O papel da comunicação química na escolha do parceiro

Prévia do material em texto

Biol. Rev. (2007), 82, pp. 265–289. 265
doi:10.1111/j.1469-185X.2007.00009.x
The role of chemical communication in
mate choice
Björn G. Johansson1* and Therésa M. Jones2
1Department of Animal Ecology, Uppsala University, Sweden
2Department of Zoology, The University of Melbourne, Australia
(Received 10 February 2005; revised 17 January 2007; accepted 30 January 2007)
ABSTRACT
Chemical signals are omnipresent in sexual communication in the vast majority of living organisms. The
traditional paradigm was that their main purpose in sexual behaviour was to coordinate mate and species
recognition and thus pheromones were conserved in structure and function. In recent years, this view has been
challenged by theoretical analyses on the evolution of pheromones and empirical reports of mate choice based
on chemical signals. The ability to measure precisely the quantity and quality of chemicals emitted by single
individuals has also revealed considerable individual variation in chemical composition and release rates, and
there is mounting evidence that prospecting mates respond to this variation. Here, we review the evidence for
pheromones as indicators of mate quality and examine the extent of their use in individual mate assessment. We
begin by briefly defining the levels of mate choice – species recognition, mate recognition and mate assessment.
We then explore the degree to which pheromones satisfy the key criteria necessary for their evolution and
maintenance as cues in mate assessment; that is, they should exhibit variation across individuals within a sex and
species; they should honestly reflect an individual’s quality and thus be costly to produce and/or maintain; they
should display relatively high levels of heritability. There is now substantial empirical evidence that pheromones
can satisfy all these criteria and, while measurements of the actual metabolic cost of pheromone production
remain to some degree lacking, trade-offs between pheromone production and various fitness-related characters
such as growth rate, immunocompetence and longevity have been reported for a range of species. In the
penultimate section, we outline the growing number of studies where the consequences of chemical-based mate
assessment have been investigated, specifically focussing on the reported direct and genetic benefits accrued by
the receiver. Finally, we highlight potential areas for future research and in particular emphasise the need for
interdisciplinary research that combines exploration of chemical, physiological and behavioural processes to
further our understanding of the role of chemical cues in mate assessment.
Key words: pheromone, chemical cues, species recognition, mate assessment, individual variation, direct bene-
fits models, honest signalling, good genes, Fisherian benefits.
CONTENTS
I. Introduction ...................................................................................................................................... 266
II. Chemical signals and levels of mate choice ..................................................................................... 267
(1) Pheromones employed in species recognition ............................................................................ 267
(2) Pheromones employed in mate recognition .............................................................................. 268
(3) The evolution of species and mate recognition pheromones .................................................... 269
(4) Pheromones employed in mate assessment ............................................................................... 269
*Address for correspondence: Björn G. Johansson, Department of Animal Ecology, Evolutionary Biology Centre, Uppsala University,
Norbyvägen 18D, SE - 752 36 Uppsala, Sweden. Tel: ]46-18-4716495. Fax: ]46-18-4716484. E-mail: bjorn.johansson@ebc.uu.se
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
III. Key criteria for mate assessment pheromones ................................................................................. 270
(1) Individual variation ..................................................................................................................... 270
(2) The cost of chemical signals ....................................................................................................... 270
(3) Additive genetic variance ........................................................................................................... 274
IV. Consequences of mate assessment .................................................................................................... 275
(1) Pheromones as direct benefits or indicators thereof ................................................................. 276
(a ) Pheromones as a resource .................................................................................................... 276
(b ) Pheromones as indicators of a resource ............................................................................... 276
( c ) Pheromones as indicators of fertility .................................................................................... 277
(2) Pheromones as indicators of indirect benefits ........................................................................... 278
(a ) Pheromones as indicators of ‘‘good genes’’ .......................................................................... 278
(b ) Pheromones as indicators of genetic compatibility .............................................................. 279
(3) Pheromones as indicators of Fisherian benefits ......................................................................... 280
V. Conclusions ....................................................................................................................................... 280
VI. Acknowledgments ............................................................................................................................. 281
VII. References ......................................................................................................................................... 281
I. INTRODUCTION
Chemical signalling is generally regarded as the most
ancient and widespread form of communication (Bradbury
& Vehrencamp, 1998; Wyatt, 2003). Pheromones (derived
from the Greek pherein to carry or transfer and hormôn to
excite) are chemical signals involved specifically in infor-
mation exchange between individuals within a species
(Karlson & Lüscher, 1959). A role for chemical signals in
promoting the process of sexual selection was proposed
initially at the end of the 19th Century (Darwin, 1871).
However, it was not until the mid-20th Century that the first
pheromone, the sex pheromone of the silk moth Bombyx
mori, was formally identified (Butenandt et al., 1959).
Extraction and analysis of the pheromone was arduous
and required sacrificing thousands of individuals. Today,
analyses of the chemicals emitted by live individuals are
commonplace and chemical-based signalling has been
reported for taxa ranging from bacteria to mammals (see
Table 1). Pheromones have been identified from about 15
orders of organisms (El-Sayed, 2006), including well over
3000 species of insects (Witzgall et al., 2004) and an
increasing number of mammals (for recent reviews, see
Burger, 2005; Marchlewska-Koj, Lepri & Müller-Schwarze,
2001). Moreover, the transmission and perception of
pheromones has been modelled extensively (e.g. Kauer &
White, 2001; Meredith, 1991; Metcalf, 1998). Given the
extent of research and the recognised importance of
pheromones as signals in sexual selection – all but two of
the reviews in Table 1 makes some reference to ‘sex
pheromones’ – it is perhaps then surprising that their role in
individual mate assessment remains relatively unexplored.
The vast majority of studies exploring pheromones in
sexual selection have focussed attention on long-range mate
attraction(Svensson, 1996) or the role of chemical signals in
species and mate recognition (Löfstedt, 1993; Ptacek, 2000).
Indeed in his seminal review of sexual selection Malte
Andersson (1994) cites just nine studies that had investigated
pheromone-based mating preferences; in striking contrast,
several hundred studies are cited that had examined the
consequences of visual or acoustic signals. While the
number of studies that explore pheromone-based signalling
in mate assessment is increasing rapidly, the above disparity
remains. The objective of this review is to demonstrate the
wide range of information conveyed by pheromones with
respect to mate quality and to highlight the potential for
pheromones to be utilised in mate assessment.
