The term pheromone was introduced by Karlson and Lüscher in
1959 (ref. 1). They defined pheromones as “substances secreted to the
outside of an individual and received by a second individual of the
same species in which they release a specific reaction, for example, a
definite behaviour or developmental process”. Originally, two types of
pheromonal signalling were recognized. Releaser pheromones elicit
an immediate behavioural response — for example, 2-methylbut-2-
enal, which is produced in rabbit milk. This single molecule elicits
stereotyped nipple-search behaviour in rabbit pups, and is vital for
them to locate the nipples during the brief daily period of suckling2.
By contrast, primer pheromones mediate more slowly developing and
longer-lasting changes to endocrine state or development. For example,
α-farnesene is one of several testosterone-dependent constituents of
male mouse urine, which accelerate puberty of pre-pubertal female
As more examples of intraspecific chemical signalling have been iden-
tified, the original definition of a pheromone has appeared excessively
rigid, and the term has frequently been used for any form of innate,
intraspecific chemical communication4. Thus, many investigators also
recognize the category of signaller pheromones. These convey infor-
mation about the sender, such as individual or group identity, which
are important for parent–offspring recognition and mate choice5,6. A
further category — modulator pheromones — has been proposed to
affect mood and thought processes in humans, although this has yet to
gain wide acceptance7.
This review seeks to highlight the diverse chemical nature of vertebrate
pheromones and recent advances in our understanding of the chemo-
sensory systems and neural mechanisms underlying such pheromonal
effects. Although many of the principles discussed apply to a wide range
of vertebrates, this review concentrates on mammalian pheromones,
particularly those of rodents, in which most of the recent advances have
occurred. These advances have revealed that both volatile molecules
and non-volatile peptides and proteins can provide sex-specific and
individual-specific cues. The same chemosignals can be detected by
diverse chemosensory systems, which have access to the same behav-
ioural and physiological outputs. However, although major advances
have been made recently, our knowledge of the roles and mechanisms of
pheromonal communication in vertebrates is still fragmentary. Future
progress will require analysis at all levels, including the use of analytical
chemistry to characterize the often complex mixture of chemical signals,
Pheromonal communication in vertebrates
Peter A. Brennan1 & Frank Zufall2
Recent insights have revolutionized our understanding of the importance of chemical signals in influencing
vertebrate behaviour. Previously unknown families of pheromonal signals have been identified that
are expanding the traditional definition of a pheromone. Although previously regarded as functioning
independently, the main olfactory and vomeronasal systems have been found to have considerable overlap in
terms of the chemosignals they detect and the effects that they mediate. Studies using gene-targeted mice
have revealed an unexpected diversity of chemosensory systems and their underlying cellular and molecular
mechanisms. Future developments could show how the functions of the different chemosensory systems are
integrated to regulate innate and learned behavioural and physiological responses to pheromones.
1Department of Physiology, University of Bristol, Medical School Building, University Walk, Bristol BS8 1TD, UK. 2Department of Anatomy and Neurobiology, University of Maryland, School of
Medicine, 20 Penn Street, Baltimore, Maryland 21201, USA.
and molecular genetics to target specific receptor systems for phenotypic
analysis at the cellular and behavioural levels.
The diversity of vertebrate pheromonal signals
Pheromones come in a wide variety of chemical forms. Their most
significant features are their size and their polarity, which are the major
factors determining their volatility in air and solubility in water, respec-
tively. Thus, in the terrestrial environment, attractant and alarm phero-
mones, which by their nature act at a distance, are typically small and
volatile. Methylthiomethanethiol (MTMT), which is present in male
mouse urine and attracts investigation by females8, is a good example.
By contrast, pheromones that convey information about specific indi-
viduals are likely to be relatively non-volatile — for instance, proteins or
peptides — so that they do not disperse and can be more reliably associ-
ated with the producer. The situation differs in the aquatic environment,
where solubility is the most important factor and even relatively high-
molecular-weight peptides and proteins can have an attractant role.
For example, the decapeptide sodefrin is produced by the abdominal
gland of the male newt Cynops pyrrhogaster, and when released into
water it attracts females of the same, but not closely related, species9. As
discussed below, peptide and protein pheromones are also used widely
in mammalian chemical communication10–15. Other examples of aquatic
attractant pheromones include the mixture of sulphated steroids and
bile acids recently identified as migratory pheromones of lampreys16.
The structural constraints imposed by common metabolic pathways
mean that similar molecules may be used as pheromones by different
species. In these instances, the species specificity of the signal may be
maintained by a pheromone blend or by the context in which it occurs.
For example, the androgen derivatives 5α-androst-16-en-3-one and 5α-
androst-16-en-3-ol, which are found in boar saliva, act synergistically
as releaser pheromones to attract sows and elicit a receptive posture17.
Similarly, although MTMT increases the attractiveness of male urine to
females in the presence of other urinary cues, it is relatively ineffectual
when presented on its own8. Male mouse urine also contains high con-
centrations of the testosterone-dependent volatiles (R, R)-3,4-dehydro-
exo-brevicomin (DHB) and (S)-2-sec-butyl-4, 5-dihydrothiazole (SBT).
When sensed in the appropriate social context, or administered in urine
from castrated males, these compounds act as releaser pheromones,
eliciting aggression in male mice18. But pheromones do not always fit
neatly into a single category and, along with their releaser effects on
NATURE|Vol 444|16 November 2006|doi:10.1038/nature05404
male aggression, DHB and SBT also have primer effects, accelerating
puberty in prepubertal female mice and inducing and synchronizing
oestrus in adults3,18.
Multiple chemosensory systems mediate pheromonal effects
Mammals have several chemosensory systems capable of detecting
pheromones (Fig. 1). The main olfactory epithelium (MOE) consists
predominantly of ciliated olfactory sensory neurons (OSNs), which
project to the main olfactory bulb (MOB). Each of these expresses a
single receptor type from a large family — mice have almost 1,300 — of
olfactory receptors19. Recently, a second family of G-protein-coupled
receptors, the trace-amine-associated receptors (TAARs), which recog-
nize volatile amines present in urine, has also been shown to be expressed
In addition to the MOE, most mammals have a vomeronasal organ
(VNO)21,22. This is a blind-ended tube located in the nasal septum, and
contains vomeronasal sensory neurons (VSNs) with microvillar mor-
phology, which project to the accessory olfactory bulb (AOB). Two
classes of vomeronasal receptor have been identified, the V1rs and
the V2rs (ref. 22). They belong to the seven-transmembrane-domain
G-protein-coupled receptor superfamily, but share little homology
with each other or with the olfactory receptors, suggesting that they
have distant evolutionary origins. Gene targeting of a cluster of V1rs
has established that these molecules are involved in responses to small
organic phero mones such as 6-hydroxy-6-methyl-3-heptanone23, which
accelerates puberty in female mice18. No such information is yet avail-
able for V2rs.
