Content uploaded by Pablo S C Santos
Author content
All content in this area was uploaded by Pablo S C Santos on May 23, 2019
Content may be subject to copyright.
New evidence that the MHC influences odor perception in humans:
a study with 58 Southern Brazilian students
Pablo Sandro Carvalho Santos, Juliano Augusto Schinemann,
Juarez Gabardo, Maria da Grac¸a Bicalho*
LIGH—Laborato´rio de Imunogene´tica e Histocompatibilidade, Departamento de Gene´tica, Setor de Cieˆncias Biolo´ gicas,
Centro Polite´cnico da Universidade Federal do Parana´ , Jardim das Ame´ ricas, Caixa Postal 19071, CEP: 81.530-990 Curitiba, Parana´, Brazil
Received 2 February 2004; revised 1 June 2004; accepted 5 November 2004
Available online 2 February 2005
Abstract
Increasing evidence suggests a correlation between mate choice, odor preference, and genetic similarity at the Major Histocompatibility
Complex (MHC) in a variety of animals, including our species. The MHC is a highly polymorphic group of genes that play an important role
in the immunological self/nonself recognition. Its products have been reported to take part on the variety of compounds and reactions that
together build an individual’s body odor. It has been suggested, therefore, that animals use body odor as a guide to identify possible mates as
MHC-similar or MHC-dissimilar from their own genotype. Preference for a MHC-dissimilar partner enhances MHC heterozygosity of an
individual’s offspring. The possible adaptive advantages are clear: it is a mechanism of avoiding inbreeding and MHC-heterozygous
offspring may have enhanced immunocompetence. The aim of this study was to search, in our species, new evidence on the correlation
between specificities at HLA-A and HLA-B and assessments of pleasantness regarding specific body odors. HLA is the name for the human
MHC. Four olfactory sessions were performed with 58 young Southern Brazilian students, in order to investigate whether assessments of
pleasantness of body odors from individuals correlate to a person’s HLA phenotype. Body odors were collected via sweat and urine from all
participants. Women smelled and scored all male odor samples and men did the same with all female samples. We found a significant
correlation only when female smellers evaluated male sweat odors.
D2004 Elsevier Inc. All rights reserved.
Keywords: HLA; Mate choice; MHC; Negative assortative mating; Odors; Reproductive behavior; Sex attractants; Smell
Introduction
Since Yamazaki and his colleagues published their
seminal paper in 1976, increasing interest has been directed
to the influence of the Major Histocompatibility Complex
(MHC) on mating preferences. Yamazaki and his colleagues
observed MHC-based disassortative mating preferences
mediated by odor in house mice, Mus musculus domesticus
(Yamazaki et al., 1976). Since then, several studies have
been published showing that the MHC is, somehow, a
source of unique personal odor, which influences individual
recognition and/or mating preferences in rodents (Beau-
champ et al., 1988; Eklund, 1997; Penn and Potts, 1998;
Yamazaki et al., 1978), in humans (Jacob et al., 2002; Ober
et al., 1997; Wedekind et al., 1995), in fish (Landry et al.,
2001; Milinski, 2003; Olse´n et al., 1998), and recently in
sand lizards (Olsson et al., 2003) and in birds (Freeman-
Gallant et al., 2003).
Brown (1979) refers very properly to individual odors
from mammals as bolfactory fingerprintsQ, as they seem to
enable detailed identification among individuals, conspecific
and not. In the same work, Brown proposes that mammals use
their body odor in order to provide a variety of information,
such as sex, species, age class (infant, juvenile, or adult),
colony or family membership, reproductive status (rut or
heat), social status, stress (fear or alarm state), maternal
status (pregnancy/lactation), and individual identity.
Implication of MHC genes in kin recognition has been
described in mice (Manning et al., 1992) and at least one
0018-506X/$ - see front matter D2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.yhbeh.2004.11.005
* Corresponding author. Fax: +55 41 2662042.
E-mail address: ligh@ufpr.br (M. da Grac¸a Bicalho).
Hormones and Behavior 47 (2005) 384 – 388
www.elsevier.com/locate/yhbeh
relatively recent report (Lewis, 1998) relates these genes
with the evolution of kin altruism.
