Content uploaded by Jan Havlicek
Author content
All content in this area was uploaded by Jan Havlicek on Aug 12, 2017
Content may be subject to copyright.
ORIGINAL PAPER
Shaving of axillary hair has only a transient effect
on perceived body odor pleasantness
Dagmar Kohoutová &Anna Rubešová &Jan Havlíček
Received: 22 November 2011 / Revised: 1 December 2011 / Accepted: 5 December 2011 / Published online: 29 December 2011
#Springer-Verlag 2011
Abstract In contrast to other apes, humans have relatively
greater amounts of armpit hair, which is thought to retain
signaling molecules. Although armpit shaving is widespread
cross-culturally, its effect on body odor has been little
investigated. In four experiments, we tested the effect of
shaving and the subsequent regrowth of axillary hair. Armpit
odors were collected from men who regularly shaved
(group S) or who had never shaved (group N) their
armpits before. The samples were subsequently rated
by women for intensity, pleasantness, and attractiveness.
In Experiments I,II (group N) and III, subjects firstly
shaved one armpit and then let the hair regrow over 6
or 10 weeks. In Experiments I,II (group S) and IV,
subjects shaved both armpits before the sampling and
subsequently shaved one armpit during the same period,
leaving the second armpit unshaved. Odors of the
shaved armpits were rated more pleasant, attractive,
and less intense compared to the unshaved armpits
(Experiment I(group N)). However, no significant dif-
ferences found in Experiments II and III (group N)
suggest the effect of shaving is relatively minor. More-
over, there were no significant differences in odor
comparing unshaved armpits with armpits after 1 week
of regrowth (Experiments I,II (group N) and III)or
comparing regularly shaved armpits with armpits after 1
or 3 weeks of regrowth (Experiments I,II (group S) and
IV). The odor of shaved armpits was rated significantly
more attractive compared to the armpits where hair had
beenregrowingfor6or10weeks.
Keywords Armpit .Attractiveness Human .Olfaction .
Scent .Smell
Introduction
The last two decades have witnessed an outburst of interest
in human chemical ecology, both in the scientific literature
and the popular media. The majority of studies have focused
on genetically related traits influencing body odor, particu-
larly in mate choice and mother–infant contexts. These
include research on individuality, gender, MHC, and other
aspects intrinsic to the odor carrier (for recent reviews see
Hays 2003; Wysocki and Preti 2004; Havlicek and Roberts
2009). However, as in other species, human body odor is
also shaped by numerous environmental factors. Most non-
genetic variation can be accounted for by differences in
reproductive status, emotional state, infections, and groom-
ing habits (such as using perfumes or deodorants) (reviewed
in Havlicek and Lenochova 2008). For instance, it has been
repeatedly demonstrated that men find the body odor of
women not taking oral contraceptives most appealing when
conception is most likely (e.g. Kuukasjärvi et al. 2004;
Havlicek et al. 2006). Dietary habits may also have a pro-
found effect on body odor. Havlicek and Lenochova (2006)
recently showed that the odor of men on a red meat diet was
judged less attractive than on a meat-free diet.
Communicated by T. Bakker
D. Kohoutová :J. Havlíček (*)
Department of Anthropology, Faculty of Humanities,
Charles University,
Husníkova 2075,
155 00 Prague 5, Czech Republic
e-mail: jan.havlicek@fhs.cuni.cz
A. Rubešová
Department of Philosophy and History of Sciences,
Faculty of Science, Charles University,
Viničná 7,
128 44 Prague 2, Czech Republic
Behav Ecol Sociobiol (2012) 66:569–581
DOI 10.1007/s00265-011-1305-0
The axillary region is the most intensely studied body part
regarding chemical communication. Human axillary odor has
an individual character dependent on specific microflora pres-
ent in the human armpit area (Shelley et al. 1953). Aerobic
coryneforms, propionibacteria, staphylococci, and micrococci
are the major axillary microbial organisms (Rennie et al.
1991). These microorganisms metabolise the products
of axillary glands, creating volatile odors (Rennie et
al. 1991; Gower et al. 1994;Natschetal.2003). The
axillary area has some unique features, such as, the
abundance of apocrine glands. This led some theorists
to speculate that in humans the axillary products specif-
ically evolved for communication (e.g. Comfort 1971).
The location might be especially advantageous due to
the upright posture of modern humans (Pawlowski
1999). Another specific feature is the presence of axil-
lary hair which has been proposed to serve to retain
chemical compounds active in communication processes
(Cohn 1994). This is supported by the study of Nixon
et al. (1988) who found 16-androstenes in the axillary
hair extracts. Some of these compounds, in particular
androstenol, androstenone and androstadienone, have
been shown to influence various social interactions
(Cowley and Brooksbank 1991; Jacob et al. 2002;
Lundstrom et al. 2003; Pierce et al. 2004;Saxtonet
al. 2008; for review see Havlicek et al. 2010). Other
odoriferous axillary chemicals, in particular various sat-
urated and unsaturated fatty acids, are also thought to
bind to armpit hair (Zeng et al. 1991;Natschetal.
2003,2006). Moreover, other great apes have apparently
less hair in their armpits (S Lhota personal communica-
tion) which is again suggestive of the idea that human
axillary hair could have evolved as an adaptation.
Shaving of axillary hair is a widespread practice not only
within the context of Western cultures but also in the Near
East, India and elsewhere. In Western cultures, armpit shav-
ing is regarded as a social norm in women, with the over-
whelming majority of women regularly performing it
(Tiggemann and Kenyon 1998; Tiggemann and Hodgson
2008). Recently, it is becoming popular among men as well.
However, actual figures, together with a record of the social
profile and motives of men engaged in this grooming activ-
ity, are missing. Interestingly, the effect of axillary hair
shaving on body odor hedonics has not attracted much
research attention. Until almost 60 years ago, as far as we
know, only one study had addressed this issue. It found that
removal of axillary hair in men resulted in a marked reduc-
tion or elimination of axillary odor for the next 24 h (Shelley
et al. 1953). However, the raters indicated only whether they
were able to smell any odor; their judgments on odor
hedonics or strength were not recorded. Moreover, the effect
of odor dynamics due to hair growth on subjective percep-
tion was not tested in the study.
The aim of our study was to test the effect of armpit
shaving in men and the consequential influence of axillary
hair growth on axillary odor hedonics, as rated by women.
The putative function of axillary hair is to retain armpit
chemicals and therefore we expected that armpit shaving
would decrease the intensity and increase the pleasantness
and attractiveness of the axillary odor. Simultaneously, we
expected that the odor of armpits with regrowing hair would
be rated more intense and less pleasant and attractive com-
pared to the odor of regularly shaved armpits. To test our
hypothesis, we performed four independent experiments
with male odor donors who regularly shaved their armpits
(group S in Experiments Iand II, and Experiment IV)as
well as those who had never shaved their armpits before
(group N in Experiments Iand II, and Experiment III). As
has been found repeatedly, both odor sensitivity and odor
preferences change in females across the course of the
menstrual cycle (Doty et al. 1981;Hummeletal.1991)
and are also affected by hormonal contraception (Caruso et
al. 2001; Roberts et al. 2008). Thus, our sample of odor
raters consisted of both users (Experiment II) and non-users
(Experiments I,III and IV) of hormonal contraception.