In addition to their distinct morphology, VSNs have a different
transduction mechanism from OSNs, involving a diacylglycerol-acti-
vated cation channel, which depends in part on the transient-receptor-
potential channel 2 (Trpc2) gene24. Analysis of the mouse genome has
identified 137 functional receptors of the V1r class25 as well as 60 or so
potentially functional V2rs26. Although a similar set-up has been found
in other rodents, and in opossums, it is not typical of all mammals, let
alone all vertebrates. Therefore, as with most aspects of pheromonal
signalling, species differences must be borne in mind.
Notable evolutionary changes have occurred to chemosensory
systems in association with the transition from aquatic to terrestrial
environments. Fish do not have a VNO. Instead, they have a single
olfactory organ containing a mixture of ciliated sensory cells expressing
olfactory-receptor-like receptors and microvillar cells expressing both
V2r-like27 and V1r-like28 receptors. These microvillar sensory neurons
seem to have segregated into a separate organ in early terrestrial ver-
tebrates, at about the same time that the MOE was adapting to detect
airborne odorants. However, the detailed story is likely to be complex29,
and the division between cell types is not absolute. Although most
sensory cells in the mammalian MOE are ciliated and express olfactory
receptors, also present are microvillar cells that seem to form a distinct
The different locations and anatomical structures of the MOE and
VNO have consequences for stimulus access. Whereas the MOE has
access to stimuli in the nasal airstream, the VNO is connected to the
nasal cavity by a narrow duct21. Stimuli are thought to gain access to
the VNO by a vascular pumping mechanism that is activated in arous-
ing situations31. This has led to the idea that the main olfactory system
is specialized for detecting volatile, airborne molecules, whereas the
vomeronasal system is specialized for the detection of non-volatile
chemosignals, such as those in urine, skin, scent glands and reproduc-
tive secretions, after direct contact with the stimulus source. However,
recent findings indicate that this functional division is not as absolute
as was once thought.
VSNs can detect small organic molecules present in urine — for exam-
ple, the male pheromones SBT and DHB32–34 — that bind to carrier
proteins such as major urinary protein35 (MUP; see below). In vivo, such
molecules probably gain access to the VNO after direct contact with the
stimulus source, which is consistent with increases in neuronal activity in
the AOB after direct contact with an anaesthetized conspecific36. There is
some debate about whether the VNO can detect volatiles in the absence
of physical contact. Katz’s group failed to find neuronal responses in the
AOB to volatile odorants or pheromones presented on cotton swabs
without direct contact. However, relatively few cells were tested, and it
is not clear whether recordings were made from the anterior subregion
of the AOB, which receives projections from the volatile-sensing V1r
class of VSNs22. By contrast, functional magnetic resonance imaging
(fMRI) in anaesthetized mice showed robust changes in the activity
of the AOB in response to urine odours delivered via the airstream37.
These responses were predominantly localized to the anterior part of
the AOB, which is consistent with the projection of the volatile-sensing
VSNs to the anterior subregion. Finally, even the seemingly reasonable
hypothesis that the MOE responds only to airborne molecules will have
to be rethought, as non-volatile stimuli have been shown to reach the
MOE after direct contact with a stimulus10.
Mice with VNO ablation show a number of behavioural and physi-
ological deficits in their response to chemosensory cues. For instance,
male urinary chemosignals fail to elicit aggression in males with VNO
lesions, and such males do not show the normal rise in luteinizing hor-
mone (LH) levels in response to female chemosignals38. Genetic ablation
of Trpc2 markedly reduces VSN responses and leads to specific deficits
in male–male and maternal aggression39,40. However, its effects on male
sexual behaviour differ from those of physical ablation, which may be
explained in part by the recent finding that some peptide-sensing VSNs
are TRPC2-independent41 (Fig. 2).
These findings have focused attention on the VNO as a conveyer
of pheromonal responses. But it has long been known that many phe-
romonal responses are mediated by the main olfactory system. For
instance, the nipple-search pheromone that guides the suckling of
rabbit pups is still effective after VNO lesions42. Similarly, in sows, the
effects of androstenone and androstenol on sexual behaviour are unaf-
fected by VNO lesions17. Furthermore, male mouse urinary attractant
MTMT selectively excites neurons in the MOB, although whether its
attractant effects are mediated by the main olfactory pathway remains
to be seen8.
Genetic ablation of the cyclic-nucleotide-gated channel subunit
CNGA2, which is vital for transduction in most mouse OSNs, causes
behavioural deficits equally as marked as VNO lesions, including lack of
chemosensory investigation of conspecifics and impaired mating43. This
does not necessarily mean that all of the effects seen in Cnga2-knockout
Figure 1 | The mouse olfactory system. Scheme highlighting the
heterogeneity of chemosensory subsystems in the nose of the mouse
(sagittal view). Sensory cells in each subsystem express specific signal
detection molecules (not shown), which enables selective targeting of these
populations22,45,46. Chemosensory functions have not yet been established for
the guanylyl cyclase type-D system (GCD) and Grueneberg ganglion (GG).
NC, nasal cavity; NG, necklace glomeruli; SOM, septal organ of Masera.
NATURE|Vol 444|16 November 2006
animals are mediated solely by the main olfactory system. Main olfac-
tory signals have an important role in attracting attention to stimuli to
be investigated by the vomeronasal system44, and possibly in activation
of the vomeronasal pump. If this is the case, then some of the deficits
seen in Cnga2-knockout mice may result from indirect effects on VNO
function. Indeed, the results of genetic-ablation experiments must be
interpreted with care, as they can be confounded by the function of other
olfactory subsystems with as yet unknown roles in pheromone detection
(Fig. 1). For instance, a subpopulation of MOE sensory neurons expresses
a guanylyl cyclase-D signalling system rather than the canonical cyclic
AMP–CNGA2 system45. This chemosensory subsystem is not affected in
Cnga2-knockout mice. There are further chemosensory systems about
which little is known at present. The septal organ of Masera is an anatomi-
cally isolated area of main olfactory-like epithelium, situated ventrally to
the MOE, and, although it projects to the ventral region of the MOB, its
function is unclear46. The Grueneberg ganglion, which is situated at the
anterior end of the nasal cavity, is also likely to have a chemosensory role
— a prediction based on its expression of olfactory marker protein and
its projections to glomeruli in the dorsal region of the MOB47.
MHC-associated cues as individuality signals
It has long been recognized that behaviours such as mate choice and
parent–offspring interactions can be influenced by major histocom-
patibility (MHC) genotype. MHC molecules are encoded by a highly
polymorphic family of genes that determine immunological identity
at the tissue level. Hence, laboratory mice are more likely to mate with
individuals of dissimilar MHC genotype from their own48. Such dis-
assortative mating is found in semi-natural conditions in which fewer
MHC-homozygous offspring are born than would be expected from
random matings6. The maternal recognition of young is also affected
by MHC genotype. Maternal mice preferentially retrieve pups of the
same MHC type as themselves5. Similarly, in a Y-maze test, mouse pups
were found to prefer nest odours of maternal and sibling MHC type to
unfamiliar MHC type5. Furthermore, mice can be trained to discrimi-
nate between urine odours of congenic mice that differ genetically only
at their MHC H2 locus49, implying that MHC genotype influences the
volatile constituents of urine.