One of the first works of this kind concerning humans
was carried out by Wedekind et al. (1995). They found that
women tend to prefer body odor from individuals carrying a
different HLA-genotype (the human MHC is called HLA,
Human Leukocyte Antigen). On their review of studies on
MHC-dependent mating preferences in many species, Penn
and Potts (1999) described three possible adaptive signifi-
cances of this phenomenon, which we summarize in two:
first, MHC-disassortative mating patterns produce MHC-
heterozygous offspring, which may have enhanced immu-
nocompetence and be more resistant against pathogens.
Second, MHC-disassortative mating preferences may help
avoiding inbreeding.
Although many works agree with the possible benefits of
maximizing MHC-heterozygosity, recent researches suggest
a different approach: there might be an optimum, inter-
mediate degree of MHC-heterozygosity, which would bring
more fitness than the strongest possible heterozygosity. One
example is the work by Milinski (2003), who studied a
species of fish, the three-spined Stickleback, and found that
female sticklebacks tend to optimize (and not to maximize)
the MHC-heterozygosity of offspring through mate choice.
Milinski (2003) demonstrated that offspring with an
intermediate number of different MHC alleles are fitter than
both strongly homozygous and strongly heterozygous
individuals. Jacob et al. (2002) came to comparable findings
in a very elegant work that had humans as subjects: their
study leads to the conclusion that women tend to prefer the
smell of a man that has an intermediate number of HLA
matches to their own, rather than the smell of men carrying
HLA alleles that are too similar or too different from theirs.
The MHC is a large chromosomal region containing
many linked, highly polymorphic genes that play a central
role in controlling immunological self/nonself recognition
(Klein, 1986), and its genetic diversity seems to be its major
feature. Many theories have been sought to explain how an
individual’s MHC-genotype shapes its body odors. In
another review, Penn and Potts summarized five hypotheses
explaining the possible ways that this could happen,
although none are proven to date (see Penn and Potts,
1998 for details).
An impressive number of the functional HLA genes are
related to the immune system, but many other HLA genes
(or HLA-linked) are not. Examples are some human
olfactory receptor-like genes, which have been reported to
be HLA-linked (Eklund et al., 2000; Fan et al., 1996;
Ziegler et al., 2000). Ziegler et al. (2000) demonstrated the
existence of more than thirty olfactory receptor genes
located at the HLA class I region, in the immediate vicinity
of the HLA-F locus. As Eklund et al. (2000) put it, HLA-
linked olfactory receptor genes provide a potential mecha-
nism for detecting HLA-specific odors.
The main goal of this work was to test whether HLA-A
and HLA-B specificities correlate to one individual’s
perception of a specific body odor or not. Moreover, we
aimed to test whether we were able to reproduce the main
findings of Wedekind et al. (1995) assessing only HLA class
I loci and using an alternative methodological approach.
Although it has been reported that the use of contra-
ceptives might mask female odor perception, we did not
distinguish between, in our statistical analysis, women who
were taking oral contraceptives from women who were not.
We found, however, a significant correlation between HLA
phenotypes from female smellers and from male represented
by sweat odor samples. We observed that when women
evaluated the smell of sweat coming from males, the level of
difficulty in deciding whether a sample was bpleasantQor
bunpleasantQincreased significantly with the increased level
of HLA-similarities between smeller and odor donor. We
did not observe a significant correlation when female
smellers scored male urine samples. No significant correla-
tion was observed when male smellers scored both kinds of
female samples (sweat and urine) either. Men neither dlikedT
nor ddislikedTa smell that was in some way related to a
woman’s HLA phenotype. Women seemed to do so, but
only when the odor sample was collected through sweat.
Methods
Subjects
A total of 58 Southern Brazilian students (29 male and 29
female) self-reported heterosexual from the Federal Uni-
versity of Parana´, city of Curitiba, took part in this study.
Curitiba is the capital city of the Southern Brazilian Federal
State of Parana´, with approximately 1.5 million inhabitants.
Mainly Germans, Italians, Polish, and Portuguese formed its
population, since the official immigration policy was
established in the XIX century. Nine female participants
(31.03% of female subjects) were taking oral contraceptives.