Experiments Iand II explored the effect of up to 6 weeks
of hair regrowth and Experiments III and IV continued for
an extended period of 10 weeks (see Fig. 1for a schedule of
the procedure). Different donors participated in each
experiment, and the same was true for the vast majority
of the raters (four raters who participated in Experiment
III and three raters in Experiment IV took also part in
Experiment I).
Methods
Subjects
To avoid body odor fluctuations across the menstrual cycle,
we chose only male subjects as odor donors. All of them
were students of various Prague universities, and none
reported dermatological or other diseases. Sexual orienta-
tion has previously been found to have an effect on hedonic
ratings of body odor (Martins et al. 2005; Sergeant et al.
2007). We therefore take this factor into account by asking
our odor donors: “What is your sexual orientation?”One
donor in Experiment III and two in Experiment IV identified
themselves as homosexual. Excluding homosexual donors
did not significantly affect the results. The donors were
awarded 1000 CZK (approximately US $55) as compensa-
tion for their time and potential inconvenience. The number
of samples used for the analyses of individual sessions and
the overall analyses varies as some of the donors did not
attend all of the sampling sessions (e.g. due to illness). We
further excluded from the analyses samples in which more
570 Behav Ecol Sociobiol (2012) 66:569–581
than half of the women reported contamination by smoke or
perfume. Female students participated as raters in all experi-
ments. They were unpaid but received a 100-g chocolate bar
after each testing session. The number of raters varies
between individual sessions as some of them did not
attend all rating sessions. Table 1summarizes the num-
ber of samples and raters used and excluded from
particular analyses.
Experiment I
Eleven men (mean age, 20.7; range, 20–23 years) served as
odor donors. They comprised two experimental conditions:
group N (n06) who had never shaved their armpits before
and group S (n05) who had been regularly shaving their
armpits for at least 1 year. Thirty female raters (mean age,
24.2; range, 18–30 years) participated in this experiment. To
Fig. 1 Male armpit treatments across sessions. Note that the control
axilla in group N of Experiments Iand II and in Experiment III was
unshaved. In contrast, the control axilla in group S of Experiments I
and II and in Experiment IV was regularly shaved. Experiments Iand II
consisted of four sampling sessions and Experiments III and IV con-
sisted of three sampling sessions
Behav Ecol Sociobiol (2012) 66:569–581 571
avoid a possible effect of menstrual cycle phase on odor
perception, we only recruited subjects using hormonal con-
traception (21 users of single-phase hormonal contraception
and nine users of another type of hormonal contraception).
Experiment II
Eleven men (mean age, 23.0; range, 20–26 years) partici-
pated as odor donors. Six had never shaved their armpits
(group N), and five had been regularly shaving their armpits
for at least 1 year (group S). Twenty-five women (mean age,
23.7; range, 19–32 years) participated as raters. To assess
possible effects of the menstrual cycle on odor perception,
we recruited only subjects not using hormonal contracep-
tion. All raters but one reported having a normal menstrual
cycle length (23–35 days). The cycle length was estimated
according to the reported usual length of the cycle. The day
of the menstrual bleeding onset was considered to be the
first day of their cycle. The menstrual cycle data were split
into fertile (days 7–14) and non-fertile phases (days 1–6 and
days 15–28) in a 28-day cycle following a simplified
protocol used by Havlicek et al. (2006). If the cycle
length was other than 28 days, the end of the fertile
phase was computed as F0L−14, where Fis the last
day of the fertile phase and Lthe length of the cycle.
Numbers in the fertile phase per session were as fol-
lows: Session 1, 10 out of 25; Session 2, 6 out of 19;
Session 3, 11 out of 22; Session 4, 5 out of 18.
Experiment III
Twelve men (mean age, 21.8; range, 19–28 years) who had
never shaved their armpits participated in this experiment as
odor donors. Seventeen women (mean age, 22.9; range, 19–
28 years) using hormonal contraception (ten of them used
single-phase hormonal contraception) participated as raters.
Experiment IV
Eleven men (mean age, 21.9; range, 19–27 years) who had
been regularly shaving their armpits for at least 1 year
before the study started participated as odor donors.
Twenty raters (mean age, 22.6; range, 19–27 years)
using hormonal contraception (15 of them used single-
phase hormonal contraception) participated in this
experiment.
Table 1 Numbers of subjects and temperatures across individual sessions in Experiments I–IV
Donors N (excluded) Donors S (excluded) Raters (absent) Outdoor temperature (°C) Indoor temperature (°C)
Experiment I Session 1 6 5 28 (2) 3.4 19
Session 2 5 (P) 5 29 (1) 4.4 17–20
Session 3 6 4 (A) 24 (6) 15.7 20–21
Session 4 6 5 21 (9) 14.5 18–19
OA 5 4 19
Experiment II Session 1 5 (S) 5 25 3.8 17–20
Session 2 4 (S, A) 5 19 (6) 7.4 18–19
Session 3 6 5 24 (1) 15.5 20–21
Session 4 6 5 18 (7) 15.1 19–19.5
OA 4 5 12
Experiment III Session 1 12 17 7.2 20–21
Session 2 11 (P) 13 (4) 2.1 17–19
Session 3 12 15 (2) 5.5 17–18
OA 11 12
Experiment IV Session 1 11 20 6.6 18–20
Session 2 10 (A) 18 (2) 4.6 16–19
Session 3 11 17 (3) 11.2 18–19
OA
a
10 15
Donors N indicates men who had never shaved their armpits before and donors S men who regularly shaved one of their armpits across the four
experimental sessions. The capital letters in brackets indicate the reason why a particular donor was excluded from the analyses. Values in brackets
(raters column) indicate the number of raters who did not attend a particular session. Analyses of individual sessions included all non-excluded
donors and raters in each session. Temperature measurements represent the average daytime temperature during the individual sampling sessions
(“Outdoor temperature”) and the range of temperatures during the rating sessions (“Indoor temperature”)
Aabsent, Pcontaminated by perfume, Scontaminated by smoke, OA overall analyses
a
Data of subjects who took part in all sessions and adhered to all instructions were included
572 Behav Ecol Sociobiol (2012) 66:569–581
Odor sampling procedure
In Experiments Iand II, odor samples were provided four
times over 6 weeks to test the effect of hair regrowth (see
below for details). Following the initial sampling session
(Session 1), the subsequent sessions took place after 1
(Session 2), 3 (Session 3) and 6 weeks (Session 4).
In Experiments III and IV, odor sampling was carried out
three times over 10 weeks. Specifically, following the initial
sampling session (Session 1), the subsequent sessions took
place after 6 (Session 2) and 10 weeks (Session 3).
In Experiments I,II (group N) and III, the effect of
shaving on formerly non-shaved armpits and the temporal
effect of hair regrowth was tested by asking each subject to
shave one of his armpits, randomly assigned by the exper-
imenter, the evening before the first odor sampling, and then
let the hair grow (see Fig. 1).
In Experiments I,II (group S) and IV, the effect of hair
growth on shaved armpit odor was tested by asking subjects to
shave both their armpits every other day for 1 week before the
sampling for Session 1. Subsequently, they shaved one armpit
every other day during the 6- or 10-week experimental period,
and the second armpit was left unshaved (see Fig. 1). The side
of armpit to regrow was again randomly assigned by the
experimenter.
