruce et al 1994; Cunningham 1986; Jones and Hill 1993; O'Toole
et al 1998; Perrett et al 1998; Rhodes 2006), these results help explain the previous
finding that female faces are more attractive with increased facial contrast than with
decreased facial contrast (Russell 2003).
In the current study, typical application of cosmetics was found to increase the
contrast between the eyes, lips, and the rest of the face
ö
precisely the manipulation
capable of making the face appear more feminine. It is extremely unlikely that this
would happen by chance. Parts of the face could be lightened or darkened in many
different spatial patterns, but only this particular pattern is related to how male and
female faces differ. Further, there is a direction to the spatial pattern
ö
increasing the
contrast makes the face appear more feminine, but decreasing it makes the face
appear more masculine. Yet cosmetics consistently increase facial contrast. Faces are
rated more feminine (Cox and Glick 1986) and more attractive (Cash et al 1989;
Cox and Glick 1986; Graham and Jouhar 1981; Huguet et al 2004; Mulhern et al
2003) when wearing cosmetics than when not wearing cosmetics, whether the cosmetics
are self-applied [as in the Cash et al (1989) study] or professionally applied (as in
the other studies). Together with the current findings, this suggests that an important
function of cosmetics may be to increase the apparent femininity, and hence attractive-
ness, of the female face by increasing facial contrast.
Accentuating sex differences to make the female face appear more feminine and
thereby more attractive is not limited to changing facial contrast. Another common
manipulation of a sexually dimorphic facial feature is eyebrow plucking. Brow thick-
ness and brow-to-eye distance are both sexually dimorphic (Burton et al 1993; Farkas
and Munro 1987), with females having thinner brows that are higher above the eye.
Standard advice (Aucoin 1997; Brown and Iverson 1997) instructs women to pluck the
eyebrows from the bottom side, resulting in a thinner brow that is also further from
the eye, making the face appear more feminine. It is likely that accentuation of sexual
dimorphism as a way to enhance facial attractiveness is a general strategy of cosmetics
and other grooming behaviors.
In addition to sexual dimorphism, there are other biologically based standards
of facial beauty, including averageness (Langlois and Roggman 1990), symmetry
(Thornhill and Gangestad 1993), and youth (Zebrowitz 1997). It is likely that cosmetics
are
ö
or could be
ö
used to manipulate all of these factors. While it is widely believed
that cosmetics are an arbitrary cultural phenomenon largely dictated by fashion, the
present findings suggest an alternative scientific explanation for the use of cosmetics,
premised upon their manipulation of biologically based factors of facial beauty.
Sex difference in facial contrast 1217
Acknowledgments. I thank Jill Jin and Charisse Massey for taking the photos, P Matt Bronstad,
Nancy Etcoff, and Ken Nakayama for comments, and Pawan Sinha for comments and support.
References
Aucoin K, 1997 Making Faces (Boston, MA: Little, Brown)
Brown B, Iverson A, 1997 Bobbi Brown Beauty (New York: HarperCollins)
Bruce V, Burton A M, Dench N, 1994 ``What's distinctive about a distinctive face?'' Quarterly
Journal of Experimental Psychology A 47 119 ^ 149
Burton A M, Bruce V, Dench N, 1993 ``What's the difference between men and women? Evidence
from facial measurement'' Perception 22 153 ^ 176
Ca
¨rdenas R A, Harris L J, 2006 ``Symmetrical decorations enhance the attractiveness of faces
and abstract designs'' Evolution and Human Behavior 27 1^18
Cash T F, Dawson K, Davis P, Bowen M, Galumbeck C, 1989 ``Effects of cosmetics use on the
physical attractiveness and body image of American college women'' Journal of Social Psychol-
ogy 129 349 ^ 355
Cox C L, Glick W H, 1986 ``Resume evaluations and cosmetics use: When more is not better''
Sex Roles 14 51 ^ 58
Cunningham M R, 1986 ``Measuring the physical in physical attractiveness: Quasi-experiments
on the sociobiology of female facial beauty'' Journal of Personality and Social Psychology 50
925^935
Darwin C, 1871 The Descent of Man, and Selection in Relation to Sex (London: John Murray)
Dobbelsteyn C, Joffres M, MacLean D, Flowerdew G, 2001 ``A comparative evaluation of waist
circumference, waist-to-hip ratio and body mass index as indicators of cardiovascular risk
factors. The Canadian Heart Health Surveys'' International Journal of Obesity 25 652 ^ 661