Participants were aged between 18 and 27 years (average
age 19.76 years). None of them had relatives taking part in
the study. All participants were informed about the goal of
this investigation: a search for a correlation between their
odor perception and their HLA phenotypes, using an
analysis of HLA matches and mismatches between smellers
and odor donors. All subjects agreed to follow the necessary
restrictions on their personal routine (described below)
throughout the time for preparation and performance of the
olfactory sessions. Ethical approval for this study was
obtained from the Ethics Committee at Hospital de Clı´nicas,
and informed consent to participate in the study was
obtained from all individuals.
Phenotypings
On August 1999, all participants had their HLA-A and
HLA-B serologically typed at the Laborato´rio de Imunoge-
ne´tica e Histocompatibilidade—LIGH, at the Federal Uni-
P. Sandro Carvalho Santos et al. / Hormones and Behavior 47 (2005) 384–388 385
versity of Parana´. Phenotypings were performed using
peripheral blood lymphocytes and a ONE LAMBDA
Special Monoclonal Typing Tray Set—SMT144A. The
privacy of the personal data of all participants was assured
by the laboratory named above.
Preparation for the olfactory sessions
On May 2000, each person was given a necklace with a
sachet that was made of absorbent cotton. The sachets were
manufactured by covering and sewing a 4 cm 4cm
0.5 cm patch of 100% untreated cotton with a cotton gauze.
All subjects carried the sachet on their breasts, at the height
of the sternum (an area of high perspiration), for 5 days,
aiming the absorbance of body odors through sweat. During
this time, they were allowed to use only neutral personal
hygiene products, all from the same lot provided by a
Brazilian cosmetics company. Participants were asked to
avoid other products that could mask their natural odors,
such as cigarettes, chewing gums, alcohol, and all strong
natural or artificial flavored food. They were allowed to take
a shower, but not to bathe, once in a day, using the provided
personal hygiene products and after placing their sachets in
a closed plastic recipient that had been given to each
participant for that purpose. After 1 h, they should be
wearing the necklace with the sachet again. Sexual contact
should be avoided. After the five preparation days, all
sachets were collected and put in sealed plastic bags. All
bags were identified with a code. The first urine of the fifth
preparation day from each participant was also collected.
Stripes of absorbent paper were put in contact with each
urine sample so that each stripe represented one individual’s
urine sample. Stripes were put in sealed plastic bags that
were also identified with a code. The olfactory tests began
on the day after the fifth preparation day.
The olfactory sessions
Tab le 1 summarizes the four olfactory sessions, which
were performed in 2 days: sessions 1 and 2 on the first day
and sessions 3 and 4 on the day after. Participants received
on both days a worksheet in which they evaluated the odor
from each sample using three possible categories: pleasant,
indifferent, or unpleasant. Before each evaluation, they
briefly sniffed a sample of coffee powder that had been
given to each of them, so that they could avoid misinter-
pretation between the odors from different samples.
Participants were aware that they were scoring samples
from individuals from the opposite sex, but they were
ignorant of the source person of each odor sample.
Statistical analysis
At each of the two HLA loci studied, there was a
possibility of up to two allele matches, for a maximum of
four matches.
We compared male HLA-A and HLA-B phenotypes with
female HLA-A and HLA-B phenotypes and stratified the
possible matches considering all possible male/female
combinations in 4 categories as follows:
0 —When a pair shared no HLA specificities.
1 —When a pair shared 1 HLA specificity.
2 —When a pair had 2 HLA specificities in common.
3 —When a pair had 3 HLA specificities in common.
The situation in which a pair shared all 4 specificities
(HLA identicals) was not observed.
Chi-square tests (independence and goodness-of-fit)
were used for all comparisons. Data analysis was performed
using the electronic tools MICROSTAT (Ecosoft, Inc.) and
EpiCalc 2000 (v.1.02, Joe Gilman and Mark Myatt, 1998,
Brixton Books).
Results and discussion
Table 2 displays the results of the first olfactory session,
in which female participants scored the sachets that had
been worn by male subjects.
Out of the four olfactory sessions, the one shown on
Table 2 was the only session in which we observed a
significant correlation (v
2
= 14.37; df =4;P= 0.0062)
between degrees of HLA-similarities and assessments of
pleasantness.
The total number of scorings on Table 2 is 838, and not
841, as it would be expected (29 smellers 29 donors)
because three female smellers failed to score one different
sachet each. All three blostQsmeller/donor pairs belonged to
the category bzero matchesQ.