Two days before each sampling session, the subjects
were provided with cotton pads, a T-shirt, a plaster, two
zip-lock plastic bags for storing the odor samples, razors
(Wilkinson Sword Extra II Sensitive), non-perfumed soap
(Sara Lee Household and Body Care, Sweden) and an
instruction sheet. The donors were instructed to refrain from
the following activities 2 days before collecting the samples
and on the day of wearing the pads: (1) using perfumes,
deodorants, antiperspirants, aftershave and shower gels; (2)
eating meals containing garlic, onion, chili, pepper, vinegar,
blue cheese, cabbage, radish, fermented milk products, mar-
inated fish; (3) drinking alcoholic beverages or using other
drugs; (4) smoking; (5) sexual activity; (6) sleeping in a bed
with a partner or pet. Whilst wearing the pads, they were
also asked to avoid intense physical activity (e.g. jogging,
playing football, etc.).
Odor samples were collected using elliptical pads approx-
imately 9× 7 cm in size madeof 100% cotton (Ebelin cosmetic
pads, DM-drogerie markt, www.dm-drogeriemarkt.cz)worn
in the armpits (cf methods in Havlicek et al. 2005,2006). In
the evening before the sampling session, the odor donors
showered without using even the non-perfumed soap. In the
morning (7 a.m.) the donors applied their cotton pads to both
armpits using 3 M™Micropore™surgical tape and wore
them for 24 h. Sampling length was kept constant
across the experiments as it can affect the quality of
the samples (Havlíček et al. 2011). To avoid odor con-
tamination from the donors’clothes or from the
background, donors wore the white 100% cotton T-shirts
provided (previously washed without washing powder) as
the first layer of their clothing. The following morning they
put the pads into the two labeled zip-lock plastic bags and
handed them back to the experimenters.
The average outdoor temperatures (source: Research
Institute of Crop Production in Prague) during individ-
ual sampling sessions are shown in Table 1. However,
the actual temperatures for individual donors might vary
slightly as they spent the sampling day in various parts
of Prague.
Odor rating procedure
The experimental procedure is based on a within-subjects
design as the same group of raters assessed the odor samples
repeatedly across all sessions. Odor ratings started within an
hour after collecting all the samples to minimize the possible
influence of further bacterial activity on the samples. The
ratings took place in a quiet, ventilated room. Indoor tem-
peratures during individual sessions are shown in Table 1.
Cotton pads were placed into 500-ml opaque glass jars
labeled by a code. The testing procedure lasted from 9 a.
m. to 6 p.m. Individual raters attended each rating session
approximately at the same time of day. This procedure
minimizes a possible diurnal effect on the olfactory abilities
and temporal changes of the odors. The total sample set
consisted of 22 axillary samples (two from each donor) in
Experiments I,II and IV andof24axillarysamplesin
Experiment III. Additionally, we used two non-human
samples to control for possible fluctuations in the raters’odor
preferences. One was of floral origin, cinnamon (75% cinna-
mal, 14% eugenol and several other minor compounds) and
one of animal origin, castoreum (a musky complex compound
that originates from the beaver’s anal gland; we used a mixture
of 33% dipropylene glycol, 16% thujopsene, 10% gurjunene,
9% benzyl benzoate, 8% cederene and several other minor
compounds manufactured by the chemical industry). In
both cases, two drops (i.e. approximately 0.1 ml) of
100% essence (Aroma Corp. Decin) were put on a
cotton pad and was placed into a 500-ml opaque glass
jar labeled by a code, thus replicating the procedure
used with the body odor samples.
Each rater assessed all samples. The samples were divid-
ed into two equal subsets. The order of the subsets was
randomized in Session 1, and in the subsequent sessions,
the order of the subsets was the same as in Session 1. The
order of samples within a subset was randomized for each of
the sessions. The raters had a break of approximately 10 min
between assessing the two subsets to avoid sensory adaptation,
when they were offered refreshments and asked to complete a
questionnaire regarding their age, health status, menstrual
cycle and partnership status.
Behav Ecol Sociobiol (2012) 66:569–581 573
Each rater washed her hands with non-perfumed soap
before the rating to avoid rating bias caused by hand
odor. They were told about the origin of the samples
(i.e. male axillary samples and the two essence sam-
ples), but they were not further informed about the aims
of the experiment.
The samples were rated on 7-point scales for their (1)
intensity, (2) pleasantness, and (3) attractiveness. Both ends
of each scale were anchored by verbal descriptions (e.g.
very unpleasant to very pleasant). For explorative purposes,
we also obtained ratings on masculinity. However, some
raters found judging this variable difficult. According to
our preliminary analyses, ratings of masculinity were in
general positively correlated with intensity ratings and
did not show any systematic trend in relation to shaving
status. For these reasons, results on masculinity are not
included here. The ratings were written down immedi-
ately after sniffing each stimulus, but the time spent by
sniffing was not restricted. Raters were also asked to
note whether they detected any contamination by smoke
or perfume.
The overall rating procedure was same across all experi-
ments. The only difference was that in Experiments III and
IV, we used a more sensitive test paradigm (the equivalent of
a forced-choice test). Within each subset, the samples were
paired. Each pair consisted of samples acquired from the left
and right armpit of a particular donor, and raters were
instructed not to use the same value within each pair (sam-
ples of one person) for any of the assessed variables (e.g.
pleasantness). This procedure is designed to detect subtle
effects as it generally strengthens the differences between
the tested groups.
Statistical analysis
The statistical package Statistica 7.1 was used for all anal-
yses. The data adhered to the requirements for parametric
tests. As our design was a within-subject one, we used a
paired ttest to test the effect of treatment within a session
and a repeated measures analysis of variance (ANOVA) to
test the dynamics of hair growth across individual sessions.
Our aim was to test possible changes in the perception of
axillary odor of shaved and unshaved armpits; therefore, we
used mean subject (raters’) ratings as the unit of the analysis.
We used post hoc tests (Fisher LSD test) only when the main
effect of the ANOVA was significant. To assess any possible
effect of the menstrual cycle in Experiment II,wealso
included a binary factor (fertile/non-fertile phase, for details
on criteria see section “Subjects”) as an independent vari-
able within a repeated measures ANOVA. We did not use
Bonferroni adjustments for multiple tests in order not to
decrease the statistical power of our relatively small sample
sizes (Nakagawa 2004). Further, we tested specific
hypotheses concerning the effect of shaving (in contrast to
the universal null hypothesis). In such cases, adjustments
are not recommended as they inflate the chance of type II
error (Perneger 1998).
We also analysed our data with mean donor ratings as the
unit of the analysis (results are not shown). The results
showed similar trends as with raters as the unit of the
analysis, but were not statistically significant and are not
reported further. Correlations among rated variables were
computed by using Pearson correlations for each session in
all four experiments.