Edwards E A, Duntley S Q, 1939 ``The pigments and color of living human skin''American Journal
of Anatomy 65 1^33
Etcoff N, 1999 Survival of the Prettiest: the Science of Beauty (New York: Doubleday)
Farkas L G, Munro I R, 1987 Anthropometric Facial Proportions in Medicine (Springfield, IL:
Charles C Thomas)
Fink B, Neave N, 2005 ``The biology of facial beauty'' International Journal of Cosmetic Science
27 317 ^ 325
Fink B, Penton-Voak I, 2002 ``Evolutionary psychology of facial attractiveness'' Current Directions
in Psychological Science 11 154 ^ 158
Frost P, 1988 ``Human skin color: A possible relationship between its sexual dimorphism and its
social perception'' Perspectives in Biology and Medicine 32 38 ^ 58
Frost P, 2005 Fair Women, Dark Men (Christchurch: Cybereditions)
Graham J A, Jouhar A J,1981 ``The effects of cosmetics on person perception'' International Journal
of Cosmetic Science 319 9 ^ 210
Huguet P, Croizet J-C, Richetin J, 2004 ``Is ``What has been cared for'' necessarily good? Further
evidence for the negative impact of cosmetics use on impression formation'' Journal of Applied
Social Psychology 34 1752 ^ 1771
Jablonski N G, Chaplin G, 2000 ``The evolution of human skin coloration'' Journal of Human
Evolution 39 57 ^ 106
Jones D M, Hill K, 1993 ``Criteria of facial attractiveness in five populations'' Human Nature 4
271 ^ 296
Langlois J H, Roggman L A, 1990 ``Attractive faces are only average'' Psychological Science 1
115^121
Law Smith M J, Perrett D I, Jones B C, Cornwell R E, Moore F R, Feinberg D R, Boothroyd L G,
Durrani S J, Stirrat M R, Whiten S, Pitman R M, Hillier S G, 2006 ``Facial appearance is
a cue to oestrogen levels in women'' Proceedings of the Royal Society of London, Series B 273
135^140
Liu C H, Collin C A, Burton A M, Chaudhuri A, 1999 ``Lighting direction affects recognition
of untextured faces in photographic positive and negative'' Vision Research 39 4003 ^ 4009
Mulhern R, Fieldman G, Hussey T, Leveque J-L, Pineau P, 2003 ``Do cosmetics enhance female
Caucasian facial attractiveness?'' International Journal of Cosmetic Science 25 199 ^ 205
O'Toole A J, Deffenbacher K A, Valentin D, McKee K, Huff D, Abdi H, 1998 ``The perception
of face gender: The role of stimulus structure in recognition and classification'' Memory
& Cognition 26 146 ^ 160
Ogden C, Fryar C, Carrol M, Flegal K, 2004 Mean Body Weight, Height, and Body Mass Index,
United States 1960 ^ 2 002 Advance data from vital and health statistics, no 347 (Hyattsville, MD:
National Center for Health Statistics)
1218 R Russell
Perrett D I, Lee K J, Penton-Voak I, Rowland D, Yoshikawa S, Burt D M, Henzi S P, Castles D L,
Akamatsu S, 1998 ``Effects of sexual dimorphism on facial attractiveness'' Nature 394 884 ^ 887
Rhodes G, 2006 ``The evolutionary psychology of facial beauty'' Annual Review of Psychology 57
199 ^ 226
Russell R, 2003 ``Sex, beauty, and the relative luminance of facial features''Perce ptio n 32 1093 ^ 1107
Russell R, Biederman I, Nederhouser M, Sinha P, 2007 ``The utility of surface reflectance for
the recognition of upright and inverted faces'' Vision Research 47 157 ^ 165
Russell R, Sinha P, Biederman I, Nederhouser M, 2006 ``Is pigmentation important for face
recognition? Evidence from contrast negation'' Perception 35 749 ^ 759
Sinha P, 2002 ``Qualitative representations for recognition'' Lecture Notes in Computer Science
(LNCS) 2525 249 ^ 262
Thornhill R, Gangestad S W, 1993 ``Human facial beauty: averageness, symmetry, and parasite
resistance'' Human Nature 4237 ^ 269
Watt R J, 1994 ``A computational examination of image segmentation and the initial stages of
human vision'' Perception 23 383 ^ 398
Zebrowitz L A, 1997 Reading Faces: Window to the Soul? (Boulder, CO: Westview Press)
Appendix
Figure A1. This androgynous face was produced by morphing together Caucasian male and female
averaged faces. It was manipulated to produce the faces in figures 3, 4, and A2.
Figure A2. No illusion. Both the right and left images appear androgynous. The entire original
image (figure A1) was lightened to produce the left image, and darkened to produce the right image.
ß 20 09 a Pion publication
Sex difference in facial contrast 1219
ISSN 0301-0 06 6 (print)
Conditions of use. This article may be downloaded from the Perception website for personal research
by members of subscribing organisations. Authors are entitled to distribute their own article (in printed
form or by e-mail) to up to 50 people. This PDF may not be placed on any website (or other online
distribution system) without permission of the publisher.
www.perceptionweb.com
ISSN 1468-423 3 (electronic)