Broadly speaking, the aim of sexual communication
(regardless of the sensory modality used) is to influence mate
choice, where mate choice is defined as ‘any behaviour that
restricts the set of potential mates’ (Wiley & Poston, 1996).
Such behaviours should be beneficial whenever they
increase the number and/or quality of offspring for choosy
individuals as compared to those individuals that mate
randomly. However, this wide definition of mate choice
necessitates some sub-categories. First, we might distinguish
between ‘direct’ and ‘indirect’ mate choice. Direct mate
choice (sensu Wiley & Poston, 1996), ‘requires discrimination
between attributes of individuals of the opposite sex’,
whereas indirect mate choice is the result of all other
behaviours that restrict a set of potential mates. (These terms
differ from the direct and indirect benefits of mate choice
discussed in Section IV below). Second, the set of potential
mates might be restricted to the right species, the right mate
recognition system (sensu Paterson, 1985) or particular
individual qualities that maximise the reproductive fitness
of the choosing sex. These three levels of choice are
mutually non-exclusive and may be viewed as a continuum
ranging from sexual isolation between species through to
individual mate assessment (Ryan & Rand, 1993).
The original definition of a pheromone was a particular
chemical(s) secreted by an animal that ‘releases a specific
reaction’ in a member of the same species (Karlson &
Lüscher, 1959). However, as the knowledge of the variety of
behaviours elicited by pheromones has increased, a plethora
of often quite vaguely defined sub-terms has developed.
Words like ‘courtship pheromone’ or ‘aphrodisiac’ are often
quite misleading in that they imply a function that is, in
many cases, inferred rather than observed. On the other
hand, ‘sex pheromone’ has been used for chemicals with
Björn G. Johansson and Therésa M. Jones266
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
very different behavioural effects, from mate attraction
through to regulation of physiological processes (such as
ovulation). For example, pheromones used to attract both
sexes to a particular feeding site are defined as aggregation
pheromones. However, an aggregation pheromone may
also increase the mating success of a particular individual
and is thus effectively a sex pheromone (Raffa, 2001).
Mate choice often involves a process of evaluation of
information that does not lead to the ‘specific reaction’
required of pheromones in the strict sense. Moreover, many
chemical signals influencing mate choice are not well-
defined blends of specific chemicals but complex mixtures
that vary greatly between individuals in the same species. In
particular, mammal chemical signals often fall outside the
traditional definition of pheromones for the above reasons,
although they obviously play an important role in behaviour
(Brennan & Keverne, 2004). Since individuality in signalling
is at the core of mate assessment, here we define a ‘sex
pheromone’ as any substance that is released by an
individual, either directly from a specialised structure or
that arises through changes in body chemistry, and that
promotes subsequent variation in the sexual behaviour of
individuals within the same species to the benefit of the
releasing individual. Note that this is broader than the
traditional definition of a pheromone (sensu Karlson &
Lüscher, 1959), but it incorporates the wide range of
chemical signals that are used as indicators of fertility or
genetic compatibility and odours that are not synthesised by
the individual per se but are acquired from an external
source such as a conspecific during mating or directly from
consumed food. Herein we use the terms pheromone and
chemical signals interchangeably and ‘chemical cue’ when
the benefits accrued by the individual releasing the
chemical (the sender) are either unknown or questionable.
Within the term sex pheromone we define three sub-
categories, which are involved in the process of mate choice:
species recognition pheromones; mate recognition pher-
omones and mate assessment pheromones. Species recog-
nition pheromones are sex pheromones that are used to
distinguish individuals of different species. Mate recognition
pheromones are sex pheromones that are used to co-
ordinate the sexual behaviour between the sexes. Finally,
mate assessment pheromones are sex pheromones that
individuals of one sex use to discriminate between in-
dividuals of the other sex. These three categories may be
mutually non-exclusive.
In Section II, we briefly review the three sub-classes of
sex pheromones and outline a limited number of examples.
In Section III, we explore the key criteria necessary for the
evolution and maintenance of pheromones employed in the
final level of mate assessment. In Section IV we review
studies that have explored specifically the consequences of
using chemical-based signals in mate assessment. Finally, in
Section V we offer potential avenues for future research.
II. CHEMICAL SIGNALS AND LEVELS OF MATE
CHOICE
(1) Pheromones employed in species recognition
To achieve a viable mating a signaller must be able to
convey accurately its species, sex and reproductive status.
While the process of mate choice is conserved by the
necessity to perceive and locate a suitable mate, chemical
signals may be highly divergent across species. The main
processes driving the evolution of pheromones involved
in species recognition are interspecific competition for
Table 1. Recent reviews on chemical communication. The
table illustrates the width of chemical communication across
various taxa by giving examples of reviews from the last ten
years. If the main focus of the review is structure, it is
concerned primarily with the chemical composition of
chemical signals; synthesis, the chemistry of the production of
chemical signals in the laboratory and/or living organisms;
function, the role of chemical signals in behaviour and ecology
Taxa
Main focus
of review Reference
Bacteria Structure and synthesis Chhabra et al. (2005)
Function Ben Jacob et al.
(2004)
Structure and function Prozorov (2001)
Plants Structure and function Sekimoto (2005)
Fungi Genetics and function Casselton (2002)
Mushrooms Function and genetics Brown & Casselton
(2001)
Ciliates Structure and
perception
Luporini et al. (2005)
Animals Function Wyatt (2003)
Nematodes Perception Perry (1996)
Rotifers Function Snell (1998)
Molluscs Structure Cummins et al. (2005)
Structure Susswein & Nagle
(2004)
Polychaetes Function Hardege et al. (1998)
Arthropods Function Greenfield (2002)
Crustacea Structure and synthesis Rittschof & Cohen
(2004)
Ticks Pest control Sonenshine (2006)
Insects Synthesis Tillman et al. (1999)
Heteropterans Structure and synthesis Millar (2005)
Beetles Structure and synthesis Francke & Dettner
(2005)
Genus Drosophila Evolution and function Ferveur (2005)
Lepidopterans Structure Ando et al. (2004)
Function Svensson (1996)
Hymenopterans Structure Keeling et al. (2004)
Structure and function Ayasse et al. (2001)
Fish Structure and function Stacey (2003)
Mammals Structure Burger (2005)
Function Gosling & Roberts
(2001)
Structureand
perception
Brennan & Keverne
(2004)
Rodents Function Johnston (2003)
Humans Function Grammer et al. (2005)
Structure and function Wysocki & Preti (2004)
Function Hays (2003)
Chemical communication in mate choice 267
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
communication channels and, for closely related species,
selection for pre-mating reproductive isolation (Cardé &
Baker, 1984; Löfstedt, 1993; Roelofs & Cardé, 1974). These
are the mechanisms proposed for the evolution of the long-
range pheromones of moths (Svensson, 1996) and mammals
(Ptacek, 2000); a view based on the idea that the specificity
of pheromones evolved in these taxa to minimise the costs
of searching for, courting and/or mating with partners of
other species.