MHC class I proteins bind a vast range of nine-amino-acid peptides
that result from proteasomal degradation of endogenous and foreign
proteins, and present them at the cell surface for immune surveillance.
The ability of mice to discriminate urine odours of MHC-congenic
strains is consistent with the polymorphism in the peptide-binding
groove of the MHC protein50. Fragments of MHC class I proteins have
been found in mouse urine, albeit at low levels51. But how could they
influence urine odour? According to the carrier hypothesis, the peptide-
binding groove releases its peptide ligand when the MHC protein is
cleaved from the cell, and becomes available for binding small, volatile
molecules52. Polymorphic variations in the peptide-binding-groove
structure would determine the composition of volatiles bound by MHC
class I proteins, and therefore influence body odour50. Although there
is no direct evidence for the hypothesis that MHC class I proteins are
able to bind volatiles, MHC genotype has been shown to influence the
profile of volatile fatty acids in urine53. However, it should be noted that
although the volatile profile affects the discriminability of mouse urine,
MHC-dependent urinary volatiles have yet to be shown to convey indi-
viduality in a natural context.
More recent evidence has demonstrated that the MHC peptide ligands
themselves can function as chemosignals, forming a direct link between
individuality at the immunological and behavioural levels11,12. VSNs
of the V2r class respond highly sensitively to these nine-amino-acid
peptides12 (Fig. 2b). Both VSN responses and the specificity of peptide
binding to MHC class I proteins are determined by the position of what
are known as anchor residues. These are amino acids with bulky side
chains that fit into pockets in the peptide-binding cleft. A particular
MHC class I molecule can bind a variety of peptides that differ in amino-
acid composition but have anchor residues constrained to specific posi-
tions along the peptide chain. The identities and positions of the anchor
residues of an MHC peptide ligand therefore convey information about
the MHC class I protein that binds it. As the VSN responses show very
similar dependence on anchor-residue position to those of the MHC
class I proteins, they convey the MHC identity of the peptides. Indeed,
the addition of synthetically produced MHC peptide ligands has been
shown to influence the perceived strain identity of mouse urine, which
determines mate recognition in the context of the pregnancy-block
effect12 (see below).
Neither the responses of the V2r-expressing VSNs to MHC peptide
ligands (Fig. 2b) nor the effectiveness of these ligands in conveying
individuality in the pregnancy-block effect require a functional TRPC2
channel41. This implies that the peptide-sensitive VSNs rely on a differ-
ent transduction mechanism from the TRPC2-dependent transduction
used by the V1r class of VSNs. The V2r class of vomeronasal receptors
are structurally very different from the V1rs. V2rs have a large extracel-
lular amino-terminal domain22, which might be involved in binding
MHC peptide ligands. Intriguingly, V2rs are co-expressed with non-
classical MHC Ib proteins54,55, with which they might form a receptor
complex54. There are nine members of this non-classical MHC Ib family.
However, their binding groove is unlikely to bind peptides, and their
Figure 2 | Both the MOE and VNO detect pheromonal ligands at low
concentrations. a, b, Electrophysiological recordings from wild-type (WT)
and gene-targeted mice showing examples of field-potential responses
to the same pheromone ligands at low concentrations (10–10 M) in the
MOE and VNO. 2-Heptanone is a urinary constituent that has a primer
pheromonal effect, extending the length of female oestrous cycles18.
SYFPEITHI is a prototypical MHC class-I peptide ligand (for the H-2d
haplotype of BALB/c mice)12. Note that both MOE responses are absent
in mice with a targeted mutation in the cyclic-nucleotide-gated channel
subunit CNGA2. The response to 2-heptanone is absent in the VNO of
mice with a targeted mutation in the transient-receptor-potential channel
TRPC2; however, the response to SYFPEITHI is relatively normal in these
mutants. c–e, Large-scale confocal Ca2+ imaging with cellular resolution
using a nasal coronal tissue slice identifies discrete populations of peptide-
detecting olfactory sensory neurons in the mouse MOE. AAPDNRETF
is a prototypical MHC class-I peptide ligand for the H-2b haplotype of
C57BL/6 mice. A portion of the olfactory epithelium delimited by the
white box in c is shown at higher magnification in d and e. d, Confocal
fluorescence image of the fluo-4-loaded olfactory epithelium acquired at
rest (grayscale). e, Pseudocolour image of the relative increase in peptide-
induced Ca2+ fluorescence, showing four responsive OSNs. S, septal wall;
DR, dorsal recess; I, endoturbinate I. (Images a, c, d and e reproduced, with
permission, from ref. 10. Image b is based on data from ref. 41.)
NATURE|Vol 444|16 November 2006
role in MHC peptide recognition is unclear56. Certain combinations of
MHC Ib proteins are expressed with particular V2rs55, and it is tempting
to speculate that they might convey innate differences in vomeronasal
signalling between different MHC types.
Highly specific responses to MHC peptide ligands have recently been
recorded from the mouse MOE, although their dependence on anchor
residues differs from the responses of VSNs10 (Fig. 2a, e). Hence, MHC-
dependent signals can be sensed by both the main olfactory system and
the vomeronasal system. MHC peptide ligands have also been shown
to influence preferences of female sticklebacks in their choice of mate,
in a manner consistent with the generation of optimum MHC diversity
in their offspring57. Given the ubiquity of class I MHC proteins, the role
of their peptide ligands in signalling individuality may be widespread
Major urinary proteins and territorial behaviour in mice
A classic example of the use of signalling pheromones is scent mark-
ing. Individuals deposit scent cues around their environment as long-
lasting advertisements of their presence to competitors or potential
mates. This is perhaps best understood in mice, in which dominant
males ‘paint’ small urine marks throughout their territory, especially
at the boundaries and along major access routes44. Mice excrete large
amounts of protein in their urine, 99% of which are MUPs. These are
18–20 kDa members of the lipocalin family of ligand-binding proteins,
which are thought to have a chemosensory signalling role in a variety of
species and behavioural contexts58. MUPs bind volatile ligands including
the testosterone-dependent male mouse pheromones DHB, SBT, (E,E)-
α-farnesene, (E)-β-farnesene and 6-hydroxy-6-methyl-3-heptanone3.
Moreover, the concentration of MUPs is four to five times higher in
male than female mouse urine, and some MUP variants are found only
in males59. The release of volatiles from urine marks is prolonged by
their binding to MUPs and serves to attract direct investigation of the
non-volatile chemosignals. As the ligands are released over a period
of time, they also convey the freshness of urine marks, which allows
a male to advertise his territorial dominance and competitive ability
to potential mates, without the potentially damaging consequences of
direct confrontation with competitors44.
