Original Article
Cues to fertility: perceived attractiveness and facial shape predict
reproductive success
Lena S. Pflügera,⁎, Elisabeth Oberzauchera
, Stanislav Katinab,c
,
Iris J. Holzleitnera
, Karl Grammera
a
Department of Anthropology, University of Vienna, Althanstrasse 14, A-1090, Vienna, Austria
b
School of Mathematics and Statistics, University of Glasgow
c
Department of Applied Mathematics and Statistics, Faculty of Mathematics, Physics and Informatics, Comenius University, Bratislava
Initial receipt 19 April 2011; final revision received 23 May 2012
Abstract
Attractive facial features in women are assumed to signal fertility, but whether facial attractiveness predicts reproductive success in
women is still a matter of debate. We investigated the association between facial attractiveness at young adulthood and reproductive life
history—number of children and pregnancies—in women of a rural community. For the analysis of reproductive success, we divided the
sample into women who used contraceptives and women who did not. Introducing two-dimensional geometric morphometric methodology,
we analysed which specific characteristics in facial shape drive the assessment of attractiveness and covary with lifetime reproductive
success. A set of 93 (semi)landmarks was digitized as two-dimensional coordinates in postmenopausal faces. We calculated the degree of
fluctuating asymmetry and regressed facial shape on facial attractiveness at youth and reproductive success. Among women who never used
hormonal contraceptives, we found attractive women to have more biological offspring than less attractive women. These findings are not
affected by sociodemographic variables. Postmenopausal faces corresponding to high reproductive success show more feminine features—
facial characteristics previously assumed to be honest cues to fertility. Our findings support the notion that facial attractiveness at the age of
mate choice predicts reproductive success and that facial attractiveness is based on facial characteristics, which seem to remain stable until
postmenopausal age.
© 2012 Elsevier Inc. All rights reserved.
Keywords: Facial attractiveness; Facial shape; Reproductive success; Fluctuating asymmetry; Geometric morphometrics
1. Introduction
Female faces are assumed to display honest cues to
fertility by means of facial attractiveness (Grammer, Fink,
Møller, & Thornhill, 2003). Previous studies support the
notion that the perception of women's facial attractiveness is
based on morphological indicators of physiological and
developmental condition. For instance, small random deviations
from perfect bilateral symmetry (Van Valen, 1962),
so-called fluctuating asymmetries (FAs), are known to
negatively affect ratings on facial attractiveness (Grammer
& Thornhill, 1994; Rhodes, Proffitt, Grady, & Sumich,
1998; Mealey, Bridgstock, & Townsend, 1999). As these
small morphological deviations develop early during
embryogenesis as an outcome of genomic and environmental
stress, a low level of FA has been linked to developmental
stability (Parsons, 1992; Gangestad, Thornhill, & Yeo, 1994;
Thornhill & Møller, 1997) and health (see meta-analysis of
Van Dongen & Gangestad, 2011). Other anthropometric
indicators are so-called hormone markers. In humans, the
basic facial proportions are affected by prenatal sex
steroides. For instance, the influence of oestrogene leads to
‘typically female’ features (e.g., less robust jaws, high
eybrows, and fuller lips), which are known to positively
affect ratings on female attractiveness and are supposed to
serve as reliable indicators of fecundity (Johnston &
Franklin, 1993; Fink et al., 2005).
Even though the impact of facial attractiveness on
human mate choice is well documented and theoretically
discussed, no distinctive facial characteristics, which
Evolution and Human Behavior 33 (2012) 708–714
⁎ Corresponding author.
E-mail address: lena.pflueger@univie.ac.at (L.S. Pflüger).
1090-5138/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.evolhumbehav.2012.05.005
covary with female reproductive success, have been
identified so far. Three previous studies investigated the
relation between perceived facial attractiveness at the age
of young adulthood and final number of children. Hill and
Hurtado (1996) showed that facial attractiveness correlates
with the number of children in a hunter and gatherer
society. Pawlowski, Boothroyd, Perrett, and Kluska (2008)
and Jokela (2009) investigated the association between
facial attractiveness at youth and postmenopausal number
of children in samples of contemporary women. Jokela's
findings indicate that, for women living in industrialized
settings, physical attractiveness is related to the number
of children, while Pawlowski et al. (2008) failed to show
any association.