Several mechanisms have evolved that reduce the
probability of mistaking pheromones from a closely related
sympatric species for those of a conspecific. Pheromones of
sister-taxa often have dramatic qualitative differences in
their chemical structure or composition (Coyne & Orr,
1997; Frey, Lonsdale & Snell, 1998; Roelofs, 1995; Shine
et al., 2002). In bark beetles, genus Ips, the pheromones used
as long-range attractants to feeding and mating sites are
equally (or more) dissimilar from one another in closely
related species as they are from more distantly related taxa
(Symonds & Elgar, 2004a, b). Species that employ the same,
or similar, chemicals may evolve additional mechanisms
that reduce the probability that they respond to phero-
mones of non-conspecifics. For example, males of some
moth species have chemical receptors sensitive to the
pheromones of sister species (Löfstedt, 1990). The function
of these receptors seems to be avoidance of heterospecific
females, since adding a component from the pheromone of
one of these to the conspecific chemical blend often inhibits
attraction (Löfstedt et al., 1990). The partitioning of
communication channels may also be achieved by temporal
or seasonal differences in the timing of pheromone release
(Greenfield & Karandinos, 1979; Schal, 1982). Differences
in pheromone composition may also be combined with
different release times, as in two closely related scarab
beetles, Anomala albopilosa and A. cuprea, which share one
major pheromone component. Species specificity is main-
tained by across-species variation in minor components
within the pheromone blend and by temporal differences in
pheromone release during the day (Leal et al., 1996).
(2) Pheromones employed in mate recognition
Mating is a complicated affair that requires close coordina-
tion between mates: sequences of behaviour must be co-
ordinated at the appropriate time and place by two
individuals who do not necessarily share exactly the same
interests. Signals advertising sex, receptivity and specific
phases in mating behaviour are common, and we have
chosen to call these mate recognition signals in allusion to
the ‘specific mate recognition system’ of Paterson (1985) and
others. The predominant factor affecting the evolution of
these signals is the environment, and although evidence from
chemical-based signals is lacking, several studies of visual
(e.g. Endler, 1995; Marchetti, 1993) and acoustic signals (e.g.
Slabbekoorn & Smith, 2002) reveal how environmental
change can transform sexual characters or traits.
Odours seem to be ideal indicators of transitory states
such as changing reproductive status, perhaps because these
physiological changes are accompanied or directed by
changes in internal chemistry. At its most extreme, the same
substance might synchronise reproduction both within and
between individuals, as seen in goldfish Carassius auratus
(Stacey et al., 2003). An individual that is able to
discriminate the reproductive status of a potential mate
from a distance will minimise search costs by avoiding less
desirable or even unavailable partners. In mammals,
pheromones frequently advertise when a female is likely
to be receptive (Johnston, 1983; Marchlewska-Koj et al.,
2001). From the female perspective, chemical advertisement
of receptivity also reduces the probability of male
harassment during the non-reproductive period. Chemical
signals indicating female receptivity are thus beneficial to
both the signaller (females) and the receiver (males). By
contrast, chemical cues that indicate the mating status (for
example, mated or virgin) of an individual are likely to be
most beneficial to the discriminating sex. By preferentially
mating with virgin females, males may reduce the risk of
sperm competition and thus increase their share of
paternity. From a female perspective this may reduce the
opportunity to select additional mates and potentially
creates conflict between the sexes.
There are many examples of male discrimination of
female mating status based on chemical cues which are often
facilitated by changes in the body chemistry of females
following copulation. For example, male brown lemmings,
Lemmus sibiricus, prefer the odour of unmated over mated
females, thus reducing the potential costs of aggression by
attempting to copulate with already mated females (Huck,
Banks & Coopersmith, 1984). Males of the copepod Tigriopus
japonicus use certain proteins on the females’ body surface to
discriminate against mated females (Ting, Kelly & Snell,
2000). Virgin females of the spider Agelenopsis aperta emit an
odour that elicits courtship behaviour in males and males
thus avoid copulating with already mated females (Riechert
& Singer, 1995). In the biting midge Culicoides nubeculosus,
virgin females release a volatile chemical cue that stimulates
an electrophysiological response from males (Mair & Black-
well, 1998); mated females continue to emit the chemical but
it elicits a lower response from males. It is unknown whether
this response to mating is male- or female-induced, but
female odour release is negatively correlated with resistance
behaviour. Female resistance may in part resolve the conflict
over mating opportunity in this species as virgin females also
use resistance to assess the vigour of their mate prior to
copulation. Discrimination by males may also occur through
chemicals that have been transferred previously to females
by other males during mating. For example, male garter
snakes, Thamnophis sirtalis, deposit a mating plug in the
female following copulation. The mating plug contains
chemical cues that are used by subsequent males to
discriminate against newly mated females (O’Donnell et al.,
2004). Similarly, within the spermatophore transferred to
a female during mating, many male butterflies include ‘‘anti-
aphrodisiac’’ substances that ultimately lower a female’s
receptivity and reduce her attractiveness to future mates
(Andersson, Borg-Karlson & Wiklund, 2000).
Response to mate recognition pheromones may vary
temporally. During the mating season, males of the stripe-
necked terrapin, Mauremys leprosa, prefer water containing
female rather than male pheromones; however they avoid
Björn G. Johansson and Therésa M. Jones268
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
water containing pheromones of conspecifics outside the
reproductive period (Munoz, 2004). The ability to perceive
mate recognition pheromones may vary between the sexes.
For example, in the crayfish Austropotamobius pallipes, females
orientate to the chemical signals of males, however males
require both chemical and visual signals to perceive females
(Acquistapace et al., 2002).
(3) The evolution of species and mate
recognition pheromones
From an evolutionary point of view, selection for species and
mate recognition produce similar results. Both are driven by
non-intraspecific processes, i.e. interspecific competition
and environmental change respectively; both are essentially
adaptive (Löfstedt, 1993); and both may be subject to
stabilizing selection (Cardé & Baker, 1984; Paterson, 1985).
This is becausea failure to advertise your species or, for
example, your receptivity, might lead to no reproduction at all.