Male mice countermark urine marks that have been left by an intruder,
a behaviour that is driven by non-volatile components of urine (presum-
ably MUPs) and is likely to be mediated by the VNO44. An individual
mouse in the wild produces a specific pattern of 5–15 variants from the
polymorphic MUP family60, and the rate of countermarking of urine
marks is associated with their MUP profile13. The differences in amino-
acid residues among MUP variants are largely localized to the surface
of the protein, and, although some are in the ligand-binding calyx, they
have so far been found to have limited influence on the relative binding
of their volatile ligands35. It remains to be determined whether MUPs
interact directly with receptors, although recent evidence suggests that
MUPs that have been stripped of their ligands have pheromonal activity
in influencing ovulation14.
Neural coding of pheromonal information
The vomeronasal system is often thought of as a labelled-line system: a
high degree of receptor-cell stimulus specificity is maintained in their
neural projections to the central structures that control the behavioural
or neuroendocrine output. Certainly, V1r-expressing VSNs seem to be
good candidates for such a system, as they respond highly selectively.
Moreover, their specificity does not broaden as the stimulus concentra-
tion is increased, as is typically found in OSNs32. However, the responses
to MHC peptide ligands of V2r VSNs to is likely to be considerably more
complicated than a ‘one ligand–one receptor–one cell’ coding strategy61.
Individual VSNs are frequently found to respond to MHC peptide lig-
ands from more than one MHC type12. It remains to be established
whether the coding of MHC individuality by VSNs relies on a pattern
recognition system, or whether the responses of individual VSNs to
particular combinations of MHC peptides are representative of specific
combinations of MHC alleles in a labelled-line manner.
Neurons in the AOB respond to specific strain and sex combinations
during direct contact with the head and anogenital regions of an anaes-
thetized conspecific36. This is interesting, because the V2r MHC-peptide
sensitive VSNs that respond to strain identity project to the posterior
subregion of the AOB, whereas the V1r VSNs that respond to testoster-
one-dependent urinary volatiles, such as SBT and DHB, project to the
anterior subregion. However, it is likely that little integration of infor-
mation occurs between these information streams at the level of the
AOB62. So, is there another source of male-specific chemosignals that
could be sensed by the V2rs? One possibility is the recently identified
7 kDa protein produced by the extraorbital lacrimal gland of male mice.
This is encoded by a member of a family of at least 23 related genes that
are expressed by other glands in the head region, including the salivary
and Harderian glands, although their behavioural effects are currently
The main olfactory system is thought to code information using a
pattern-recognition system in which chemosensory specificity arises out
of comparisons among input from receptor types with relatively broad
and overlapping molecular-receptive ranges. Hence, the relative propor-
tions of the MHC-dependent volatile constituents of urine odour are
represented by broad and overlapping but statistically different patterns
of activity in the MOB53. However, there is evidence that OSN responses
can be as selective as those seen in VSNs. For example, individual OSNs
in the MOE have been found to respond to MHC peptide ligands from
only a single MHC type10, and recordings from the MOB have shown
highly selective responses to the male attractant pheromone MTMT at
naturally occurring concentrations8.
Perhaps one of the most significant recent advances is the growth
of appreciation of the complementary roles of the main olfactory and
vomeronasal systems. OSNs and VSNs not only respond to overlap-
ping sets of social chemosignals, but do so with extraordinarily high
sensitivity (Fig. 2). The mouse MOE responds robustly to urinary
volatiles, such as 2-heptanone, at concentrations of 10–10 M (ref. 10),
which is comparable to the sensitivity of V1rb2-expressing VSNs34. The
responses of the two systems to MHC peptide ligands are also highly
sensitive, with robust responses at 10–10 M for OSNs10 and 10–12 M for
VSNs12. Moreover, both the main olfactory system and the vomero-
nasal system have access to neural systems controlling endocrine and
behavioural responses. For example, the vomeronasal pathway has long
been recognized as projecting to the region of the anterior hypothala-
mus, where luteinizing-hormone releasing hormone (LHRH) neurons
are located. A recent study using a barley lectin transynaptic tracer
expressed specifically in LHRH neurons has shown that they receive
direct synaptic connections from the vomeronasal pathway63. Moreo-
ver, this study and another using a viral transynaptic tracer targeted to
LHRH neurons have both shown abundant inputs from areas handling
main olfactory information63,64.
Pheromones and learning
Pheromones are generally thought to elicit innate responses that occur
in naive animals without prior learning. However, one of the best-
known examples of pheromonal effects in vertebrates is the selective
pregnancy block (or Bruce effect)21, which is associated with the forma-
tion and maintenance of an olfactory recognition memory by the vom-
eronasal system21,41,65. A female mouse learns the individual identity of
the male’s urinary chemosignals during mating. This is vital to prevent
implantation failure and abortion of her mate’s offspring upon subse-
quent exposure to his MHC peptide ligand-associated chemosignals12
(Fig. 3). Learning depends on a mating-induced noradrenaline release
in the AOB, which results in inhibition of her mate’s pheromonal sig-
nal at the level of the AOB65. This is proposed to disrupt transmission
of the pregnancy-blocking signal, and results in decreased activation
of central brain areas by her mate’s pheromones66,67, preventing the
activation of the neuroendocrine mechanisms leading to pregnancy
failure (Fig. 3).
Learning also has an important role in governing social odour pref-
erences in mice. The preference of mouse pups for maternal odours is
NATURE|Vol 444|16 November 2006
influenced by neonatal learning in the nest environment5. Such neona-
tal learning of maternal and sibling odours can have lasting effects on
MHC-associated mate preferences in adulthood, which can be partly
reversed by cross-fostering to a mother of dissimilar MHC type68,69. An
intriguing question is the extent to which MHC-related social prefer-
ences are innate, and whether these pheromonal effects are mediated
by MHC peptide ligands rather than by the MHC-dependent profile of
volatiles that is thought to constitute individual body odour.
Learning is an important component of many other pheromonal
responses. For example, the pheromone 2-methylbut-2-enal, found
in rabbit milk, elicits highly stereotyped nipple-search behaviour in
rabbit pups through the main olfactory system. However, if the doe’s
ventrum has been painted with an artificial odour, such as a perfume,
then the pups learn, in a single exposure, to associate that odour with
the pheromonal response70. Subsequent presentation of the conditioned
odour is sufficient to elicit full nipple-search behaviour in the absence
of the nipple-search pheromone. Such an association of chemosen-
sory cues and information from other sensory modalities is likely to
play an important part in reinforcing pheromonal responses in natural
Vomeronasal stimulation seems to have intrinsic reinforcing and
rewarding properties that not only maintain interest in social investi-
gation but could also act as an unconditioned stimulus in the associative
learning of main olfactory cues71. It is thought that odours conveyed by
the main olfactory system become conditioned by vomeronasal stim-
uli and subsequently become sufficient to drive the same behavioural
response in the absence of vomeronasal stimulation. For instance, VNO
lesions severely disrupt mating in sexually naive hamsters, whereas VNO
lesions in sexually experienced hamsters are less effective, as learned
chemosensory cues sensed through the main olfactory system have
become associated with mating72.