The aim of the present study was to shed light on the
signalling value of female facial attractiveness in terms of
reproductive success. In contrast to previous research, we
applied a new approach of landmark-based geometric
morphometric methodology (GMM) in order to analyse
female facial shape corresponding to reproductive success.
We hypothesized that if perceived attractiveness at the age of
young adulthood predicts female reproductive success, faces
must possess anthropometric measurable indicators, which
have driven the perception at young age.
2. Materials and methods
2.1. Data collection
Questionnaire data and facial photographs were collected
from 88 women of a rural community of 1868 inhabitants in
Carinthia (Austria). All women were at postmenopausal age
(mean±S.D.=61.04±8.70 years) and in a long-term relationship
with their current partner (mean age at the start of the
partnership±S.D.=20.07±4.01 years). Subjects were interviewed
about their total number of children. Besides the
biological offspring, reproductive success was measured by
estimating number of pregnancies, taking (induced) abortions
into account.
In order to control for potentially confounding variables,
we collected the following information: age of
participants, years of marriage to their husband (father of
their children), education of subject and corresponding
husband, income of subject and corresponding husband
before pension, family contact of subject while raising
children, and religiosity of subject. Additionally, women
were asked about the use of hormonal contraceptives (HC)
during their fertile years.
We collected facial portraits showing participants at
young adulthood (aged 19–23 years). Images were mostly
wedding pictures taken by the same photographer. Additionally,
we took standardized frontal pictures (constant in
lightning condition, camera-to-subject distance, and focal
length) of women's present-day faces while participants
showed neutral expressions (i.e., having mandible relaxed
with teeth not touching and lips closed freely).
2.1.1. Sample subdivision
For the analysis of facial shape, portraits which were not
exactly frontally imaged (N=10) were excluded (e.g., when
heads were rotated vertically or horizontally). This resulted
in a total of 78 women used for the assessment of the relation
between FA and facial attractiveness. Two women had to be
excluded in the analyses considering reproductive success
due to missing data (husbands did not fill out the
questionnaire). For the analysis considering facial attractiveness
and reproductive success, the sample was divided
into women who used HCs (N=46) and women who never
used HCs (N=40) during their fertile years.
2.2. Attractiveness rating
Facial pictures showing participants at young adulthood
were aligned in colour, size, tonal value, and contrast. Hair
and accessories were masked with an oval frame. One
hundred twenty-five male students (mean age±S.D.=23.47±
3.90 years) of the University of Vienna evaluated the female
faces on the items ‘attractive’ and ‘sexy’ with sliders ranging
from 1 (not attractive/sexy) to 100 (highly attractive/sexy).
From past experience, we expected that the evaluation of
88 stimuli easily overtaxes participants and leads to
exhaustion. In order to prevent such bias, each male
participant was requested to evaluate a subset of 20
randomly chosen faces. The rating study resulted in at least
20 ratings per woman.
2.3. Statistical analyses
Statistical analyses were performed in R (R Development
Core Team, 2010). All univariate two-variable associations
were estimated by Spearman correlation coefficient rs, and
the test of zero correlation (9999 permutations) was used to
test this association. We controlled for the confounding
effect of age, years of marriage, income of husband, and own
income while estimating the association between facial
attractiveness and number of children by Spearman partial
correlation coefficient rs. The test of zero partial correlation
(9999 permutations) was used to test this association. The
mean levels of facial attractiveness (attractiveness/sexy)
between women who used HCs and women who did not
were compared by the two-sample Wilcoxon test (9999
permutations). The effect of family contact, education of
subject, education of husband, and religiosity (church
appendance) on number of children was tested by the
Kruskal–Wallis test (9999 permutations).