Species and mate recognition signals should thus be
evolutionarily stable and show little variation among indi-
viduals, and we should not expect the exaggeration that is
often associated with mate assessment signals. Furthermore,
since these signals merely coordinate the sexual behaviour
of males and females, they could be sent by either sex.
Many of the most well-studied pheromones of moths
conform to these predictions (for reviews, see Greenfield,
2002, Chapter 3; Svensson, 1996). According to the
traditional view, a male following a pheromonal track
emitted by a female has not chosen this specific signal and
discarded others, but is rather engaged in scramble
competition with other males over mates (sensu Svensson,
1996). From a female perspective, this amounts to a form of
indirect mate choice for the males that move fastest, have
the best orientating abilities or the greatest sensitivity to the
pheromone (Cardé & Baker, 1984). There are some
characteristics of moth mating systems supporting this view.
First, the pheromones are produced in minute quantities,
which seems an unlikely result if they were subject to mate
choice, but could trigger indirect mate choice for sensitive
males. Second, males of some moth species respond to
a broader spectrum of pheromonal compositions than their
conspecific females actually emit (Löfstedt, 1990). These
two factors suggest that the selection pressure driving the
evolution of perception by the receiver (i.e. the male moth)
is likely to be stronger than that driving the evolution of
female signalling capacity.
The early classical view for the role of sex pheromones
was as highly conserved species (Cardé & Baker, 1984;
Roelofs & Cardé, 1974) or mate recognition (sensu Paterson,
1985) signals that had limited potential for mate assessment.
These theories were based on the idea that species or mate
recognition are crucial for fitness and require tight co-
adaptation between the sexes. In either case, chemical
signals would be maintained by stabilizing selection and
should be largely resistant to change. This view has
gradually given way to theoretical and empirical studies
suggesting that, far from being conserved and stable,
pheromones may evolve rapidly in large saltational shifts
and as a consequence may be extremely divergent,
particularly across closely related species (Baker, 2002;
Roelofs et al., 2002; Roelofs & Rooney, 2003; Symonds &
Elgar, 2004a, b). These studies provide some of the best
evidence to date of the potential for rapid evolution of
chemical signals and as such pave a theoretical pathway for
their application in mate assessment in addition to species
and sex recognition.
(4) Pheromones employed in mate assessment
Mate assessment is promoted, at least in part, through the
benefits accrued from choosing a mate that is in some way
superior to its conspecifics, assuming that this superiority
translates into more or better quality offspring. In contrast
to species and mate recognition signals, mate assessment
signals must advertise the identity of the sender and its
potential quality as a mate. Moreover, it should do this in
a way that is genuine. Mate assessment signals are thus
expected to be costly and vary qualitatively with the
condition of the sender in order to provide honest
information in a stable signalling system (Grafen, 1990;
Rowe & Houle, 1996). They should furthermore be highly
variable between individuals and are thus prone to a kind of
exaggeration that may be detrimental to the survival of the
sender. This arises because mate assessment signals are not
constrained by the necessity to communicate an unequiv-
ocal and absolute message of, for example, species identity,
but are assessed on a relative scale where misjudgement
merely affects the quality or number of offspring.
Species recognition, mate recognition and mate assess-
ment are mutually non-exclusive functions of a signal. For
example, male bark beetles of the genus Ips produce
pheromones that attract both sexes from a distance and that
females use for mate assessment (Schlyter & Zhang, 1996).
Similarly, the male pheromone of the cockroach Nauphoeta
cinerea attracts females, has a role in male dominance
hierarchy and acts as a signal of male quality in female mate
choice (Moore & Moore, 1999). Interestingly, in both these
cases the different functions of the pheromones are to some
extent mediated by different components in their phero-
monal blends. In a cross-breeding experiment with the fruit-
flies Drosophila birchii and D. serrata, Blows & Allan (1998)
found that specific (and different) cuticular hydrocarbons
were responsible for species and sex recognition in the two
species, but that hydrocarbons important in both these
functions in the parents were used in mate choice in the
hybrids. This suggests that signals used in species recogni-
tion could evolve from signals with mate recognition or
mate assessment functions (Blows & Allan, 1998). Thus, it
might be helpful to keep the distinctions between these
functions in mind when investigating the role of a specific
sexual signal. In the study of chemical signals in particular,
where signal function is often summarily described by the
word ‘sex pheromone’, additional insights might be gained
by investigating at which level a specific pheromone acts.
The number of studies that have investigated whether
species signal or can perceive sex pheromones has been
growing steadily for the last forty years (for a comprehensive
recent review see Wyatt, 2003). Fewer studies have actually
Chemical communication in mate choice 269
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
investigated whether pheromones are used in mate
assessment. For species and mate recognition it is required
that an individual recognises and orientates towards
a particular chemical cue. By contrast, to demonstrate that
pheromones are used in mate assessment requires that an
individual is able to discriminate between the chemical
signals emitted by individuals within a sex and species.
Table 2 reviews those studies that have demonstrated an
ability to both discriminate between and move towards the
chemical signals of one particular individual over another.
Mate assessment pheromones are correlated either directly
or indirectly with traits such as condition, fertility, female
reproductive status, age, parasite load, nutritional status,
maturity or immunocompetence and have been identified
in a range of taxa (see Table 2 and Section IV).
III. KEY CRITERIA FOR MATE ASSESSMENT
PHEROMONES
There are a number of characteristics that could be used
as indicators when identifying putative mate assessment
pheromones. (1) Mate assessment pheromones must vary
across individuals within a sex and species, thus making
individual mate choice possible. (2) To ensure that a mate
assessment signal honestly reflects an individual’s quality,
pheromones used in mate assessment should be costly to
produce and/or maintain. (3) In general, the additive
genetic variance of a trait determines its evolutionary
potential, and sexually selected traits are expected to show
higher additive genetic variance than traits more important
to survival. In this section, we review the evidence that
pheromones satisfy each of these criteria. It should be noted
that a chemical signal does not qualify as a mate assessment
pheromone just because it possesses any, or indeed all, of
the above characteristics. Ultimately, mate assessment
pheromones must be shown to guide receivers to the mates
that produce the most or/and best offspring.
(1) Individual variation
Exploration of individual variation in pheromone pro-
duction was, until recently, limited by our inability to
identify the often minute quantities of chemicals produced
by a specific individual (Schlyter & Birgersson, 1989).