Due to the small number of smeller/donor pairs belong-
ing to the category b3 matchesQ(there were 7 pairs in this
category), we decided to merge that category with its upper
immediate, the category b2 matchesQ.
Table 1
Summary of the olfactory sessions performed
Olfactory session Smellers Donors Odor
source
No. of
assessments
1 29 Female 29 Male Sweat 838
a
2 29 Male 29 Female Sweat 838
b
3 29 Female 29 Male Urine 837
c
4 27 Male 29 Female Urine 781
d
a
Number differs from 841 (29
2
) because three smellers failed to score one
different sample each, all belonging to the category bno matchesQ.
b
Number differs from 841 because three smellers failed to score one
different sample each, all belonging to the category bno matchesQ.
c
Number differs from 841 because four smellers failed to score one
different sample each, three in the category bno matchesQand one in the
category bone matchQ.
d
Number differs from 841 because two male smellers did not come to this
olfactory session and one smeller failed to score 2 urine samples, category
bone matchQ.
P. Sandro Carvalho Santos et al. / Hormones and Behavior 47 (2005) 384–388386
A preliminary analysis of the results could suggest
similar findings as those by Wedekind et al. (1995): female
smellers are more likely to find the odor of a male’s sweat
pleasant when they share less HLA specificities. This would
be, however, a misinterpretation. Chi-squared goodness-of-
fit tests were performed in each score category (shown in
Table 2), testing against null hypothesis that our data fit the
expected data. The results of the goodness-of-fit tests
indicate that the samples scored as bindifferentQare
responsible for driving the statistical association of the first
olfactory session (v
2
= 10.12; df =2;P= 0.0064), and not
the samples evaluated as bpleasantQ(v
2
= 1.46; df = 2; n.s.),
or bunpleasantQ(v
2
= 2.79; df = 2; n.s.).
The other three olfactory sessions showed results that,
according to our statistical analysis, correspond to a pattern
of odor assessments regarding HLA phenotypes similar to
the pattern that would be expected by chance (also shown
on Table 1). The results of the chi-squared independence
tests are for session 2 (v
2
= 4.632; df = 6; n.s.), 3 (v
2
=
6.308; df = 6; n.s.), and 4 (v
2
= 3.852; df = 6; n.s.) not
significant.
As Koelega and Kfster (1974) concluded from a series of
experiments in odor perception, our results also suggest that
men seem to be less sensitive to smelling stimuli than
women.
The work by Herz and Inzlicht (2002) supports this view,
showing that this kind of stimulus seems to be much more
important for women than for men, when many variables for
sexual interest are compared.
Our results regarding the sweat samples are not in
agreement with the findings of Wedekind et al. (1995) and
Wedekind and Fu¨ri (1997), who found that perceived
pleasantness of body odors tend to correlate negatively
with the degree of HLA similarity between female smellers
and male sample donors. Instead, our study indicates that a
significant increase of difficulty to decide between pleasant
or unpleasant regarding sweat odors correlates positively
with the increase of HLA-class I similarities between odor
smeller and donor. Moreover, works by Wedekind et al.
(1995) and Wedekind and Fu¨ri (1997) did find an asso-
ciation when male acted as smellers, whereas we did not.
Both differences could be related to the fact that subjects in
our study were typed only for HLA class I loci. As a result,
smeller/donor pairs that we classified in the category bzero
matchesQcould actually match for one or two class II loci.
According to the review by Penn and Potts (1999), no study
demonstrating MHC-dependent mating and odor preferen-
ces had ignored MHC class II genes until that time. The
most consistent evidence linking the HLA class I region
with odor perception are works that relate olfactory receptor
genes with the HLA class I region (Eklund et al., 2000; Fan
et al., 1996; Ziegler et al., 2000). It is notable that a
significant correlation between odor perception and HLA-
similarities was observed in this work in spite of the fact that
we did not investigate HLA class II genes.