2.4. Landmark-based GMM
Landmark-based GMM deals with the analysis of
geometrical information about the form (also called
geometric morphometrics; Bookstein, 1991). Digitizing
(semi)landmarks on facial representations [two-dimensional
(2D) portraits] provides individual geometric information of
each specimen based on (semi)landmark coordinates. The
main idea of GMM analysis is to adjust Cartesian
709L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714
coordinates of (semi)landmarks with respect to position,
orientation (rotation), and scale (size) using Procrustes
superimposition [Generalized Procrustes Analysis (GPA),
Bookstein, 1991]. The resulting centred Procrustes shape
coordinates (centred PSCs; Procrustes tangent coordinates)
capture shape information only and can be used for
subsequent multivariate statistical analyses. One major
advantage of using GMM is that statistical results (e.g.,
multivariate linear regression) emerge in terms of estimated
(semi)landmark coordinates and findings can be visualized
(Schaefer et al., 2006).
2.4.1. Landmark setup and Procrustes superimposition
The shape of women's present-day faces was determined
by digitizing 2D coordinates of 35 anatomical landmarks,
which were based on biologically homologous facial traits.
Fifty-eight semilandmarks (slidings; Gunz, Mitteroecker, &
Bookstein, 2005) on curves (eyebrows, pupils, lips, and
chin) were added to improve the shape description of the
faces (see supplementary material, available on the journal's
website at www.ehbonline.org). With respect to a reference
face, the semilandmarks on the target faces were iteratively
slid along the particular curves until bending energy is
minimized (Slice, 2005). (Semi)landmarks were digitized by
one person only. In order to assess the intraobserver error, a
randomly selected face was digitized repeatedly 10 times
(with at least 24-h time lapse). This resulted in a median
intraobserver error of 1.44% per (semi)landmark with
respect to the whole sample median variance.
In order to transform Cartesian coordinates into shape
coordinates, a GPA was performed. First, the centroid of
each form was computed, and all (semi)landmark configurations
were translated to the same origin. Next, landmark
configurations were scaled to the same size (centroid size)
and finally rotated around the origin in order to minimize the
sum of squared distances between corresponding (semi)
landmarks (Rohlf & Slice, 1990). This resulted in PSCs, i.e.,
coordinates that are invariant to change in location, rotation,
and scale (Slice, Bookstein, Marcus, & Rohlf, 1996).
2.4.2. Measuring FA
To analyse facial FA, we quantified the dissimilarities in
shape (deviations of PSCs) between the original configurations
and their relabelled reflections (RRs; Mardia, Bookstein,
& Moreton, 2000).
The original (semi)landmarks and their RR counterparts
were jointly submitted to GPA in order to register both in the
same shape space (Bookstein, 1991). Both configurations
(original and RR) were centred to their Procrustes mean
resulting in original and RR centred PSCs, respectively. FA
expresses the extent to which the sample fluctuates around its
own mean asymmetry (Mardia et al., 2000). Thus, the
individual degree of FA was assessed as the total sum of
squared distances (Procrustes shape distances) between
centred PSCs of original and RR configurations (Bock &
Bowman, 2006).
2.4.3. Shape regression
We determined the linear association between facial
shape and three independent variables (attractiveness scores
at youth, number of children, and number of pregnancies) for
women who never used HCs only (not exactly frontally
imaged facial portraits excluded, N=3). Based on a
multivariate ordinary linear regression model (shape regression),
we calculated a linear regression function for each
shape coordinate separately. Estimated R2
describes the
quality of linear relationship between shape and independent
variables [mean R2
=mean relation of all (semi)landmarks;
min R2
=R2
of single (semi)landmark with the weakest
relation to the independent variable; max R2
=R2
of single
(semi)landmark with the strongest relation to the independent
variable]. Results of shape regression were visualized
via thin-plate spline (TPS) deformation grids (Bookstein,
1991). Estimated slopes correspond to the shape changes of a
particular (semi)landmark (displacement of its positions) that
occur in one unit change of the independent variable
(Schaefer et al., 2006). In this study, TPS deformation
grids illustrate the changes in shape from the Procrustes
mean shape to the estimated shape corresponding to the most
negative and the most positive change of the independent
variables. To ease the visualization of a shape change, TPS
grids were extrapolated 1.2 times in the negative and positive
directions of the independent variable.