However, an increasingbody of research has revealed
substantial individual variation in pheromones. Variation
may be achieved through quantitative or qualitative differ-
ences in the pheromone blend, or via differential release
rates. In general, the quantity of pheromone released varies
more than the ratios of its components (Table 3).
In a comparison among several species of moths and bark
beetles, Schlyter & Birgersson (1989) found greater
individual variation in the amounts of pheromones released
than in the composition of the pheromone released for both
groups. However, bark beetles produced greater absolute
amounts of pheromones than moths, and individual
variation in this character was significantly higher in bark
beetles. It is suggested that these differences may arise
because of the different mating strategies of moths and bark
beetles. In moths, isolated females release pheromones to
attract males, thus sexual selection is more likely to operate
on the male response to the signal than on the female
pheromone (Svensson, 1996). By contrast, male bark beetles
release pheromones from large aggregations, making mate
choice for the most elaborate signal possible (Schlyter &
Birgersson, 1989).
Variation in the number of components contained within
a pheromone blend may also contribute to variation in the
behavioural responses to pheromone cues. Insect phero-
mones are generally assumed to have a fixed number of
active components, and although a large number of ‘minor
components’ are often revealed in chemical analysis of
pheromone glands or effluvia, these are frequently dismissed
as ‘residuals’ or ‘precursors’. By contrast, studies on
mammal chemical signals have been directed at exploring
the varying numbers and types of components within the
chemical blend. For example, chemical extracts from the
sub-caudal glands of the European badger, Meles meles, are
highly varied across individuals providing each with a unique
chemical signature (Buesching & Macdonald, 2004; Buesching,
Waterhouse & Macdonald, 2002a). Similar patterns are
observed in chemical secretions produced by the shrew
Crocidura russula (Cantoni et al., 1996) and several species of
deer (Lawson, Putman & Fielding, 2000). It should be noted
that these studies were searching for signals communicating
individuality per se rather than mate quality. In most cases,
the number of possible components in the blend is large
enough to provide an individual identity, although there
are a few mammal studies indicating that individuality
is signalled by the proportions of a fixed number of
components (Gorman, 1976; Smith et al., 2001). Individual
recognition is important in many social interactions apart
from mate choice in mammals, for example in dominance
hierarchies, territorial defence and parent-infant interac-
tions. However, an obvious example of where individual
odours may be utilised as mate assessment signals are those
implicated in mate selection of specific major histocompat-
ibility complex-profiles reported from a growing number of
vertebrates (see Section IV 2b).
(2) The cost of chemical signals
From an evolutionary perspective, the signals that we
observe represent the optimum balance between the relative
costs and benefits of signal production (Bergstrom &
Lachmann, 1997; Johnstone, 1999). The relative cost of
producing and responding to pheromones is likely to
determine when an individual should signal (Bergstrom &
Lachmann, 1997; Gosling & Roberts, 2001) and which
sex produces the signal. In general, when pheromone
production is cheap but search costs are high, such as seen
in the majority of moths (Greenfield, 1981; Svensson, 1996)
it is the female that is most likely to produce the signal. By
contrast, when pheromone production is costly, then it is
males that are likely to produce the pheromone. This is
based on the assumption that females invest most in
production of offspring and should invest least effort or risk in
mate attraction (Kokko & Monaghan, 2001; Trivers, 1972).
Björn G. Johansson and Therésa M. Jones270
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
Table 2. Experimental studies reporting pheromone-based mate assessment. Sex is the pheromone-emitting sex. Message is the
information that the receiver extracts from the signal according to the respective authors; FA ¼ fluctuating asymmetry; Imm. ¼
immunocompetence; MHC ¼ major histocompatibility complex. Benefit is the benefit of choosing for the receiver according to the
respective authors
Taxon Sex Message Benefit Reference
Fungi
Saccharomyces cerevisiae Both1 Mate quality – Jackson & Hartwell (1990)
Spiders
Agelenopsis aperta Female Virgin vs. mated Direct Riechert & Singer (1995)
Schizocosa ocreata Female Virgin vs. mated Direct Roberts & Uetz (2005)
Insects
Cockroach
Nauphoeta cinerea Male Male aggressiveness Direct Moore & Moore (1999)
N. cinerea Male Male aggressiveness Direct Moore et al. (2001)
Orthoptera
Gryllus integer Male Dominant vs. subord. Indirect Kortet & Hedrick (2005)
Beetles
Dermestes maculatus Female Virgin vs. mated Direct McNamara et al. (2004)
Ips pini Male Predator avoidance Direct Teale et al. (1994)
Neopyrochroa flabellata Male Resource Direct Eisner et al. (1996)
Nicrophorus orbicollis Male Male size Direct Beeler, Rauter & Moore (2002)
Osmoderma eremitica Male Resource Direct Larsson et al. (2003)
Prostephanus truncatus Male – – Birkinshaw & Smith (2001)
Tenebrio molitor Male Male quality (Imm.) Indirect Rantala et al. (2002)
T. molitor Male Male quality (Imm.) Indirect Rantala et al. (2003a)
T. molitor Female Virgin vs. mated Direct Carazo et al. (2004)
Tribolium castaneum Male Male quality Indirect Lewis & Austad (1994)
Scorpionfly
Panorpa japonica Male Male quality (FA) Indirect Thornhill (1992)
Hymenoptera
Andrena nigroaenea Female Virgin vs. mated Direct Schiestl & Ayasse (2000)
Lasioglossum figueresi Female Fertility/relatedness Dir./ind. Wcislo (1992)
Diptera
Ceratitis capitata Male Male quality – Shelly & Kennelly (2003)
Drosophila grimshawi Male Male quality Dir./ind. Droney & Hock (1998)
Lutzomyia longipalpis Male Male attractiveness Fisherian Jones et al. (1998)
Lepidopteran
Utetheisa ornatrix Male Resource Direct Dussourd et al. (1991)
U. ornatrix Male Male quality/size Indirect Iyengar et al. (2001)
Crustacea
Alpheus angulatus Female Female receptivity Direct Mathews (2003)
Homarus americanus Male Dominant vs. subord. Indirect Bushmann & Atema (2000)
Rhynchocinetes typus Male Dominant vs. subord. Indirect Diaz & Thiel (2004)
Tigriopus japonicus Female Female age Direct Ting et al. (2000)