Results
Experiment I
In Session 1, we found a significant difference between
the odors of shaved and unshaved armpits of donor
group N. The axillary odor of shaved armpits was rated
as more pleasant (t
27
02.80; p00.009), more attractive
(t
27
02.28, p00.03) and less intense (t
27
06.55; p<0.001)
than the odor of unshaved armpits. No significant dif-
ference was found between the axillary odors from
unshaved armpits and armpits with regrowing hair in
sessions 2, 3 and 4 (all pvalues> 0.08) (Table 2). To
test the effect of axillary hair growth on odor percep-
tion, we compared the ratings from all four sessions by
a repeated measures ANOVA. The ratings of unshaved
axillae served as control. We did not find a significant effect of
armpit treatment, repeated measure or any interaction with
pleasantness (all pvalues>0.055), or attractiveness (all pval-
ues>0.11). There was a significant repeated measure effect of
intensity (F
3,36
02.70; p00.05), but no interaction with armpit
treatment. Post hoc analyses showed that the odor samples
from Session 3 were rated significantly more intense compared
to Session 1 (p00.03) and Session 2 (p00.01).
Each person from group S shaved both armpits the
day before the first sampling session. We found no
significant differences between the ratings of the left
and right shaved armpits in Session 1, nor in pleasant-
ness (all pvalues>0.07), attractiveness (all pvalues >
0.09) or intensity (all pvalues> 0.12) between repeatedly
shaved armpits and armpits with regrowing hair in the other
three sessions (Table 2). To test the effect of axillary hair
growth during all four sessions in group S, we again
performed a repeated measures ANOVA. We did not
find a significant effect of armpit treatment, repeated
measure nor any interaction with pleasantness (all p
values > 0.24), attractiveness (all pvalues>0.21) or intensity
(all pvalues>0.07).
To test possible fluctuations in the raters’odor preferen-
ces, we included two non-human samples (castoreum,
574 Behav Ecol Sociobiol (2012) 66:569–581
cinnamon) in each testing session. We performed a repeated
measures ANOVA for castoreum ratings and did not find a
significant effect for any of the tested characteristics (e.g.
pleasantness) (all pvalues>0.54). For cinnamon, we found
a significant repeated measure effect for pleasantness
(F
3,18
04.10; p00.01) and attractiveness (F
3,18
04.11;
p00.005). Based on post hoc analyses in Session 1,
the cinnamon essence was rated as more pleasant (p00.05)
and more attractive (p00.008) than in Session 2. In Session 3,
it was rated as less pleasant (p00.002) and less attrac-
tive (p00.003) compared to Session 1. The cinnamon
samples from Session 4 were rated as more pleasant
(p00.02) than samples from Session 3 and less attrac-
tive (p00.002) than in Session 1. No significant
changes in ratings of intensity of the cinnamon samples
were found.
Experiment II
In Sessions 1 and 2, we found no significant difference
between the ratings of shaved and unshaved armpit odors
from donor group N (all pvalues>0.18). In Session 3, the
axillary odor of the unshaved armpits was rated as more
pleasant (t
23
02.37; p00.03) and less intense (t
23
02.95;
p00.007) than the odor of the armpits with regrowing
hair. On the contrary, in Session 4, the axillary odor of
the unshaved armpits was rated as less pleasant (t
17
0
2.76; p00.01) and more intense (t
17
08.38; p<0.001)
than the odor of the armpits with regrowing hair (Table 3).
When controlling for the raters’menstrual cycle phase,
qualitatively identical results were obtained.
To test the effect of axillary hair growth on odor ratings,
we compared the ratings from all four sessions. Ratings of
the unshaved axillae served as a control. We did not find a
significant effect of armpit treatment, repeated measure or
any interaction with pleasantness (all pvalues > 0.31) and
attractiveness (all pvalues> 0.37). A significant repeated
measure effect (F
3,22
08.45; p<0.001) and a significant inter-
action between armpit treatment and repeated measure for
intensity (F
3,22
04.01; p00.01) was found. Based on post
hoc analyses, we found that in Session 4, the samples of
unshaved armpits were judged as significantly more intense
compared to Session 1 (p<0.001), Session 2 (p<0.001) and
Session 3 (p<0.001). The other armpit of each donor was
shaved for Session 1, and then the hair was left to regrow (i.e.
in Sessions 2, 3 and 4). Further, changes in ratings of odor
intensity were found. The samples of armpits with regrowing
axillary hair from Session 2 were rated as less intense than the
samples of the same armpits from Session 3 (p00.02) and
Session 4 (p00.008).
To test possible differences between the ratings of axil-
lary odors of the right and left armpit, each person from
group S shaved both armpits before the onset of the
Table 2 Mean (standard deviations) ratings of pleasantness, attractiveness and intensity in Experiment I
Session 1 Session 2 Session 3 Session 4
Group N U S U R U R U R
Pleasantness 3.235 (0.965) 3.532 (0.957) 3.072 (0.944) 3.190 (0.898) 3.380 (0.808) 3.331 (0.908) 3.182 (1.155) 3.233 (1.104)
Attractiveness 3.216 (0.973) 3.480 (1.046) 3.090 (0.982) 3.147 (1.005) 3.277 (0.809) 3.163 (0.908) 3.125 (1.143) 3.373 (1.040)
Intensity 4.255 (1.050) 3.589 (1.190) 4.096 (0.966) 3.864 (1.156) 3.954 (1.118) 3.913 (1.173) 3.927 (1.042) 3.811 (1.078)
Group S S S S R S R S R
Pleasantness 3.649 (0947) 3.526 (1.008) 3.647 (0.839) 3.365 (0.941) 3.250 (0.892) 3.356 (1.218) 3.528 (1.032) 3.593 (0.921)
Attractiveness 3.469 (0.983) 3.354 (0.981) 3.540 (0.775) 3.279 (1.039) 3.190 (0.907) 3.235 (1.186) 3.528 (1.016) 3.500 (0.932)
Intensity 3.630 (1.159) 3.849 (0.995) 3.523 (1.090) 3.842 (1.132) 4.298 (1.011) 4.377 (1.286) 3.590 (1.163) 3.624 (1.324)
Group N indicates donors who had never shaved their armpits before and group S donors who regularly shaved one of their armpits across the four experimental sessions. Values in bold significantly
differ at P<0.05. The ratings are based on the judges using hormonal contraception
Umean values in the unshaved armpits, Rmean values in the armpits with regrowing hair, Smean values in the shaved armpits
Behav Ecol Sociobiol (2012) 66:569–581 575
experiment. First we compared ratings separately for each of
the four sessions. The only significant differences found
were that in Session 4, where the odor of the repeatedly
shaved armpits was rated as more attractive (t
17
03.07;
p00.007) than the odor of the armpits with hair regrowth.
Again when controlling for the raters’menstrual cycle phase,
qualitatively identical results were obtained.
Subsequently, we tested the effect of axillary hair growth
during all four sessions in group S. We did not find any
significant effect of armpit treatment nor any significant
interaction between armpit treatment and repeated measure
for pleasantness, attractiveness and intensity (all pvalues >
0.35). However, there was a significant repeated measure
effect for pleasantness (F
3,22
06.33; p00.001), attractiveness
(F
3,22
08.41; p<0.001) and intensity (F
3,22
012.06; p<0.001).