The brain’s dopaminergic reward system is likely to be heavily involved
in odour conditioning because both ejaculation and exposure to chemo-
signals from a receptive female are effective in increasing dopamine
release in the nucleus accumbens of male rats73. The increased activity
of neurons in the nucleus accumbens in response to a conditioned odour
to which the animal was exposed during copulation is mediated through
a main olfactory neural pathway74, in contrast to the vomeronasal areas
activated by chemosignals from a female in oestrus75.
The posterior medial amygdala responds to socially relevant conspe-
cific but not heterospecific stimuli76. Moreover, homeodomain tran-
scription factors delineate separate networks of neurons in the medial
amygdala that are activated by reproductive or defensive chemosensory
stimuli and project to hypothalamic areas controlling the appropriate
behavioural responses77. The cortical and medial regions of the amyg-
dala receive direct input from the AOB and MOB, and form a hub in
the networks governing social behaviour in mammals (Fig. 4). These
regions have been implicated in social-recognition learning in rodents78
and infant recognition in sheep79.
Do human pheromones exist?
Can human pheromones be identified? What are their effects? And can
they be exploited commercially? These are undoubtedly questions of
great public interest, and a controversial area of research. This is partly
because human social systems are too complex for human behaviour
to be dominated by the drastic releaser effects of pheromones seen in
other species. However, this does not mean that pheromonal signals
did not influence the behaviour and physiology of our recent human
ancestors, nor that they do not have subtle influences on humans
One of the main candidates for a source of human pheromones is the
apocrine-gland secretions of the axilla, which, when subject to microbial
action, produce the complex mixture of odorants responsible for body
odour, including androgen derivatives and volatile acids80. One of the
major constituents of axillary secretions is (E)-3-methyl-2-hexanoic
acid81. This is bound to apolipoprotein D, a member of the lipocalin
family of ligand-binding proteins82, which are important transporters
of pheromones in other species58. However, although apocrine secre-
tions are a likely source of potential human pheromones, little is known
about the chemosensory systems involved in their detection. The VNO
No pregnancy block
(low pregnancy failure rate)
(high pregnancy failure rate)
No pregnancy block
Figure 3 |The Bruce effect, one of the best-
known examples of olfactory imprinting in adult
vertebrates. Urine from unfamiliar males
(yellow) — or familiar urine (blue) supplemented
with unfamiliar (disparate) MHC class-1 peptide
ligands — activates neuroendocrine mechanisms
leading to pregnancy failure (a, b) by means of
the AOB, amygdala and medial hypothalamus
(c). The mated female learns to recognize
the pheromones of the mating male during a
sensitive period around mating. Subsequent
exposure to her mate’s urine activates mitral cells
(M) in the AOB that are subject to enhanced
inhibition from granule cell (G) interneurons.
This increased inhibition is thought to disrupt
the transmission of the mate’s pheromonal signal
to the amygdala, preventing pregnancy block.
(Images a and b are based on data from ref. 12.
Image c is based on data from refs 65–67, 99.)
NATURE|Vol 444|16 November 2006
is present early in human fetal development and the vomeronasal nerves
have a vital role in guiding the migration of developing LHRH neurons
to the hypothalamus83. However, the epithelial diverticulum in the adult
human nasal septum, referred to as the VNO, does not have a simi-
lar structure or function to the rodent VNO7,84. Moreover, in humans,
the gene encoding the TRPC2 cation channel is a pseudogene, hav-
ing been impaired and removed from functional constraints about 23
million years ago, shortly before the separation of hominoids and Old
World monkeys85,86. This had been seen as the final nail in the coffin for
the human VNO. However, the recent finding that V2r-type peptide-
responding VSNs do not require TRPC2 channel function41 reminds
us that our understanding of vomeronasal signal transduction is far
from complete and that this evidence is not as conclusive as originally
thought. More problematic for advocates of the existence of a human
VNO are the consistent failure to find vomeronasal nerves projecting to
the brain and the failure to find an AOB in adult humans83.
The main olfactory system is therefore the most likely pathway for
conveying any pheromonal effects in humans. Analysis of the human
genome has revealed that most V1r-like genes are pseudogenes87.
Intriguingly, however, four functional human V1r-like genes have been
found88. One of these, V1RL1, is expressed in the olfactory mucosa and
may well have a receptor function, although this does not necessarily
mean that it responds to a pheromone89. Indeed, members of the olfac-
tory-receptor-gene family, the newly discovered TAAR family20 or other
as yet unknown receptors could function as pheromone receptors in
humans as well as other species.
There is relatively good evidence for primer pheromonal effects in
humans. Exposure to axillary secretions from females in the late fol-
licular and ovulatory phases of their menstrual cycle have been found to
shorten and lengthen, respectively, the cycles of recipient females90. Fur-
thermore, male axillary extracts seem to influence female reproductive
state by affecting the frequency of LH release91. Evidence for signalling
pheromones in humans is also reasonably strong. Several studies have
found that females prefer odours from individuals of dissimilar MHC
type92,93, and this might even affect their choice of sexual partner94. It
remains to be determined whether MHC-peptides are detected by the
MOE in humans and, if so, whether this has a similar role to its media-
tion of individual odour preferences in mice10.
The breast odour of human mothers has been reported to attract the
attention of newborn babies and elicit movement towards the odour95,
which is comparable with the releaser effects of pheromones in other
vertebrate species. However, whether pheromones mediate sexual attrac-
tion in adult humans is a much more complex issue7. Although there are
reports of chemosensory effects on attractiveness ratings under labora-
tory conditions, there have not been many rigorous, placebo-controlled,
double-blind studies on sexual activity in natural social situations. Those
that have been performed must be interpreted with caution owing to the
large differences between individuals in their opportunities for, and the
nature of, their sexual interactions. The situation is further complicated
by fluctuations in female olfactory sensitivity during the menstrual cycle,
and reports that pheromones can have a modulatory effect in humans
to alter their mood96. This poses the question of whether changes in
measures of sexual attraction or behaviour are actually secondary to
changes in mood or general motivation. Functional MRI of brain activity
provides a promising technique with which to study the neural basis of
putative pheromonal effects in humans97.
Recent developments have led to an appreciation of the diversity of
chemosensory systems and their complementary roles in influencing
vertebrate physiology and behaviour. However, our understanding of
the nature of pheromones and the neural mechanisms underlying their
effects is still fragmentary. Future developments in this fast-moving
field will increasingly require researchers to integrate gene-target-
ing approaches with physiological and behavioural characterization.