Fishes
Gasterosteus aculeatus Male Male quality (MHC) Indirect Reusch et al. (2001)
Poecilia reticulata Male Male quality Dir./ind. Shohet & Watt (2004)
Amphibia
Plethodon vehiculum Both Mate size Dir./Ind. Marco et al. (1998)
Reptiles
Lacerta monticola Male Male quality (FA) Indirect Martı́n & López (2000)
L. monticola Male Male age Dir./Ind. López et al. (2003)
Podarcis hispanica Female Pregnant vs. non-pregn. Direct Cooper & Pèrez-Mellado (2002)
P. hispanica Male Male quality (Imm.) Indirect López & Martı́n (2005)
Thamnophis sirtalis Female Female size Direct Lemaster & Mason (2002)
T. sirtalis Female Virgin vs. mated Direct O’Donnell et al. (2004)
Mammals
Homo sapiens Male Male quality (FA) Indirect Thornhill & Gangestad (1999)
H. sapiens Male Male quality (MHC) Indirect Wedekind et al. (1995)
Lemmus sibiricus Female Virgin vs. mated Direct Huck et al. (1984)
Mus musculus Male Dominant vs. subord. Dir./ind. Drickamer (1992)
M. musculus Male Male infection status Direct Kavaliers & Colwell (1995)
Chemical communication in mate choice 271
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
In species and mate recognition signalling, both signallers
and receivers usually share the same interest - the
coordination of mating in time and space - and there is
thus no need for specific mechanismsto ensure the honesty
of the signal. When advertising mate quality however, there
is ample incentive for deceit, since signallers have a general
interest in communicating a greater quality than they
actually possess. The most commonly proposed mechanism
for ensuring the honesty of a signal is costliness, and mate
assessment signals are consequently thought to be costly to
produce or maintain (Zahavi, 1975, 1977). Many sexual
signals are assumed to be costly; however, to ensure the
honesty of the signal, the associated cost must reduce the
signallers’ fitness. It is thus important to distinguish between
cost and expenditure, the latter being costs of time or
energy that do not affect the signallers’ longevity or
reproductive success (Kotiaho, 2001). Furthermore, honest
signals require that the cost is condition dependent, so that
individuals in good condition are better able to afford
exaggerations of the signal. Finally, the marginal cost of the
signal must be greater for individuals in poorer condition to
ensure that the individuals releasing the most attractive
signals also incur the lowest relative costs (Grafen, 1990;
Rowe & Houle, 1996). In spite of the many studies
exploring the costs of sexual signals, these assumptions
have rarely been met (Kotiaho, 2001).
A key limitation in the exploration of pheromones in
mate assessment was that the cost of producing pheromones
was regarded to be low (Alberts, 1992; Cardé & Baker,
1984) and thus it was assumed that pheromones were
unlikely to act as honest signals in mate choice. There are
considerable difficulties in measuring the metabolic costs of
pheromone production, and the few studies that have made
an attempt are based on measurements of the amount of
pheromone released during a specified time period and the
body mass of the species in question. For example, it is
estimated that an average male bark beetle, Ips typographus,
releases the equivalent of 2 % of his body mass in just one of
the components of his pheromone blend over a 24 h period
during excavation of his nuptial chamber (Schlyter &
Birgersson, 1989). However, bark beetles release unusually
large amounts of pheromones compared to most insects,
perhaps partly because a large part of bark beetle
pheromones are not synthesised de novo but extracted from
the host tree (Landolt & Phillips, 1997) which might reduce
metabolic costs. By comparison, the approximate lifetime
cost of pheromone production for a male boll weevil,
Anthonomus grandis, is 0.2 % of his body mass (Hedin et al.,
1974). A search for more general costs of pheromone
production needs to account for the amount of resources
and energy needed for synthesising small amounts of
chemicals. More precise estimates of the metabolic costs of
pheromone production in insects should be quite feasible
since the biochemical pathways used for synthesising
pheromones are known for several species (see Jurenka,
2004, for a recent review). In unicellar organisms, these
costs alone might be proportionally greater and possibly
substantial enough to ensure that the pheromone is
a reliable, honest signal (Nahon et al., 1995; Pagel, 1993).
Animals that use larger molecules as chemical signals, such
as the proteins found in the urine of mice (Gosling &
Roberts, 2001; Hurst et al., 2001) and femoral gland
secretions of lizards (Alberts, 1993), may also invest
considerable amounts of resources in their synthesis.
There could be indirect social costs to pheromone
production, as has been proposed for the American lobster,
Homarus americanus. In this species, females visit male shelters
for mating and protection during the moulting period.
Dominant males are located by pheromones in the males’
urine and are preferred by females as they are more likely to
provide protection (Bushmann & Atema, 2000). However,
this pheromone also attracts other males, resulting in
potentially costly antagonistic encounters as males compete
for access to females, thus ensuring that only dominant
males can afford to produce and maintain the signal. The
territorial scent marks of many mammals may have a similar
function (Gosling, 1990; Gosling & Roberts, 2001). Because
they increase the number of intraspecific aggressive
encounters a male incurs, scent marks allow prospective
mates to gather information about the resource-holding
capabilities of the territory owner (see Gosling & Roberts,
2001, for a comprehensive review). Most male mammals
scent-mark, particularly when holding a territory, and in
rabbits Oryctolagus cuniculus (Reece-Engel, 1988), hamsters
Mesocricetus auratus (Steel, 1984) and mice Mus musculus (Rich
& Hurst, 1998) females choose the marks of males that they
have smelled previously, thus indicating a preference for
dominant and/or resource-holding males.
It has been suggested that the increased risk of predation
on sexually displaying individuals infers a cost on the
display that only individuals in good phenotypic condition
can afford (Zuk & Kolluru, 1998). In the context of
chemical signalling, cues produced by one species may also
be exploited by other species (for a recent review see Zuk &
Kolluru, 1998). For example, sex pheromones produced by
bark beetles (genus Ips) attract not only mates but also
Table 2 (cont.)