3. Results
3.1. Perceived facial attractiveness and
reproductive success
Table 1 presents the association between perceived facial
attractiveness (sexy and attractive) and reproductive success
(number of children and pregnancies) of women who never
used HC compared to women who used HC during their
fertile years. Results are presented with and without
adjustments for sociodemographic variables. We observed
no relation between facial attractiveness and the use of HCs
(N=88; attractive: p=.256; sexy: p=.278).
3.1.1. Women who never used HC (N=40)
Women who never used HC showed a positive association
between facial attractiveness at youth and reproductive
success. This association remained equally significant with
adjustment for age, years of marriage, and income (Table 1).
Education of subject (p=.945) and corresponding husband
(p=.867), family contact of subject (p=.562), and religiosity
of subject (p=.360) were not associated with the number of
children. For descriptive statistics of the tested sociodemographic
variables, see Tables 2 and 3.
As illustrated in Fig. 1, the lowest number and the highest
number of children and pregnancies are represented by few
data points. In order to test the stability of this finding, we
performed the same analysis after excluding extreme values
(women with zero, five, or six children/women with zero or
710 L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714
six to eight pregnancies, N=5). The positive correlation with
number of children remained but was no longer significant
(Fig. 1, left panel, regression line adjusted), whereas the
significant positive association with number of pregnancies
persisted (Fig. 1, right panel, regression line adjusted).
3.1.2. Women who used HC (N=46)
In women who used contraceptives during their fertile
years, we found no association between perceived facial
attractiveness and reproductive success. This lack of an
association remained after adjustment for age, years of
marriage, and income (Table 1). Education of subject
(p=.081) and corresponding husband (p=.517), family
contact of subject (p=.058), and religiosity of subject
(p=.359) were also not associated with the number of
children born to women in this group.
3.2. FA and reproductive success
We found a negative correlation between the degree of
postmenopausal FA and facial attractiveness at youth
(attractive: rs=−0.205, N=78, p=.034; sexy: rs=−0.222,
N=78, p=.03). The degree of facial FA was not related to
subjects` present age [rs=−0.108, N=78, not significant
(NS)]. We observed no association between the degree of
postmenopausal FA and reproductive success either in
women who used HC (children: rs=−0.10 NS; pregnancies:
rs=−0.02 NS) during their fertile years or in women who
never used HC (children: rs=0.15 NS; pregnancies: rs=0.13
NS). The negative association between facial attractiveness
and FA in the subsamples did not persist. In both groups, FA
was independent of subject's age.
3.3. Shape regression
As women with extremely high and low reproductive
success showed to have a strong effect on the association
between attractiveness and reproductive success, we excluded
women with no children/pregnancies and women with
more than four children or more than five pregnancies (N=4,
plus three not exactly frontally imaged facial portraits, see
Section 2.4.3). This resulted in a sample size of 33 women.
The shape regression revealed covariations between the
shape of postmenopausal faces and all three independent
variables. The TPS deformations grids in Figure 2 illustrate
facial shape associated with facial attractiveness at youth
(upper row), number of children (middle row), and number
of pregnancies (lower row). The middle column represents
the consensus (Procrustes mean shape).
Faces corresponding to a decreasing level of attractiveness,
number of children, and number of pregnancies (left
column) have thinner lips, flatter noses, broader eyebrows,
and more angular jaws than the consensus. Faces
corresponding to an increasing level of attractiveness,
number of children, and number of pregnancies (right
column) have fuller lips, small and more tapered noses,
higher arched eyebrows, and less angular jaws compared to
the consensus.