Subsequent post hoc analyses showed that odor samples from
Session 2 were rated significantly less pleasant (all pvalues<
0.01), less attractive (all pvalues<0.007) and more intense (all
pvalues<0.02) compared to Sessions 1, 3 and 4,and that odor
samples from Session 4 were rated significantly more intense
compared to Session 1 (p00.002). We also performed a
repeated measures ANOVA for castoreum and cinnamon
ratings and did not find a significant effect for any of
the tested characteristics (all pvalues> 0.09).
Experiment III
We found no significant differences in ratings of pleasant-
ness, attractiveness or intensity between the odors of the
shaved (or the armpits with regrowing hair) and unshaved
armpits in any of the three sessions (all pvalues > 0.47)
(Table 4). To test the effect of axillary hair growth on odor,
we compared the ratings from all three sessions. We did not
find any significant effect of armpit treatment, repeated
measure or any interaction with any of the rated character-
istics (all pvalues>0.15).
For castoreum ratings, we found a significant repeated
measure effect for pleasantness (F
3,18
03.83; p00.04). In
Session 2, the castoreum sample was rated as less pleasant
compared to Session 1 (p00.03) and Session 3 (p00.02).
For cinnamon, we did not find a significant effect for any of
the tested characteristics (all pvalues>0.21).
Experiment IV
Each donor had regularly shaved both armpits before the
onset of the experiment. In Session 1, we found no signif-
icant difference between the axillary odors from the right
and left shaved armpits (all pvalues> 0.06) (Table 5). No
significant differences in pleasantness or intensity of the
axillary odors were found between the repeatedly shaved
armpits and armpits with regrowing hair in Sessions 2 and 3
(all pvalues>0.10). In Session 3, the odor of the repeatedly
Table 3 Mean (standard deviations) ratings of pleasantness, attractiveness and intensity in Experiment II
Session 1 Session 2 Session 3 Session 4
Group N U S U R U R U R
Pleasantness 3.359 (1.149) 3.493 (0.955) 3.434 (1.243) 3.557 (1.118) 3.490 (1.062) 3.220 (1.030) 3.198 (1.245) 3.559 (1.081)
Attractiveness 3.347 (1.057) 3.318 (1.112) 3.320 (1.263) 3.478 (1.035) 3.579 (1.054) 3.437 (1.066) 3.269 (1.292) 3.473 (1.148)
Intensity 4.008 (1.028) 4.161 (0.766) 4.070 (1.173) 3.763 (1.148) 3.955 (0.964) 4.432 (1.069) 5.261 (1.129) 4.550 (1.201)
Group S S S S R S R S R
Pleasantness 3.985 (1.045) 3.896 (1.036) 3.087 (0.963) 3.108 (1.228) 4.044 (0.999) 3.748 (1.079) 4.068 (1.235) 3.657 (1.220)
Attractiveness 3.717 (1.148) 3.694 (1.125) 3.016 (1.088) 2.855 (1.205) 3.869 (0.931) 3.693 (0.873) 3.975 (1.166) 3.381 (1.032)
Intensity 3.589 (0.869) 3.665 (0.796) 4.913 (0.906) 5.129 (905) 3.587 (0.969) 3.414 (1.061) 4.142 (1.210) 4.233 (1.170)
Group N indicates donors who had never shaved their armpits before and group S donors who regularly shaved one of their armpits across the four experimental sessions. Values in bold significantly
differ at P<0.05. The ratings are based on the judges not using hormonal contraception
Umean values in the unshaved armpits, Rmean values in the armpits with regrowing hair, Smean values in the shaved armpits
576 Behav Ecol Sociobiol (2012) 66:569–581
shaved armpit was rated to be more attractive (t
16
03.13;
p00.006) than the odor of armpits with growing hair.
Subsequently, we tested the effect of axillary hair growth
during all three sessions. There was a significant repeated
measure effect for pleasantness (F
2,28
06.09; p<0.01),
attractiveness (F
2,28
04.48; p00.02) and intensity (F
2,28
0
5.58; p00.006), but no interaction with armpit treatment.
Odor samples from Session 1 were rated less pleasant com-
pared to Session 2 and Session 3. The odor samples from
Session 1 were rated less attractive compared to Session 3, and
odor samples from Session 1 and Session 2 were rated more
intense compared to Session 3. Repeated measures ANOVA
for castoreum and cinnamon ratings did not reveal any signif-
icant effect for any of the tested characteristics (all pvalues>
0.07), suggesting that there were no systematic fluctuations in
the raters’odor preferences.
Additional analyses
To explore relations between rated characteristics, we
carried out correlational analyses between individual
variablesforeachsession.Wefoundhighlysignificant
positive correlations between attractiveness and pleasantness
in all four experiments (all r00.81–0.92). Conversely,
negative significant correlations between intensity and attrac-
tiveness (all r00.13–0.57) or pleasantness (all r00.28–0.63)
were significant (p<0.05) in all but one case.
Some of the differences between shaved and unshaved
armpits could be attributed to the particularly intense odor of
unshaved armpits in some individuals. To check this possi-
bility, we compared the ratings of the unshaved armpits
between individual experiments. The ratings of intensity in
Experiment Icompared to other experiments were higher,
though not significantly so.
We further tested whether the proportion of raters
having previous experience with the odor studies varied
across individual experiments. The proportion of such
raters was 25%, 20%, 29% and 25% in Experiments I,
II,III and IV, respectively. This factor could potentially
explain the discrepancies between the experiments (see
“Discussion”for details). However, we did not find
significant differences.
Discussion
The main aim of this study was to test whether armpit
shaving and subsequent hair growth influences the subjec-
tive perception of quality and/or intensity of axillary odor.
We used two different approaches by comparing (1) ratings
of axillary odor of one-shot shaved armpits and the
unshaved armpits of the same donors who had never shaved
their armpits before (Experiments Iand II (group N) and
Experiment III) and (2) ratings of the odor of regularly
shaved axilla and axilla with regrowing hair (Experiments
Iand II (group S) and Experiment IV). The dynamics of
odor development during hair growth was tested for 6 weeks
in Experiments Iand II, and for 10 weeks in Experiments III
and IV.