Taxon Sex Message Benefit Reference
M. musculus Male Male condition Dir./ind. Meikle et al. (1995)
M. musculus Male Male quality (MHC) Indirect Penn & Potts (1998c)
M. musculus Male Good genes and MHC Indirect Roberts & Gosling (2003)
M. musculus Male Male infection status Indirect Zala et al. (2004)
Nycticebus pygmaeus Male Competitive ability Dir./ind. Fisher et al. (2003)
1 Two mating types.
Björn G. Johansson and Therésa M. Jones272
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
Table 3. Studies investigating individual variation in pheromone release. Sex is the emitting sex. Variable is what was measured in
the study: absolute amount indicates that the total amount of substance(s) from the animal was measured; proportion that the ratio
between different isomers/components was measured; amount of isomer(s)/component(s) that one or more of the pheromonal
constituents was measured; kind of components/number of components that the measurements was qualitative. %C.V. is the
coefficient of variation, calculated as the standard deviation in per cent of the mean (%C.V. ¼ (S.D./mean)100). Rep. indicates
repeatability: X indicates if any trial on the individual constancy of the signal was included in the study
Species Sex Variable %C.V. Rep. Reference
Moths
Adoxophyes sp. Female Absolute amount 90 – Kou & Chow (1991)
Proportion 12 –
Agrotis segetum Female Absolute amount 73 – Löfstedt et al. (1985)
Proportion 32 –
Female Absolute amount 156 – Svensson et al. (1997)
Amount of components 90-202 X
Proportion 26-59 X
Argyrotaenia velutinana Female Amount of isomers 58 – Miller & Roelofs (1980)
Proportion 20 –
Cacoecimorpha pronubana Female Proportion -2 X Witzgall & Frerot (1989)
Choristoneura fumiferana Female Amount of isomers 83 – Morse et al. (1982)
Colias eurytheme Male Absolute amount 56-80 – Sappington & Taylor (1990)
Amount of components 69-126 –
Proportion 30-1806 –
Heliothus virescens Female Amount of isomers 70-101 – Pope et al. (1982)
Proportion 5-56 –
Pectinophora gossypiella Female Amount of isomers 52 – Haynes et al. (1984)
Proportion 8 –
Female Amount of isomers 59 – Collins & Cardé (1985)
Proportion 5 –
Phthorimaea operculella Female Absolute amount 45-51 – Ono et al. (1990)
Proportion 23-31 –
Planotortrix excessana Female Absolute amount 111 – Foster et al. (1989)
Proportion 45 –
Platyptilia carduiactyla Female Amount of isomer 104 – Haynes et al. (1983)
Pseudaletia unipuncta Male Absolute amount 30 – Fitzpatrick et al. (1985)
Yponomeuta spp. Female Absolute amount 46-55 X Du et al. (1987)
Beetles
Dendroctonus ponderosae Female Amount of isomer 143 – Borden et al. (1986)
Male Amount of component 114 – Hunt et al. (1986)
Proportionof isomer 1 –
Ips cembrae Male Proportion 64-140 – Zhang et al. (2000)
I. pini Male Amount of component 111 – Borden et al. (1986)
Proportion of isomer 10 –
I. typographus Male Amount of components 85-191 – Birgersson et al. (1984)
Proportion 16 –
Male Amount of components 91-130 – Birgersson et al. (1988)
Proportion 15-21 –
Prostephanus truncatus Male Amount of components 58-75 X Hodges et al. (2002)
Proportion 12 X
Heteropteran
Nezara viridula Male Proportion of isomers 20-23 – Ryan et al. (1995)
Male Proportion 14-169 X Miklas et al. (2000)
Tick
Amblyomma variegatum Male Amount of components -2 – Pavis & Barre (1993)
Proportion -2 –
Mammals
Callithrix jacchus1 Female Kind of components -2 – Smith et al. (2001)
Proportion -2 –
Cervidae spp.1 Both Kind of components -2 – Lawson et al. (2000)
Crocidura russula1 Male Kind of components -2 – Cantoni et al. (1996)
Number of components 38 –
Dama dama1 Both Kind of components -2 X Lawson et al. (2000)
Herpestes auropunctatus1 Both Proportion 13-113 – Gorman (1976)
Chemical communication in mate choice 273
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
a range of predators and parasitoids (Dixon & Payne, 1980).
Similarly, yellowjacket wasps, Vespula germanica, locate males
of the Mediterranean fruit-fly, Ceratitis capitata, through
olfactory cues alone, and signalling males are consequently
subject to more attacks than either non-signalling males or
females (Hendrichs & Hendrichs, 1998). Recent experi-
mental manipulations reveal that the chemical signals of
vertebrates may also be vulnerable, although the exploita-
tion is visual rather than odour-based. Mammal scent
marks are visible in the ultraviolet spectrum and have been
implicated in prey location by a number of predatory bird
species (Kellie, Dain & Banks, 2004; Koivula & Korpimäki,
2001; Probst, Pavlicev & Viitala, 2002). The direct cost to
the signaller from increased predation risk has not been
measured, however one study in insects (Prostephanus
truncatus; Birkinshaw & Smith, 2001) and one in mammals
(Mus musculus; Roberts et al., 2001) suggest a reduction in
chemical signalling with increased risk of predation.
An alternative way to estimate the costs of pheromone
production is to explore the trade-off between signalling
and fitness parameters such as survival and reproductive
success. A number of recent laboratory studies have
explored the relationships between condition, signalling
frequency and fitness. A study investigating the role of
natural and sexual selection on the evolution of cuticular
hydrocarbons (CHCs) in hybrids of the two Drosophila
species D. birchii and D. serrata revealed substantial fitness
costs of CHC production (Blows, 2002). In a breeding
experiment that permitted natural selection but not sexual
selection, the total amount of CHCs decreased in both
males and females, indicating a trade-off between CHC
production and reproduction. By contrast, when the
constraints on sexual selection were relaxed, the composi-
tion of the pheromone changed in males but not in females,
indicating female preference for males with particular CHC
profiles (Blows, 2002). A field-based study highlighted the
costliness of male CHCs in D. serrata; the relative
concentrations of three CHCs were positively correlated
with male body size indicating condition dependence of
these characters (Hine, Chenoweth & Blows, 2004). The
trade-off may be expressed only when signal expression is at
its upper limit and/or when it is condition dependent. In
mice, Mus musculus, small males were able to maintain
dominance over larger males by increasing their investment
in scent-marking, but the increased rate of signalling had an
associated cost in terms of reduced growth rate and body
size (Gosling et al., 2000). The cost of signalling in mice is
likely to be mediated by the production of major urinary
proteins (MUPs) found in male scent marks (Gosling &
Roberts, 2001; Hurst et al., 2001). However, the costs of
signalling, in terms of decreased growth rates, are expressed
only when the signal rate is increased substantially. In
a subsequent experiment, increased levels of scent marking
did not affect growth rates, although they were associated
with enlarged urogenital (pheromone producing) glands
(Collins et al., 2001). The authors conclude that the
magnitude of increase was not sufficient to cause mice to
trade-off signal rate with growth.