Table 1
Correlations of perceived facial attractiveness and reproductive success with and without adjustments for sociodemographic variables (SDVs)
Adjustments
for SDVs
No contraceptives (N=40) Use of contraceptives (N=46)
No. of children No. of pregnancies No. of children No. of pregnancies
Attractive Not adjusted rs=0.374* rs=0.514** rs=−0.081 NS rs=−0.079 NS
Age rs=0.364* rs=0.505*** rs=−0.082 NS rs=−0.080 NS
Years of marriage rs=0.372** rs=0.509*** rs=−0.078 NS rs=−0.076 NS
Income rs=0.374** (s) rs=0.516*** (s) rs=−0.058 NS (s) rs=−0.067 NS (s)
rs=0.379* (h) rs=0.517*** (h) rs=−0.043 NS (h) rs=−0.040 NS (h)
Sexy Not adjusted rs=0.425** rs=0.568** rs=−0.102 NS rs=−0.076 NS
Age rs=0.414** rs=0.553*** rs=−0.103 NS rs=−0.079 NS
Years of marriage rs=0.423** rs=0.563*** rs=−0.092 NS rs=−0.068 NS
Income rs=0.424** (s) rs=0.567*** (s) rs=−0.079 NS (s) rs=−0.065 NS (s)
rs=0.440** (h) rs=0.575*** (h) rs=−0.060 NS (h) rs=−0.031 NS (h)
NS: not significant; *.01≤pb.05; **.001≤pb.01; ***pb.001. Association of perceived facial attractiveness and reproductive success (test of zero correlation,
9999 permutations, Spearman correlation coefficient) in women who did not use contraceptives and women who used contraceptives. Values are presented with
and without adjustments for age, years of marriage, and income (test of zero partial correlation, 9999 permutations, Spearman partial correlation coefficient).
(s)=subject, (h)=husband.
Table 2
Descriptive statistics for sociodemographic variables (age, years of marriage, and income)
Description No contraceptives (N=40) Use of contraceptives (N=46)
Age (in years) Age at data collection (subject) 65.60 (9.15) 57.87 (6.04)
Years of marriage Years between marriage and data collection 43.15 (9.09) 35.28 (6.51)
Income (after tax per month in Euro) Subject 362.47 (600.34) 592.78 (672.68)
Husband 1214.45 (491.73) 1788.98 (769.25)
Sociodemographic variables collected in continuous scale. Values are: means (S.D.).
711L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714
4. Discussion
In the first part of our research, contemporary male
students evaluated facial attractiveness of female participants
based on pictures taken at the age of young adulthood. In
order to determine the relation of facial attractiveness and
reproductive success, we separated women who used HCs
from women who did not. For the latter, facial attractiveness
was positively related to the number of children and
pregnancies (Table 1), supporting our hypothesis that
physical attractiveness predicts female reproductive success.
While this effect was strongly driven by women with very
high or low reproductive success, it remained significant
after exclusion of women with no or more than five
pregnancies. For women who utilized birth control via
HCs, the association was not present (Table 1), indicating
that hormonal influence was decisive for our results.