Our results (and those by Wedekind and Fu¨ ri, 1997;
Wedekind et al., 1995) differ from the findings by Jacob
et al. (2002), who reported that women prefer odors coming
from men with an boptimalQ, intermediate number of HLA
matches with their own. An important difference between
their methods (Jacob et al., 2002) and ours is that we did not
perform a preference test, whereas they did. Subjects in our
study were judging pleasantness (blikingQ), when they
scored the odor samples, while subjects in their study
(Jacob et al., 2002) were actually choosing (bwantingQ)a
specific odor. bLikingQand bwantingQare, according to the
review of Berridge and Robinson (1998), rather indepen-
dent. The paradoxical bwantingQwithout blikingQis quite
possible, as blikingQand bwantingQseem to involve different
brain substrates (Berridge and Robinson, 1998). Other
differences concerning the study design, such as the
statistical treatment of data and the selection of smellers,
make it difficult to directly compare their study (Jacob et al.,
2002) with ours.
Still, our results differ from those of Wedekind et al.
(1995) and Wedekind and Fu¨ri (1997) in that we did not
stratify, in our statistical analysis, women who were taking
oral contraceptives from women who were not. Notably, a
significant correlation was observed in spite of this fact,
which, according to Pause et al. (1996), modulates olfactory
perception. Similarly, the study by Jacob et al. (2002) found
that women can discriminate between odors from donors
with different HLA types, testing women without regard to
menstrual cycle phase.
Our results show that a significant correlation between
HLA-similarities and female odor perception takes place
even without assessing two features that are had to be
decisive on olfactory perception: HLA class II genes and
menstrual cycle phase/use of oral contraceptives.
As Wedekind and Fu¨ri (1997) put it, odor preferences in
humans appear to be context-dependent. Personal history,
hormonal status, and psychological context are a few of the
possible variables that are likely to influence human
Table 2
Olfactory session 1: number of HLA matches (M) between female smellers
and male donors versus female assessments of male sweat odor
Number Score categories Total
of HLA
matches
(M)
Pleasant* Indifferent** Unpleasant***
0M
a
192 (48.24%) 82 (20.60%) 124 (31.16%) 398 (100%)
1M
b
154 (43.63%) 90 (25.50%) 109 (30.88%) 353 (100%)
2or3M
c
35 (40.23%) 34 (39.08%) 18 (20.69%) 87 (100%)
Total 381 (45.47%) 206 (24.58%) 251(29.95%) 838 (100%)
v
2
-test for independence = 14.37; df =4;P= 0.0062.
a
29 smellers and 29 donors involved.
b
29 smellers and 29 donors involved.
c
23 smellers and 21 donors involved.
*v
2
-goodness-of-fit test in the bpleasantQcategory: v
2
= 1.46; df = 2; n.s.
** v
2
-goodness-of-fit test in category bindifferentQ:v
2
= 10.12; df =2;P=
0.0064.
*** v
2
-goodness-of-fit test in the bunpleasantQcategory: v
2
= 2.79; df =2;
n.s.
P. Sandro Carvalho Santos et al. / Hormones and Behavior 47 (2005) 384–388 387
preferences. It is our understanding that in studies with
humans, we can at best conclude that HLA loci or HLA-
linked genes influence body odor production and/or per-
ception. Extrapolative considerations on human sexual
attraction or mating preferences still do not have empirical
support. In this concern, new studies with alternative
approaches need to be carried out, in order to test a broader
spectrum of different contexts that might be likely to
influence human mating behavior.
References
Beauchamp, G.K., Yamazaki, K., Bard, J., Boyse, E.A., 1988. Preweaning
experience in the control of mating preferences by genes in the major
histocompatibility complex of the mouse. Behav. Genet. 18, 537 – 547.
Berridge, K.C., Robinson, T.E., 1998. What is the role of dopamine in
reward: hedonic impact, reward learning, or incentive salience? Brain
Res. Rev. 28, 309 – 369.
Brown, R.E., 1979. Mammalian social odours. Adv. Stud. Behav. 10,
103 – 109.
Eklund, A., 1997. The effect of early experience on MHC-based mate
preferences in two B10.W strains of mice (Mus domesticus). Behav.
Ecol. 8, 630 – 634.
Eklund, A.C., Belchak, M.M., Lapidos, K., Raha-Chowdhury, R., Ober, C.,
2000. Polymorphisms in the HLA-linked olfactory receptor genes in the
Hutterites. Hum. Immunol. 61, 711– 717.
Fan, W., Weiwen, C., Parimoo, S., Lennon, G.G., Weissman, S.M., 1996.
Identification of seven new MHC class I region genes around the HLA-
F locus. Immunogenetics 44, 97 – 103.