A greater understanding of the transduction mechanisms involved
should allow selective and inducible ablation of specific receptor
systems. However, the phenotypes produced are likely to be complicated
by the interdependence and redundancy of the chemosensory systems
and the important role of learning. An understanding of the behavioural
effects needs to be coupled with gene-targeted tracing of specific recep-
tor systems using transynaptic tracers to map their neural projections.
Developments in imaging, telemetry and multiple-channel electrophys-
iological recording techniques promise to allow investigations of these
neural systems to be conducted in semi-natural behavioural situations
rather than an artificial laboratory environment.
The revolution in our understanding of vertebrate pheromonal
communication has been, and will continue to be, driven by molecu-
lar-genetic approaches. However, an important limitation is that our
understanding is based mainly on work in rodents, particularly mice. In
some respects this is highly appropriate, given the importance of chemo-
sensory information to them. However, pheromones are, by definition,
species-specific signals and caution must be exercised in extrapolating
between species, and between sexes, that have different reproductive
strategies and behavioural priorities98.
Integrated behavioural, endocrine
and autonomic responses
Figure 4 | The medial amygdala is an important hub that controls social
behaviour. The medial amygdala (MeA) receives input from both the
vomeronasal structures (red) and main olfactory structures (blue), and
is a major site for the integration of chemosensory information with
other sensory cues and hormonal states. The MeA elicits behavioural,
endocrine and autonomic responses to pheromones through its output to
the hypothalamus and central nucleus of the amygdala (CEA). Projections
to the basolateral nuclei of the amygdala (BLA) and nucleus accumbens
are likely to be involved in learned behavioural responses to conditioned
stimuli. Interconnections of the individual structures not involving the
MeA have been omitted for clarity. The central projections and possible
integration sites from the septal organ of Masera, the Grueneberg ganglion
and the GCD-cell system are unclear. ACo, anterior cortical amygdala;
AON, anterior olfactory nucleus; BAOT, bed nucleus of the accessory
olfactory tract; BNST, bed nucleus of the stria terminalis; NLOT, nucleus
of the lateral olfactory tract; PIR, piriform cortex; PLCo, posterior
lateral cortical amygdala; PMCo, posterior medial cortical amygdala.
This simplified version of the connectivity of the MeA is based on
neuroanatomical data from rats100.
NATURE|Vol 444|16 November 2006
1. Karlson, P. & Lüscher, M. Pheromones: a new term for a class of biologically active
substances. Nature 183, 55–56 (1959).
Schaal, B. et al. Chemical and behavioural characterization of the rabbit mammary
pheromone. Nature 424, 68–72 (2003).
Novotny, M. V., Weidong, M., Wiesler, D. & Zidek, L. Positive identification of the puberty-
accelerating pheromone of the house mouse: the volatile ligands associating with the
major urinary protein. Proc. R. Soc. Lond. B 266, 2017–2022 (1999).
Wyatt, T. D. Pheromones and Animal Behaviour (Cambridge Univ. Press, Cambridge, 2003).
Yamazaki, K., Beauchamp, G. K., Curran, M. & Boyse, E. A. Parent–progeny recognition as a
function of MHC odortype identity. Proc. Natl Acad. Sci. USA 97, 10500–10502 (2000).
Potts, W. K., Manning, C. J. & Wakeland, E. K. Mating patterns in seminatural populations
of mice influenced by MHC genotype. Nature 352, 619–621 (1991).
Wysocki, C. & Preti, G. Facts, fallacies, fears and frustrations with human pheromones.
Anat. Rec. A 281, 1201–1211 (2004).
Lin, D. Y., Zhang, S.-Z., Block, E. & Katz, L. C. Encoding social signals in the mouse main
olfactory bulb. Nature 434, 470–477 (2005).
Kikuyama, S. et al. Sodefrin: a female-attracting peptide pheromone in newt cloacal glands.
Science 267, 1643–1645 (1995).
10. Spehr, M. et al. Essential role of the main olfactory system in social recognition of major
histocompatibility complex peptide ligands. J. Neurosci. 26, 1961–1970 (2006).
11. Boehm, T. & Zufall, F. MHC peptides and the sensory evaluation of genotype. Trends
Neurosci. 29, 100–107 (2006).
12. Leinders-Zufall, T. et al. MHC class I peptides as chemosensory signals in the vomeronasal
organ. Science 306, 1033–1037 (2004).
13. Hurst, J. L. et al. Individual recognition in mice mediated by major urinary proteins. Nature
414, 631–634 (2001).
14. Morè, L. Mouse major urinary proteins trigger ovulation via the vomeronasal organ. Chem.
Senses 31, 393–401 (2006).
15. Kimoto, H., Haga, S., Sato, K. & Touhara, K. Sex-specific peptides from exocrine glands
stimulate mouse vomeronasal sensory neurons. Nature 437, 898–901 (2005).
16. Sorensen, P. et al. Mixture of new sulfated steroids functions as a migratory pheromone in
the sea lamprey. Nature Chem. Biol. 1, 324–328 (2005).
17. Dorries, K. M., Adkins, R. E. & Halpern, B. P. Sensitivity and behavioral responses to the
pheromone androstenone are not mediated by the vomeronasal organ in domestic pigs.
Brain Behav. Evol. 49, 53–62 (1997).
18. Novotny, M. V. Pheromones, binding proteins and receptor responses in rodents. Biochem.
Soc. Trans. 31, 117–122 (2003).
19. Zhang, X. & Firestein, S. The olfactory receptor gene superfamily of the mouse. Nature
Neurosci. 5, 124–133 (2002).
20. Liberles, S. D. & Buck, L. B. A second class of chemosensory receptors in the olfactory
epithelium. Nature 42, 645–650 (2006).
21. Halpern, M. & Martínez-Marcos, A. Structure and function of the vomeronasal system: an
update. Prog. Neurobiol. 70, 245–318 (2003).
22. Dulac, C. & Torello, A. T. Molecular detection of pheromone signals in mammals: from
genes to behaviour. Nature Rev. Neurosci. 4, 551–562 (2003).
23. Del Punta, K. et al. Deficient pheromone responses in mice lacking a cluster of vomeronasal
receptor genes. Nature 419, 70–74 (2002).
24. Lucas, P., Ukhanov, K., Leinders-Zufall, T. & Zufall, F. A diacylglycerol-gated cation channel
in vomeronasal neuron dendrites is impaired in TRPC2 mutant mice: mechanism of
pheromone transduction. Neuron 40, 551–561 (2003).
25. Rodriguez, I., Del Punta, K., Rothman, A., Ishii, T. & Mombaerts, P. Multiple new and
isolated families within the mouse superfamily of V1r vomeronasal receptors. Nature
Neurosci. 5, 134–140 (2002).
26. Yang, H., Shi, P., Zhang, Y. & Zhang, J. Composition and evolution of the V2r vomeronasal
receptor gene repertoire in mice and rats. Genomics 86, 306–315 (2005).