In a series of experiments, we demonstrated that axillary
hair grooming affects the perception of odor intensity, pleas-
antness and attractiveness. In general, our results show that
Table 4 Mean (standard deviation) ratings of pleasantness, attractiveness and intensity in Experiment III
Session 1 Session 2 Session 3
U SURUR
Pleasantness 3.484 (0.552) 3.419 (0.552) 3.521 (0.614) 3.591 (0.600) 3.559 (0.713) 3.594 (0.609)
Attractiveness 3.331 (0.671) 3.390 (0.519) 3.387 (0.678) 3.449 (0.595) 3.487 (0.807) 3.570 (0.558)
Intensity 4.086 (0.679) 4.066 (0.533) 3.873 (0.584) 3.868 (0.704) 3.736 (0.840) 3.658 (0.705)
No significant differences at P<0.05 were found. The ratings are based on the judges using hormonal contraception
Umean values in the unshaved armpits, Rmean values in the armpits with regrowing hair, Smean values in the shaved armpits
Table 5 Mean (standard deviations) ratings of pleasantness, attractiveness and intensity in Experiment IV
Session 1 Session 2 Session 3
SSSRSR
Pleasantness 3.162 (0.865) 3.180 (0.786) 3.403 (0.669) 3.540 (0.682) 3.580 (0.739) 3.451 (0.867)
Attractiveness 3.137 (0.873) 3.167 (0.915) 3.394 (0.800) 3.441 (0.745) 3.645 (0.803) 3.343 (0.906)
Intensity 4.472 (0.718) 4.253 (0.734) 4.501 (0.536) 4.459 (0.567) 4.084 (0.701) 4.181 (0.548)
Values in bold significantly differ at P<0.05. The ratings are based on the judges using hormonal contraception
Rmean values in the armpits with regrowing hair, Smean values in the shaved armpits
Behav Ecol Sociobiol (2012) 66:569–581 577
the axillary odor of shaved armpits is rated as more pleasant,
more attractive and less intense compared to the unshaved
armpits of the same individual. However, the magnitude of the
observed effect is presumably not very high as we found
differences between the shaved and unshaved armpits only
in Experiment Ibut not in Experiments II and III.Moreover,
no differences were found between regularly shaved armpits
(group S) and armpits where hair had been regrowing for 1 or
3weeks.Onlyafter6(ExperimentII)or10(ExperimentIV)
weeks of hair regrowth were unshaved armpits judged to
smell less attractive than shaved armpits. The lack of the effect
in Experiment III and partially in Experiment IV is particularly
striking as we used the more sensitive force choice paradigm
which strengthens the differences between the tested groups.
Effect of one-shot shaving
Significant differences found in Experiment I(group N) but
not in Experiments II and III may be due to differences in
the donor’s odor intensity between the experiments. It is
possible that by chance the donors in Experiment Ihad
stronger axillary odors compared to the other experiments.
The higher intensity of the unshaved armpit may have resulted
in higher differences compared to the shaved armpit. Howev-
er, comparison of intensity of unshaved armpits across the
experiments brings only limited support for this claim.
The negative results in group N obtained in Experiments
II and III can be attributed to several factors. For instance,
the raters participating in a human odor study for the first
time might judge most of the samples rather negatively
which in turn may obscure the differences between odors
of shaved and unshaved axillae. This could result in lower
variability in using our 7-point scale and lead to the so-
called floor effect. Less intense and thus also more positive
(those two scales are usually negatively correlated) ratings
of previously experienced stimuli are a widely recognized
phenomenon (e.g. O’Connell et al. 1994). However, the
proportion of raters having previous experience with the
odor studies did not vary between particular experiments.
The effect of one-shot shaving observed in Experiment I
was only transient. One week after shaving the axillary hair,
we did not find any significant differences compared to the
odor of the unshaved armpit. Further, no significant differ-
ences in pleasantness or intensity were found between the
axillary odors from the unshaved armpits and the armpits
with regrowing hair for a period of up to 10 weeks in
Experiments Iand III. These results are in accord with the
study of Shelley et al. (1953). They indicate that the axillary
odor is not present for 24 h after armpit shaving, but is again
perceivable after 48 h. However, the raters in the study of
Shelley et al. indicated only whether they could smell any
odor at all and did not judge its strength or pleasantness.
Moreover, from the method description, it is not clear
whether the raters could see the targets while smelling their
armpits, a factor which might have biased the results.
Effect of regular shaving
The next aim of our study was to compare the odor of a
regularly shaved armpit and the odor of an armpit with
regrowing hair (group S). To our knowledge, this effect
has not been studied previously. To avoid possible odor
instability due to changes in microbial colonization in freshly
shaved axilla we chose male donors who had been shaving
their armpits regularly for at least 1 year before the beginning
of the study.
We found no significant difference in pleasantness, attrac-
tiveness and intensity of the axillary odors between repeatedly
shaved armpits and armpits with 1 or 3 weeks of hair
regrowth. On the other hand, the odor of the repeatedly
shaved armpits was rated as more attractive than the odor of
armpits with 6 (Experiment II)or10(ExperimentIV)weeks
of hair regrowth. The results suggest that only relatively long
hair causes differences perceived by humans. Anthropomet-
rical observations state 0.9–1.0cm per month axillary hair
growth in Caucasian populations with minor individual and
seasonal variability (Martin and Saller 1961). Although we
did not measure hair length, we can estimate that it was about
1.5 cm after 6 or 2.5 cm after 10 weeks. Alternatively,
a longer time may be needed to re-establish different
bacterial population following long-term shaving com-
pared with one-shot shaving.
The odor quality and intensity are probably not influ-
enced only by the presence or removal of hair but also by
changes in or damage to the skin due to shaving. Regular
axillary shaving causes graver axillary skin damage than
one-shot shaving (Marti et al. 2003). The modification of
the skin surface in the armpit area may cause changes in the
composition of axillary microflora and subsequently it may
modify the intensity and/or quality of the axillary odor.
Whether such changes can be perceived by human subjects
should be addressed in future studies by employing subjects
with one regularly shaved axilla and the other shaved singly.
Moreover, in our study, we did not test the differences
between the odor ratings of the unshaved and regularly
shaved armpit. Nevertheless, as we found differences between
the ratings of the regularly shaved armpit and the armpit with
hair growing for 6 and subsequently 10 weeks and also
between the one-shot shaved axilla and unshaved axilla, we
may expect similar or greater differences between the
unshaved and regularly shaved armpits.
Effect of hair growth
The influence of hair growth dynamics on body odor was
tested by comparing the ratings across individual sessions.
578 Behav Ecol Sociobiol (2012) 66:569–581
We employed the same protocol in all testing sessions to
make conditions as standard as possible. It included restric-
tions in the consumption of certain food and alcohol and in
activities, the same sampling length, ratings attended at
similar time of day and so on. However, it is important to
note that there were still a number of external factors which
could not be controlled and therefore the results of between
session analyses should be interpreted with caution. One
such intervening factor is the environmental temperature
and humidity during the sampling sessions. Temperature
positively influences intensity of perspiration and subse-
quently increases the humidity in the axillary area. Increase
in temperature and humidity results in higher growth of a
number of skin microflora (Hartmann 1983) as well as in the
rate of colonization by axillary microorganisms which
changes the intensity of axillary odor (Hopwood et al.
2005). In Experiments Iand II, there was a relatively high
increase in the environmental temperature (cca 10°C) between
sampling Sessions 2 and 3 (see Table 1). This might be the
reason why the odor samples(provided by men from group N)
in Experiment Iwere rated more intense in Session 3 com-
pared to Sessions 1 and 2. Another example of the possible
environmental temperature influences might be the ratings in
Experiment II. For instance, odor samples (provided by men
from group S) from Session 4 (outdoor temperature, 15.1°C)
were rated more intense compared to Session 1 (outdoor
temperature, 3.8°C).
Yet another factor which might have a systematic effect
on between-session comparisons is seasonal variation in
eating habits. As mentioned above, we tried to control the
diet components presumed to influence body odor. However,
the effect of the diet is still a largely unexplored field (for a
review, see Havlicek and Saxton 2009), and for instance the
amount of meat consumed, which is known to influence the
quality and intensity of body odor (Havlicek and Lenochova
2006), might have varied across the sessions. Other important
factors, which influence the results in olfactory studies, are
changes in mental condition (e.g. in mood) and subsequent
changes in the olfactory preferences of raters. Also, possible
habituation to the samples might influence the results. The
results of the longitudinal part of our study suggest that the
aforementioned factors probably had the main influence on
changes in ratings of odor samples and as already mentioned
above, the between session analyses should be interpreted
with caution.