Freeman-Gallant, C.R., Meguerdichian, M., Wheelwright, N.T., Sollecito,
S.V., 2003. Social pairing and female mating fidelity predicted by
restriction fragment length polymorphism similarity at the major
histocompatibility complex in a songbird. Mol. Ecol. 12, 3077 – 3083.
Herz, R.S., Inzlicht, M., 2002. Sex differences in response to physical and
social factors involved in human mate selection. The importance of
smell for women. Evol. Hum. Behav. 23, 359 – 364.
Jacob, S., McClintock, M.K., Zelano, B., Ober, C., 2002. Paternally
inherited HLA alleles are associated with women’s choice of male odor.
Nat. Genet. 30, 175 – 179.
Klein, J., 1986. Natural History of the Major Histocompatibility Complex.
Wiley, New York.
Koelega, H.S., Kfster, E.P., 1974. Some experiments on sex differences in
odor perception. Ann. N. Y. Acad. Sci. 237, 234–246.
Landry, C., Garant, D., Duchesne, P., Bernatchez, L., 2001. Good genes
as heterozygosity: the major histocompatibility complex and mate
choice in Atlantic salmon (Salmo salar). Proc. R. Soc. Lond., B 268,
1279 – 1285.
Lewis, K., 1998. Pathogen resistance as the origin of kin altruism. J. Theor.
Biol. 193, 359 – 363.
Manning, C.J., Wakeland, E.K., Potts, W.K., 1992. Communal nesting
patterns in mice implicate MHC genes in kin recognition. Nature 360,
581 – 583.
Milinski, M., 2003. The function of mate choice in sticklebacks: optimizing
Mhc-genetics. J. Fish Biol. 63 (Suppl. A), 1 – 16.
Ober, C., Weikamp, L.R., Cox, N., Dytch, H., Kostyu, D., Elias, S., 1997.
HLA and mate choice in humans. Am. J. Hum. Genet. 61, 497 – 504.
Olse´n, K.H., Grahn, M., Lohmt, J., Langefors, 2., 1998. MHC and kin
discrimination in juvenile Artic charr, Salvelinus alpinus (L.). Anim.
Behav. 56, 319 – 327.
Olsson, M., Madsen, T., Nordby, J., Wapstra, E., Ujvari, B., Wittsell, H.,
2003. Major histocompatibility complex and mate choice in sand
lizards. Proc. R. Soc. Lond., B Biol. Sci. 270 (Suppl. 2), 254 – 256.
Pause, B.M., Sojka, B., Krauel, K., Fehm-Wolfsdorf, G., Ferstl, R., 1996.
Olfactory information processing during the course of the menstrual
cycle. Biol. Psychol. 44, 31 – 54.
Penn, D.J., Potts, W.K., 1998. How do major histocompatibility genes
influence odor and mating preferences? Adv. Immun. 69, 411 – 435.
Penn, D.J., Potts, W.K., 1999. The evolution of mating preferences and
major histocompatibility complex genes. Am. Nat. 153, 145 – 164.
Wedekind, C., Fqri, S., 1997. Body odour preferences in men and women:
do they aim for specific MHC combinations or simply heterozygosity?
Proc. R. Soc. Lond., B 264, 1471 – 1479.
Wedekind, C., Seebeck, T., Bettens, F., Paepke, A.J., 1995. MHC-
dependent mate preferences in humans. Proc. R. Soc. Lond., B Biol.
Sci. 260, 245 – 249.
Yamazaki, K., Boyse, E.A., Mike, V., Thaler, H.T., Mathieson, B.J., Boyse,
E.A., Zoyas, Z.A., Thomas, L., 1976. Control of mating preferences in
mice by genes in the major histocompatibility complex. J. Exp. Med.
144, 1324 – 1335.
Yamazaki, K., Yamaguchi, M., Andrews, P.W., Peake, B., Boyse, E.A.,
1978. Mating preferences of F2 segregants of crosses between MHC-
congenic mouse strains. Immunogenetics 6, 253– 259.
Ziegler, A., Ehlers, A., Forbes, S., Trowsdale, J., Volz, A., Younger, R.,
Beck, S., 2000. Polymorphisms in olfactory receptor genes: a cau-
tionary note. Hum. Immunol. 61, 1281 – 1284.
P. Sandro Carvalho Santos et al. / Hormones and Behavior 47 (2005) 384–388388