27. Hansen, A., Anderson, K. & Finger, T. Differential distribution of olfactory receptor neurons
in goldfish: structural and molecular correlates. J. Comp. Neurol. 477, 347–359 (2004).
28. Pfister, P. & Rodriguez, I. Olfactory expression of a single and highly variable V1r pheromone
receptor-like gene in fish species. Proc. Natl Acad. Sci. USA 102, 5489–5494 (2005).
29. Eisthen, H. The goldfish knows: olfactory receptor cell morphology predicts receptor gene
expression. J. Comp. Neurol. 477, 341–346 (2004).
30. Elsaesser, R., Montani, G., Tirindelli, R. & Paysan, J. Phosphatidyl-inositide signalling
proteins in a novel class of sensory cells in the mammalian olfactory epithelium. Eur. J.
Neurosci. 21, 2692–2700 (2005).
31. Meredith, M. Chronic recording of vomeronasal pump activation in awake behaving
hamsters. Physiol. Behav. 56, 345–354 (1994).
32. Leinders-Zufall, T. et al. Ultrasensitive pheromone detection by mammalian vomeronasal
neurons. Nature 405, 792–796 (2000).
33. Sam, M. et al. Odorants may arouse instinctive behaviours. Nature 412, 142 (2001).
34. Boschat, C. et al. Pheromone detection mediated by a V1r vomeronasal receptor. Nature
Neurosci. 5, 1261–1262 (2002).
35. Sharrow, S., Vaughn, J., Zídek, L., Novotny, M. & Stone, M. Pheromone binding by
polymorphic mouse major urinary proteins. Protein Sci. 11, 2247–2256 (2002).
36. Luo, M. M., Fee, M. S. & Katz, L. C. Encoding pheromonal signals in the accessory olfactory
bulb of behaving mice. Science 299, 1196–1201 (2003).
37. Xu, F. et al. Simultaneous activation of the mouse main and accessory olfactory bulbs by
odors or pheromones. J. Comp. Neurol. 489, 491–500 (2005).
38. Wysocki, C. J. & Lepri, J. L. Consequences of removing the vomeronasal organ. J. Steroid
Biochem. Mol. Biol. 39, 661–669 (1991).
39. Stowers, L., Holy, T. E., Meister, M., Dulac, C. & Koentges, G. Loss of sex discrimination and
male–male aggression in mice deficient for TRP2. Science 295, 1493–1500 (2002).
40. Leypold, B. G. et al. Altered sexual and social behaviors in trp2 mutant mice. Proc. Natl Acad.
Sci. USA 99, 6376–6381 (2002).
41. Kelliher, K., Spehr, M., Li, X.-H., Zufall, F. & Leinders-Zufall, T. Pheromonal recognition
memory induced by TRPC2-independent vomeronasal sensing. Eur. J. Neurosci. 23, 3385–
42. Hudson, R. & Distel, H. Pheromonal release of suckling in rabbits does not depend on the
vomeronasal organ. Physiol. Behav. 37, 123–129 (1986).
43. Mandiyan, V., Coats, J. & Shah, N. Deficits in aggressive behaviors in Cnga2 mutant mice.
Nature Neurosci. 8, 1660–1662 (2005).
44. Hurst, J. & Beynon, R. Scent wars: the chemobiology of competitive signalling in mice.
BioEssays 26, 1288–1298 (2004).
45. Zufall, F. & Munger, S. From odor and pheromone transduction to the organization of the
sense of smell. Trends Neurosci. 24, 191–193 (2001).
46. Ma, M. et al. Olfactory signal transduction in the mouse septal organ. J. Neurosci. 23,
47. Fuss, S., Omura, M. & Mombaerts, P. The Grueneberg ganglion of the mouse projects
axons to glomeruli in the olfactory bulb. Eur. J. Neurosci. 22, 2649–2654 (2005).
48. Jordan, W. C. & Bruford, M. W. New perspectives on mate choice and the MHC. Heredity
81, 127–133 (1998).
49. Boyse, E. A., Beauchamp, G. K. & Yamazaki, K. The genetics of body scent. Trends Genet. 3,
50. Carroll, L. S., Penn, D. J. & Potts, W. K. Discrimination of MHC-derived odors by untrained
mice is consistent with divergence in peptide-binding region residues. Proc. Natl Acad. Sci.
USA 99, 2187–2192 (2002).
51. Singh, P. B., Brown, R. E. & Roser, B. MHC antigens in urine as olfactory recognition cues.
Nature 327, 161–164 (1987).
52. Singh, P. B. Chemosensation and genetic individuality. Reproduction 121, 529–539 (2001).
53. Schaefer, M. L., Yamazaki, K., Osada, K., Restrepo, D. & Beauchamp, G. K. Olfactory
fingerprints for major histocompatibility complex-determined body odors II: relationship
among odor maps, genetics, odor composition, and behavior. J. Neurosci. 22, 9513–9521
54. Loconto, J. et al. Functional expression of murine V2R pheromone receptors involves
selective association with the M10 and M1 families of MHC class 1b molecules. Cell 112,
55. Ishii, T., Hirota, J. & Mombaerts, P. Combinational coexpression of neural and immune
multigene families in mouse vomeronasal sensory systems. Curr. Biol. 13, 394–400
56. Olson, R., Huey-Tubman, K., Dulac, C. & Bjorkman, P. Structure of a pheromone receptor-
associated MHC molecule with an open and empty groove. PLoS 3, e257 (2005).
57. Milinski, M. et al. Mate choice decisions of stickleback females predictably modified by
MHC peptide ligands. Proc. Natl Acad. Sci. USA 102, 4414–4418 (2005).
58. Flower, D. R. The lipocalin protein family: structure and function. Biochem. J. 318, 1–14 (1996).
59. Armstrong, S., Robertson, D., Cheetham, S., Hurst, J. & Beynon, R. Structural and functional
differences in isoforms of mouse major urinary proteins: a male-specific protein that
preferentially binds a male pheromone. Biochem. J. 391, 343–350 (2005).
60. Beynon, R. et al. Polymorphism in major urinary proteins: molecular heterogeneity in a wild
mouse population. J. Chem. Ecol. 28, 1429–1446 (2002).
61. Luo, M. & Katz, L. Encoding pheromonal signals in the mammalian vomeronasal system.
Curr. Opin. Neurobiol. 14, 428–434 (2004).
62. Sugai, T., Sugitani, M. & Onoda, N. Subdivisions of the guinea pig accessory olfactory bulb
revealed by the combined method with immunohistochemistry, electrophysiological, and
optical recordings. Neuroscience 79, 871–885 (1997).
63. Boehm, U., Zou, Z. & Buck, L. Feedback loops link odor and pheromone signaling with
reproduction. Cell 123, 683–695 (2005).
64. Yoon, H., Enquist, L. & Dulac, C. Olfactory inputs to hypothalamic neurons controlling
reproduction and fertility. Cell 123, 669–695 (2005).