Other confounds and implications
Some of the results may be further attributed to higher noise
caused by the fluctuations in both odor sensitivity, peaking
around ovulation and preferences across the rater’smenstrual
cycle (Doty et al. 1981;Hummeletal.1991;Grammer1993;
Pause et al. 1996). In Experiment II, where the raters did not
use hormonal contraception, we unexpectedly found that the
odor of the unshaved armpits was judged significantly more
pleasant and less intense compared to the armpits with
regrowing hair in Session 3. To check for this con-
founding factor, we included the data on menstrual
cycle phase in the analyses. However, we did not find
any effect of the cycle. Our further inspection of possi-
ble confounding factors (e.g. the donors’physical activ-
ities) also did not reveal any potential reason for the
observed effect. In view of the negative results in Ses-
sion 1 and the surprising result of Session 3, the overall
picture in Experiment II is mixed. To increase the
sensitivity of our experiments, we recruited only women
using hormonal contraception in three other experi-
ments. In contrast to non-pill users, olfactory functions
in women taking contraceptive pills are expected to be
relatively stable (Caruso et al. 2001, although see Doty
et al. 1981 for different results). Analyses including
menstrual cycle data in raters using hormonal contra-
ception again did not affect the results (data not shown).
Another possible contributing factor might be the differ-
ences between axillary odor of the right and left armpit
caused by the differences in the local microflora. It is not
yet clear from the literature whether the left and right armpit
microflora differs (Hopwood et al. 2005) or not (Leyden et
al. 1981; Rennie et al. 1991). The side of armpit which was
shaved was assigned randomly, although it is possible that
the majority of donors shaved the stronger axilla, particu-
larly given the small sample size, and this could obscure the
results. As laterality may influence the intensity of the odor
produced in each armpit, we checked our subjects’handed-
ness, but did not find any difference among the individual
experiments. Further, the comparison of the left and right
shaved armpits (Session 1, group S in Experiments I,II and
Experiment IV) and unshaved armpits between individual
sessions in Experiments I,II and III (group N) showed its
relative stability. Contrary to expectation, we found signif-
icant fluctuations in the intensity ratings among axillary
odors of the unshaved armpits in Experiment II. However,
in general, these comparisons support the assumption of
similarity between both armpits and justify our within-
subject design. Thus, the differences we found in particular
sessions were probably not caused by distinctions between
the axillary odor of the right and left armpits. This is in
agreement with a recently published study which in general
suggests no perceptional variation between armpit sides
(Ferdenzi et al. 2009).
In our study as well as in Shelley et al. (1953), the effect
of shaving was tested only in men. We chose this approach
to avoid body odor fluctuations across the menstrual cycle
(e.g. Kuukasjärvi et al. 2004; Havlicek et al. 2006). It is
reported in most studies that adult women’s body odor is
generally weaker than men’s (e.g. Hold and Schleidt 1977).
Behav Ecol Sociobiol (2012) 66:569–581 579
The magnitude of the observed effect of shaving is rather
small and therefore it is a matter of debate whether conducting
the same experiments with female donors would lead to
similar conclusions. As in Western cultural settings, axillary
hair removal is much more widespread in women (Tiggemann
and Kenyon 1998; Tiggemann and Hodgson 2008)itshould
be tested experimentally, too.
As a consequence of this study, we suggest that shaving
habits should be considered in human odor studies. This is a
particularly important issue in studies employing a between-
subjects design on individuals who might systematically
differ in their grooming habits. For instance, Martins et al.
(2005) tested the odor attractiveness of male and female
homosexuals compared to heterosexual men and women.
The number of individuals shaving their armpits, a habit
more widespread in women and in the gay community, was
not reported.
The differences observed between shaved armpits and
armpits with axillary hair might be in fact far greater than
our results show. Cotton pads worn by donors in their
armpits represent a mechanical barrier for the retention of
chemical compounds and therefore this treatment might
partly substitute the assumed function of axillary hair
(i.e. retention of volatile compounds). In our previous studies
(e.g. Havlicek et al. 2006), we advocated the use of armpit
pads instead of T-shirts as media for body odor collection as
with T-shirts, the source of the odor cannot be specified. It is
possible that for testing the effect of armpit hair the T-shirt
method would be more suitable. At this point this issue
remains only speculative as the validity of methods used in
human chemical ecology has scarcely been tested (for a full
discussion see, Lenochova et al. 2009).
In sum, our study shows that axillary hair shaving has an
impact on body odor quality and intensity. However, the
effect is relatively weak and was not observed in all experi-
ments and also not when the hair was relatively short. One-
shot shaving seems to have a relatively short-term effect as
the development of an odor similar to unshaved axilla in
one-shot shaved axilla was observed as early as after 1 week
of hair regrowth. On the other hand, odor development of
regularly shaved axilla after regrowth seems to be more
profound as it differed from shaved axilla after 6 weeks.
Our results are, in general, consistent with the idea that
axillary hair developed for the retention of chemicals which
may serve in chemical communication (Cohn 1994)and
contribute to our emerging understanding of the complex
nature of human chemical ecology.
Acknowledgements We wish to thank all volunteers for their partic-
ipation in the study, Jindra Havlickova, Robert Martin, Tamsin K.
Saxton and Jarka Valentova for the valuable advice and language
corrections, Pavlina Lenochova for helping with data collection and
Aroma Corp. Decin for providing samples of essences. The study was
supported by Grant Agency of Charles University (GAUK 57010),
Grant Agency of Czech Republic (GACR 406/09/0647) and Czech
Ministry of Education grant 0021620843.
Ethical standards This study complied with the current laws of the
Czech Republic.
Conflict of interest The authors declare that they have no conflict of
interest.
References
Caruso S, Grillo C, Agnello C, Maiolino L, Intelisano G, Serra A
(2001) A prospective study evidencing rhinomanometric and
olfactometric outcomes in women taking oral contraceptives.
Hum Reprod 16:2288–2294
Cohn BA (1994) In search of human skin pheromones. Arch Dermatol
130:1048–1051
Comfort A (1971) Likelihood of human pheromones. Nature 230:432–
433
Cowley JJ, Brooksbank BW (1991) Human exposure to putative
pheromones and changes in aspects of social behaviour. J Steroid
Biochem Mol Biol 39:647–659
Doty RL, Snyder PJ, Huggins GR, Lowry LD (1981) Endocrine,
cardiovascular and psychological correlates of olfactory sensitiv-
ity changes during the human menstrual cycle. J Comp Phys
Psychol 98:45–60
Ferdenzi C, Schaal B, Roberts SC (2009) Human axillary odor: are
there side-related perceptual differences? Chem Senses 34:565–
571
Gower DB, Holland KT, Mallet AI, Rennie PJ, Watkins WJ (1994)
Comparison of 16-androstene steroid concentrations in sterile
apocrine sweat and axillary secretions—interconversions of 16-
androstenes by the axillary microflora—a mechanism for axillary
odor production in man. J Steroid Biochem Mol Biol 48:409–418
Grammer K (1993) 5-α-androst-16en-3α-on: a male pheromone? A
brief report. Ethol Sociobiol 14:201–208
Hartmann AA (1983) Effect of occlusion on resident flora, skin-
moisture and skin-pH. Arch Dermatol Res 275:251–254
Havlicek J, Roberts SC, Flegr J (2005) Women’s preference for dom-
inant male odour: effects of menstrual cycle and relationship
status. Biol Lett 1:256–259
Havlicek J, Dvorakova R, Bartos L, Flegr J (2006) Non-advertized
does not mean concealed: body odor changes across the human
menstrual cycle. Ethology 112:81–90
Havlicek J, Lenochova P (2006) The effect of meat consumption on
body odor attractiveness. Chem Senses 31:747–752
Havlicek J, Lenochova P (2008) Environmental effects on human body
odour. In: Hurst JL, Beynon RJ, Roberts SC, Wyatt TD (eds)
Chemical signals in vertebrates XI. Springer, New York, pp 199–
212
Havlicek J, Roberts SC (2009) MHC related mate choice in humans: a
review. Psychoneuroendocrino 34:497–512
Havlicek J, Saxton T (2009) The effect of diet on human bodily odors.