Hormone supplements are a common way to exert birth
control in contemporary women and have to be considered in
further research. During a fertility window of approximately
40 years, other contraceptive methods have been used (e.g.,
condoms, temperature measurements), and subjects might
also have differed in terms of reproductive behaviour (e.g.,
copulation frequency). As these intervening factors were
Table 3
Descriptive statistics for categorical sociodemographic variables (family contact, education, and religiosity)
Variable Contraceptives Characteristics Level 0 Level 1 Level 2 Level 3 Level 4 Level 5
a
Family contact
(subjects' contact
to family members while
raising children)
No contraceptives
(N=40)
Frequencies 0 8 5 27 – –
Percentages (S.D.) 0.00 20 (6.32) 12.5 (5.23) 67.5 (7.41) – –
Use of contraceptives
(N=46)
Frequencies 1 12 3 30 – –
Percentages (S.D.) 2.17 (2.15) 26.09 (6.47) 6.52 (3.64) 65.22 (7.02) – –
b
Education of subject
(level of graduation,
apprenticeship)
No contraceptives
(N=40)
Frequencies 0 2 12 22 1 3
Percentages (S.D.) 0.00 5.00 (3.45) 30.00 (7.25) 55.00 (7.87) 2.50 (2.47) 7.50 (4.16)
Use of contraceptives
(N=46)
Frequencies 0 2 5 31 5 3
Percentages (S.D.) 0.00 4.35 (3.01) 10.87 (4.59) 67.39 (6.91) 10.87 (4.59) 6.52 (3.64)
b
Education of husband
(level of graduation,
apprenticeship)
No contraceptives
(N=40)
Frequencies 0 1 5 31 2 1
Percentages (S.D.) 0.00 2.50 (2.47) 12.50 (5.23) 77.5 (6.60) 5.00 (3.45) 2.50 (2.47)
Use of contraceptives
(N=46)
Frequencies 1 0 0 38 3 4
Percentages (S.D.) 2.17 (2.15) 0.00 0.00 82.61 (5.59) 6.52 (3.64) 8.7 (4.15)
c
Religiosity of subject
(church attendance)
No contraceptives
(N=40)
Frequencies 6 8 11 13 2 –
Percentages (S.D.) 15.00 (5.65) 20.00 (6.32) 27.50 (7.06) 32.50 (7.41) 5.00 (2.47) –
use of contraceptives
(N=46)
Frequencies 3 20 18 5 0 –
Percentages (S.D.) 6.52 (3.64) 43.48 (7.31) 39.13 (7.20) 10.87 (4.59) 0.00 –
Frequencies and percentages (with standard deviations) of sociodemographic variables collected in ordinal scale. a
Family members (FM) including FM of
subject and FM of corresponding husband (parents, grandparents, siblings, cousins, uncle, aunt). Levels: 0=no contact with FM, 1= contact with few FM,
2=contact with half of the FM, 3=contact with almost all FM. b
Levels: 0=no education, 1=primary school, 2=low-level secondary school, 3=middle-level
secondary school, 4=high-level secondary school, 5=university degree. c
Religiosity (church attendance). Levels: 0=never, 1=on holy days, 2=once a month,
3=once a week, 4=several times a week.
Fig. 1. Casual relation between reproductive success and perceived facial attractiveness in women who never used HCs (N=40). Left panel: correlation between
number of children and facial attractiveness (sexy: rs=.425, p=.001; attractive: rs=.374, p=.01). Right panel: correlation between number of pregnancies and
facial attractiveness (sexy: rs=.568, pb.001; attractive: rs=.514, pb.001). Regression lines (N=35) are calculated after excluding extreme values, resulting in
women with: one to four children in the left panel (sexy: rs=.265, p=.061; attractive: rs=.240, p=.071) and women with one to five pregnancies in the right panel
(sexy: rs=.468, p=.002; attractive: rs=.432, p=.005).
712 L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714
impossible to retrace, a future longitudinal study is needed in
which those parameters are constantly recorded throughout
the fertile period.
Almost all of our participants were mountain farmers and
were therefore supposed to be quite homogenous in terms of
socioeconomic factors. Nevertheless, we have an age range
of approximately 30 years in our sample. Also, subjects' own
income in both subsamples and religiosity of women who
never used contraceptives were highly variable (Tables 2 and
3). We controlled for the effect of all sociodemographic
variables even though none of these variables were related to
the number of children or influenced our findings in either of
our subsamples.