65. Kaba, H. & Nakanishi, S. Synaptic mechanisms of olfactory recognition memory. Rev.
Neurosci. 6, 125–141 (1995).
66. Halem, H. A., Cherry, J. A. & Baum, M. J. Central forebrain responses to familiar male odors
are attenuated in recently mated female mice. Eur. J. Neurosci. 13, 389–399 (2001).
67. Binns, K. E. & Brennan, P. A. Changes in electrophysiological activity in the accessory
olfactory bulb and medial amygdala associated with mate recognition in mice. Eur. J.
Neurosci. 21, 2529–2537 (2005).
68. Yamazaki, K. et al. Familial imprinting determines H-2 selective mating preferences. Science
240, 1331–1332 (1988).
69. Penn, D. & Potts, W. MHC-disassortative mating preferences reversed by cross-fostering.
Proc. R. Soc. Lond. B 265, 1299–1306 (1998).
70. Hudson, R. Do newborn rabbits learn the odor stimuli releasing nipple-search behavior?
Dev. Psychobiol. 18, 575–585 (1985).
71. Moncho-Bogani, J., Martinez-Garcia, F., Novejarque, A. & Lanuza, E. Attraction to sexual
pheromones and associated odorants in female mice involves activation of the reward
system and basolateral amygdala. Eur. J. Neurosci. 21, 2186–2198 (2005).
72. Meredith, M. Vomeronasal, olfactory, hormonal convergence in the brain. Ann. NY Acad.
Sci. 855, 349–361 (1998).
73. Mitchell, J. & Gratton, A. Opioid modulation and sensization of dopamine release elicited
by sexually relevant stimuli: a high speed chronoamperometric study in freely behaving
rats. Brain Res. 551, 20–27 (1991).
74. West, C., Clancy, A. & Michael, R. Enhanced responses of nucleus accumbens neurons in
male rats to novel odors associated with sexually receptive females. Brain Res. 585, 49–55
75. Kippin, T., Cain, S. & Pfaus, J. Estrous odors and sexually conditioned neutral odors activate
separate neural pathways in the male rat. Neuroscience 117, 971–979 (2003).
76. Meredith, M. & Westberry, J. M. Distinctive responses in the medial amygdala to same-
species and different-species pheromones. J. Neurosci. 24, 5719–5725 (2004).
77. Choi, G. B. et al. Lhx6 delineates a pathway mediating innate reproductive behaviors from
the amygdala to the hypothalamus. Neuron 46, 647–660 (2005).
78. Ferguson, J. N., Aldag, J. M., Insel, T. R. & Young, L. J. Oxytocin in the medial amygdala is
essential for social recognition in the mouse. J. Neurosci. 21, 8278-8285 (2001).
79. Keller, M., Perrin, G., Meurisse, M., Ferreira, G. & Lévy, F. Cortical and medial amygdala are
both involved in the formation of olfactory offspring memory in sheep. Eur. J. Neurosci. 20,
80. Leyden, J., McGinley, K., Hölzle, E., Labows, J. & Kligman, A. The microbiology of the human
axilla and its relationship to axillary odor. J. Invest. Dermatol. 77, 413–416 (1981).
81. Zeng, X.-N. et al. Analysis of characteristic odors from human male axillae. J. Chem. Ecol. 17,
NATURE|Vol 444|16 November 2006
82. Zeng, C. et al. A human axillary odorant is carried by apolipoprotein D. Proc. Natl Acad. Sci.
USA 93, 6626–6630 (1996).
83. Witt, M. & Hummel, T. Vomeronasal versus olfactory epithelium: is there a cellular basis
for human vomeronasal perception? Int. Rev. Cytol. 248, 209–259 (2006).
84. Meredith, M. Human vomeronasal organ function: a critical review of best and worst
cases. Chem. Senses 26, 433–445 (2001).
85. Zhang, J. & Webb, D. Evolutionary deterioration of the vomeronasal pheromone
transduction pathway in catarrhine primates. Proc. Natl Acad. Sci. USA 100, 8337–8341
86. Liman, E. R. & Innan, H. Relaxed selective pressure on an essential component of
pheromone transduction in primate evolution. Proc. Natl Acad. Sci. USA 100, 3328–3332
87. Kouros-Mehr, H. et al. Identification of non-functional human VNO receptor genes
provides evidence for vestigiality of the human VNO. Chem. Senses 26, 1167–1174 (2001).
88. Rodriguez, I. & Mombaerts, P. Novel human vomeronasal receptor-like genes reveal
species-specific families. Curr. Biol. 12, R409–R411 (2002).
89. Rodriguez, I., Greer, C. A., Mok, M. Y. & Mombaerts, P. A putative pheromone receptor
gene expressed in human olfactory mucosa. Nature Genet. 26, 18–19 (2000).
90. Stern, K. & McClintock, M. K. Regulation of ovulation by human pheromones. Nature 392,
91. Preti, G., Wysocki, C., Barnhart, K., Sondheimer, S. & Leyden, J. Male axillary extracts
contain pheromones that affect pulsatile secretion of luteinizing hormone and mood in
women recipients. Biol. Reprod. 68, 2107–2113 (2003).
92. Wedekind, C. & Furi, S. 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 (1997).
93. Jacob, S., McClintock, M. K., Zelano, B. & Ober, C. Paternally inherited HLA alleles are
associated with women’s choice of male odor. Nature Genet. 30, 175–179 (2002).
94. Ober, C. et al. HLA and mate choice in humans. Am. J. Hum. Genet. 61, 497–504
95. Varendi, H. & Porter, R. H. Breast odour as the only maternal stimulus elicits crawling
towards the odour source. Acta Paediatr. 90, 372–375 (2001).
96. Jacob, S. & McClintock, M. Pyschological state and mood effects of steroidal chemosignals
in women and men. Horm. Behav. 37, 57–78 (2000).
97. Savic, I., Berglund, H., Gulyas, B. & Roland, P. Smelling of odorous sex hormone-like
compounds causes sex-differentiated hypothalamic activations in humans. Neuron 31,
98. Keverne, E. Odor here, odor there: chemosensation and reproductive function. Nature
Neurosci. 8, 1637–1638 (2005).
99. Brennan, P. A., Kendrick, K. M. & Keverne, E. B. Neurotransmitter release in the accessory
olfactory bulb during and after the formation of an olfactory memory in mice. Neuroscience
69, 1075–1086 (1995).
100. Pitkänen, A. in The Amygdala: a Functional Analysis (ed. Aggleton, J.) 31–115 (Oxford
University Press, Oxford, 2000).
Acknowledgements In memory of L. C. Katz. This review and our research were
supported by grants from the NIH/NIDCD (F.Z.) and the MRC/BBSRC (P.A.B.).
Apologies to our colleagues whose work we could not cite owing to space
Author Information Reprints and permissions information is available at
npg.nature.com/reprintsandpermissions. The authors declare no competing
financial interests. Correspondence should be addressed to the authors
NATURE|Vol 444|16 November 2006