In: Hasegawa K, Takahashi H (eds) New research on food habits.
Nova Science Publishers, New York, pp 35–44
Havlicek J, Murray AK, Saxton TK, Roberts SC (2010) Current issues in
the study of androstenes in human chemosignalling. In: Litwack G
(ed) Pheromones, vitamins & hormones, vol 83. Academic Press,
London, pp 47–81
Havlíček J, Lenochová P, Oberzaucher E, Grammer K, Roberts SC
(2011) Does length of sampling affects quality of body odor
samples? Chemosens Percept 4:186–194
580 Behav Ecol Sociobiol (2012) 66:569–581
Hays WST (2003) Human pheromones: have they been demonstrated?
Behav Ecol Sociobiol 54:89–97
Hold B, Schleidt M (1977) The importance of human odor in non-
verbal communication. Z Tierpsychol 43:225–238
Hopwood D, Farrar MD, Bojar RA, Holland KT (2005) Microbial
colonization dynamics of the axillae of an individual over an
extended period. Acta Derm Venereol 85:363–364
Hummel T, Gollisch R, Wildt G, Kobal G (1991) Changes in olfactory
perception during the menstrual cycle. Experientia 47:712–715
Jacob S, Garcia S, Hayreh D, McClintock MK (2002) Psychological
effects of musky compounds: comparison of androstadienone
with androstenol and muscone. Horm Behav 42:274–283
Kuukasjärvi S, Eriksson CJP, Koskela E, Mappes T, Nissinen K,
Rantala MJ (2004) Attractiveness of women’s body odors over
the menstrual cycle: the role of oral contraceptives and receiver
sex. Behav Ecol 15:579–584
Lenochova P, Roberts SC, Havlicek J (2009) Methods of human
body odour sampling: the effect of freezing. Chem Senses
34:127–138
Leyden JJ, McGinley KJ, Holzle E, Labows JN, Kligman AM (1981)
The microbiology of the human axilla and its relationship to
axillary odor. J Invest Dermatol 77:413–416
Lundstrom JN, Hummel T, Olsson MJ (2003) Individual differences in
sensitivity to the odor of 4,16-androstadien-3-one. Chem Senses
28:643–650
Marti VPJ, Lee RS, Moore AE, Paterson SE, Watkinson A, Rawlings
AV (2003) Effect of shaving on axillary stratum corneum. Int J
Cosmet Sci 25:193–198
Martin R, Saller K (1961) Lehrbuch der Anthropologie in System-
atischer Darstelung. Gustav Fisher Verlag, Sttutgart
Martins Y, Preti G, Crabtree CR, Vainius AA, Wysocki CJ (2005)
Preference for human body odors is influenced by gender and
sexual orientation. Psychol Sci 16:694–701
Nakagawa S (2004) A farewell to Bonferroni: the problems of
low statistical power and publication bias. Behav Ecol
15:1044–1045
Natsch A, Gfeller H, Gygax P, Schmid J, Acuna G (2003) A specific
bacterial aminoacylase cleaves odorant precursors secreted in the
human axilla. J Biol Chem 278:5718–5727
Natsch A, Derrer S, Flachsmann F, Schmid J (2006) A broad diversity
of volatile carboxylic acids, released by a bacterial aminoacylase
from axilla secretions, as candidate molecules for the determina-
tion of human-body odor type. Chem Biodivers 3:1–20
Nixon A, Mallet AI, Gower DB (1988) Simultaneous quantification of
5 odorous steroids (16-androstenes) in the axillary hair of men. J
Steroid Biochem 29:505–510
O’Connell RJ, Stevens DA, Zogby LM (1994) Individual-
differences in the perceived intensity and quality of specific
odors following self-adaptation and cross-adaptation. Chem
Senses 19:197–208
Pause BM, Sojka B, Krauel K, Fehm-Wolfsdorf G, Ferstl R (1996)
Olfactory information processing during the course of the men-
strual cycle. Biol Psychol 44:31–54
Pawlowski B (1999) Loss of oestrus and concealed ovulation in human
evolution—the case against the sexual-selection hypothesis. Curr
Anthropol 40:257–275
Pierce JD, Cohen AB, Ulrich PM (2004) Responsivity to two odorants,
androstenone and amyl acetate, and the affective impact of odors
on interpersonal relationships. J Comp Psychol 118:14–19
Perneger TV (1998) What’s wrong with Bonferroni adjustments. BMJ
316:1236–1238
Rennie PJ, Gower DB, Holland KT (1991) In vitro and in vivo studies
of human axillary odor and the cutaneous microflora. Br J Der-
matol 124:596–602
Roberts SC, Gosling LM, Carter V, Petrie M (2008) MHC-correlated
odour preferences in humans and the use of oral contraceptives.
Proc R Soc Lond B 275:2715–2722
Saxton TK, Lyndon A, Little AC, Roberts SC (2008) Evidence that
androstadienone, a putative human chemosignal, modulates wom-
en’s attributions of men’s attractiveness. Horm Behav 54:597–601
Sergeant MJT, Dickins TE, Davies MNO, Griffiths MD (2007) Women’s
hedonic ratings of body odor of heterosexual and homosexual men.
Arch Sex Behav 36:395–401
Shelley WB, Hurley HJ, Nichols AC (1953) Axillary odor; experimen-
tal study of the role of bacteria, apocrine sweat, and deodorants.
Arch Dermatol Syphiol 68:430–446
Tiggemann M, Kenyon SJ (1998) The hairlessness norm: the removal
of body hair in women. Sex Roles 39:873–885
Tiggemann M, Hodgson S (2008) The hairlessness norm extended:
reasons for and predictors of women’s body hair removal at
different body sites. Sex Roles 59:889–897
Wysocki CJ, Preti G (2004) Facts, fallacies, fears, and frustrations with
human pheromones. Anat Rec 281A:1201–1211
Zeng XN, Leyden JJ, Lawley HJ, Sawano K, Nohara I, Preti G (1991)
Analysis of characteristic odors from human male axillae. J Chem
Ecol 17:1469–1492
Behav Ecol Sociobiol (2012) 66:569–581 581