In line with previous research (e.g., Grammer & Thornhill,
1994; Mealey et al., 1999), we identified a negative
correlation between FA and facial attractiveness in our
pooled sample of N=78 women. While previous studies used
the same images for attractiveness ratings and determination
of FA, our ratings were based on pictures taken at youth,
while the facial shape was measured from photos taken
approximately 40 years later. This indicates that the link
between FA and attractiveness does not dissociate in life
history. In our subsamples of women who used HC and
women who did not, FA was neither related to attractiveness
nor to reproductive success. The former can be explained by
an insufficient subsample size compared to the pooled
sample of 78 women. Nonfindings for the latter are less
surprising because for women who reported a use of HC
results are biased by hormone intake. In our sample of
women who did not take contraceptives, we additionally
regressed facial shape on evaluated attractiveness and
lifetime reproduction. Visualizations of shape regression
via TPS deformation grids revealed distinctive facial
proportions in postmenopausal faces corresponding to a
highly attractive appearance at youth: full and large lips, a
small and tapered nose, high eyebrows, big pupils, and a less
Fig. 2. Shape regression of women who never used HCs visualized by TPS deformation grids (N=33). Top: facial shape and (perceived) attractiveness at youth
(mean R2
=.178, min R2
=.011, max R2
=.456). Middle: facial shape and number of children (mean R2
=.236, min R2
=.029, max R2
=.463). Bottom: facial shape
and number of pregnancies (mean R2
=.242, min R2
=.035, max R2
=.459). Middle panels of each row correspond to the Procrustes mean shape (no deformation
of the grid). Panels on the left correspond to a low (minimal) value of independent variable. Panels on the right correspond to a high (maximal) value of the
independent variable. To ease the visualization, TPS grids were extrapolated 1.2 times in negative and positive direction of each independent variable.
713L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714
angular jaw. Those sexually dimorphic or neotenous features
were not found in faces corresponding to a low level of
perceived attractiveness. Interestingly, postmenopausal faces
corresponding to a high reproductive success (high number
of children and pregnancies) displayed the same ‘typically
female’ features. As mentioned above, sexually dimorphic
proportions or so-called hormone markers (Fink et al., 2005)
are known to affect judgments of facial attractiveness in mate
choice (Cunningham, 1986; Jones et al., 1995; Grammer et
al., 2003) and are supposed be honest cues to fertility and
health (Grammer, Fink, Moller, & Manning, 2005). All three
independent variables used for shape regression were found
to be independent of current age and explained approximately
the same amount of shape changes (attractiveness at
youth: 18%; number of children/pregnancies: 24%). In
conclusion, our findings underline our assumption that facial
attractiveness at the age of mate choice (young adulthood) is
related to female reproductive success unless HCs were used.
Our findings also indicate that postmenopausal faces still
possess anthropometric indicators, which have driven the
perception of attractiveness at young age. In this manner, we
assume that a low degree of FA does not explain female
attractiveness and reproductive success as a single feature, as
it seems to be interconnected with other ‘honest’ cues
to fertility.
Supplementary Materials
Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.evolhumbehav.2012.05.005.
Acknowledgments
We thank all our participants for their willingness to take
part in our research. Special thanks goes to M. and E.
Oberzaucher for their generous help and hospitality during
data collection. We also thank our colleagues S.F. Schmehl,
K. Meindl, M. Coquerelle, and S. Senck for their support.
References
Bock, M. T., & Bowman, A. W. (2006). On the measurement and analysis of
asymmetry with applications to facial modelling. Journal of the Royal
Statistical Society: Series C (Applied Statistics), 55, 77–91.
Bookstein, F. L. (1991). Morphometric tools for landmark data: geometry
and biology. Cambridge: Cambridge University Press.
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.
Fink, B., Grammer, K., Mitteroecker, P., Gunz, P., Schaefer, K., Bookstein,
F., & Manning, J. T. (2005). Second to fourth digit ratio and face shape.
Proceedings of the Royal Society B, 272, 1995–2001.
Gangestad, S. W., Thornhill, R., & Yeo, R. A. (1994). Facial attractiveness,
developmental stability, and fluctuating asymmetry. Ethology and
Sociobiology, 15, 73–85.
Grammer, K., Fink, B., Moller, A. P., & Manning, J. T. (2005). Physical
attractiveness and health: comment on Weeden and Sabini (2005).
Psychological Bulletin, 131, 658–661.
Grammer, K., Fink, B., Møller, A. P., & Thornhill, R. (2003). Darwinian
aesthetics: sexual selection and the biology of beauty. Biological
Reviews, 78, 385–407.
Grammer, K., & Thornhill, R. (1994). Human (Homo sapiens) facial
attractiveness and sexual selection: the role of symmetry and
averageness. Journal of Comparative Psychology, 108, 233–242.
Gunz, P., Mitteroecker, P., & Bookstein, F. (2005). Semilandmarks in three
dimensions. In D. E. Slice, Ed. Modern morphometrics in physical
anthropology (pp. 73–97). New York: Springer US.
Hill, K., & Hurtado, M. (1996). Ache life history: the ecology and
demography of a foraging people. New York: Aldine de Gruyter.
Johnston, V. S., & Franklin, M. (1993). Is beauty in the eye of the beholder?
Ethology and Sociobiology, 14, 183–199.
Jokela, M. (2009). Physical attractiveness and reproductive success in
humans: evidence from the late 20th century United States. Evolution
and Human Behavior, 30, 342–350.
Jones, D., Brace, C. L., Jankowiak, W., Laland, K. N., Musselman, L. E.,
Langlois, J. H., Roggman, L. A., Pérusse, D., Schweder, B., & Symons,
D. (1995). Sexual selection, physical attractiveness, and facial neoteny:
cross-cultural evidence and implications [and comments and reply].
Current Anthropology, 36, 723–748.
Mardia, K. V., Bookstein, F. L., & Moreton, I. J. (2000). Statistical
assessment of bilateral symmetry of shapes. Biometrika, 87, 285–300.
Mealey, L., Bridgstock, R., & Townsend, G. C. (1999). Symmetry and
perceived facial attractiveness: a monozygotic co-twin comparison.
Journal of Personality and Social Psychology, 76, 151–158.
Parsons, P. A. (1992). Fluctuating asymmetry: a biological monitor of
environmental and genomic stress. Heredity, 68, 361–364.
Pawlowski, B., Boothroyd, L. G., Perrett, D. I., & Kluska, S. (2008). Is
female attractiveness related to final reproductive success? Collegium
Antropologicum, 32, 457–460.
R Development Core Team. (2010). R: a language and environment for
statistical computing. Vienna, Austria: R Foundation for Statistical
Computing.
Rhodes, G., Proffitt, F., Grady, J. M., & Sumich, A. (1998). Facial symmetry
and the perception of beauty. Psychonomic Bulletin & Review, 5, 659–669.
Rohlf, F. J., & Slice, D. (1990). Extensions of the Procrustes method for the
optimal superimposition of landmarks. Systematic Biology, 39, 40–59.
Schaefer, K., Fink, B., Grammer, K., Mitteroecker, P., Gunz, P., &
Bookstein, F. L. (2006). Female appearance: facial and bodily
attractiveness as shape. Psychology Science, 48, 187–204.
Slice, D. E. (2005). Mordern morphometrics in physical anthropology. New
York: Kluwer Academic, Plenum.
Slice, D. E., Bookstein, F. L., Marcus, L. F., & Rohlf, F. J. (1996). A
glossary for geometric morphometrics. In L. F. Marcus, M. Corti, A.
Loy, G. J. P. Naylor, & D. E. Slice (Eds.), Advances in morphometrics
(pp. 531–552). New York and London: Plenum Press, in cooperation
with NATO Scientific Affairs Division.
Thornhill, R., & Møller, A. P. (1997). Developmental stability, disease and
medicine. Biological Reviews, 72, 497–548.
Van Dongen, S., & Gangestad, S. W. (2011). Human fluctuating asymmetry
in relation to health and quality: a meta-analysis. Evolution and Human
Behavior, 32, 380–398.
Van Valen, L. (1962). A study of fluctuating asymmetry. Evolution, 16,
125–142.
714 L.S. Pflüger et al. / Evolution and Human Behavior 33 (2012) 708–714