Condition-dependent expression of pheromones has also
been demonstrated in the grain beetle, Tenebrio molitor
(Rantala et al., 2002, 2003a). In a series of laboratory trials,
these authors demonstrated that both male attractiveness
(which is pheromone based) and immunocompetence
(measured as encapsulation rate) were correlated positively
with male condition. Similarly, females of the Iberian wall
lizard, Podarcis hispanica, prefer male femoral gland excre-
tions containing a high proportion of a specific steroid, and
males with a high proportion of this substance show
a greater T-cell-mediated immune response (López &
Martı́n, 2005). Likewise, in the cockroach, Nauphoeta cinerea,
both quantity and the relative proportions between three
components of the male pheromone have been shown to
vary with condition (Clark, DeBano & Moore, 1997).
However, there was no assessment of the costs of
pheromone production in these studies. By contrast, in
the Hawaiian picture-wing fruit fly Drosophila grimshawi male
pheromone production is both condition dependent
(Droney & Hock, 1998) and leads to reduced survival
(Johansson, Jones & Widemo, 2005).
(3) Additive genetic variance
In a review of the quantitative genetic literature, Pomian-
kowski & Møller (1995) found that the level of additive
genetic variation was generally higher in sexually selected
traits compared to non-sexually selected traits. The reason
for this might be that directional sexual selection favours an
increase in the number of genes, such that sexually selected
traits may be more polygenic than other traits (Pomian-
kowski & Møller, 1995). Alternatively, Rowe & Houle (1996)
suggested that sexually selected traits often evolve to be
costly and their expression should thus reflect phenotypic
condition. Because phenotypic condition is influenced by
a large proportion of the genome, sexually selected traits
should exhibit high levels of genetic variance (the genic
capture hypothesis; Rowe & Houle, 1996). Given that the
amount of resources available for allocation to traits is
potentially limited, if condition dependent, we also expect
Table 3 (cont.)
Species Sex Variable %C.V. Rep. Reference
Meles meles1 Both Kind of components -2 X Buesching et al. (2002b)
Number of components 24 X
Mustela vison1 Both Kind of components -2 X Brinck et al. (1978)
1 No specific sexual function of pheromone proposed.
2 Not possible to calculate.
Björn G. Johansson and Therésa M. Jones274
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
the expression of sexually selected traits to compete with
that of other traits and we may see a trade-off with certain
life-history traits (Tomkins et al., 2004).
Since the evolutionary potential of traits is determined to
a large extent by their level of additive genetic variation, the
heritability (i.e. the ratio between the additive genetic
variance of the character and its total phenotypic variance,
see Falconer & Mackay, 1996) of phenotypic traits is a key
criterion in evolutionary studies. Although heritability has
been criticized as an index for comparing additive genetic
variance (e.g. by Pomiankowski & Møller, 1995), we employ
it here, as it is still the most commonly used. Low levels of
heritability are generally expected for characters closely
linked with fitness, since genes coding for such traits are
predicted to reach fixation in populations (Falconer &
Mackay, 1996). In the context of sexual signalling,
pheromonal traits shouldbe subject to stabilising selection
if they function mainly as mate or species recognition
signals (Löfstedt, 1993). If this condition holds, the
heritability of pheromonal traits would be expected to be
relatively low. By contrast, if such traits are used as
indicators of mate quality, heritability might be high as
they are subject to trade-offs with traits that are more
important to immediate survival.
There are relatively few studies that focus on the
heritability of pheromonal traits, but some studies on insects
have shown that pheromonal blend/composition generally
exhibits moderate to high levels of heritability (see Table 4).
This is perhaps surprising, considering the expected low
heritability of species recognition signals, and since most of
the studies are done on female moth pheromones that are
generally assumed to function primarily in species recogni-
tion. However, the number of studies is still too small to
allow more specific conclusions. The studies outlined in
Table 4 have primarily been concerned with qualitative
aspects, that is, the composition of the pheromones in
question, whereas quantitative variation has received less
interest. Since species recognition in general is presumed to
be encoded by qualitative aspects of the pheromone, the
amount of pheromone released might be a more promising
trait to study when looking for mate quality indicators.
An alternative indicator of the upper limit of the
heritability of a character is to estimate the repeatability
of the signal from consecutive measurements of the same
individual (Boake, 1989). There are several studies on the
repeatability of pheromone emission across insects and
mammals (Table 3). In insects, the ratio between different
components generally shows high repeatability, whereas the
quantities released are more likely to vary over time. For
example, individual female turnip moths, Agrotis segetum,
vary extensively in the absolute and relative amounts of the
three major components in their pheromone. The repeat-
ability of the ratio between two of the components released
was higher than that of the released amounts (Svensson,
Bengtsson & Löfqvist, 1997). In the bostrichid beetle
Prostephanus truncatus males vary both in the amount of the
two components within their aggregation pheromone and
in the ratio of the components to one another. These
differences between males were repeatable over a two-week
period (Hodges, Birkinshaw & Farman, 2002). Mammal
studies indicate high repeatability for the kind of compo-
nents in the individual blend, although proportions and
quantity may vary with season or reproductive status of the
individual. For example, Lawson et al. (2000) studied
individual signatures in the scent gland secretions of
a number of deer species. Their study revealed that odour
signatures were not only individually distinct but also, at
least in the fallow deer, Dama dama, consistent across years.
Apart from hinting at the heritability of various aspects of
pheromones, these studies confirm the consistency of
individual variation in pheromones.
IV. CONSEQUENCES OF MATE ASSESSMENT
A field of chemical communication that has received com-
paratively little attention is that exploring the mechanisms
Table 4. Estimated heritability of aspects of pheromones. Sex is the emitting sex. h2 R is the heritability of the blend ratio of the
pheromone. h2 A is the heritability of the amount of pheromone. Partly adapted from Svensson et al. (2002)
Species Sex h2 R h2 A Reference
Lepidoptera
Agrotis ipsilon Female 0.56–1.09 0.56–1.15 Gemeno et al. (2000)
Argyrotaenia velutinana Female 0.41–0.64 – Sreng et al. (1989)
Colias eurytheme Male 0.05–0.89 0.20–0.89 Sappington & Taylor (1990)
Pectinophora gossypiella Female 0.50 – Collins & Cardé (1989)
– 0.71 Collins et al. (1990)
Plodia interpunctella Female 0.65 – Svensson et al. (2002)
Trichoplusia ni Female – 0.20–1.13 Gemeno et al. (2001)
Beetle
Ips pini Male 0.95 – Hager & Teale (1996)
Fruit fly
Drosophila serrata Male 0.06–0.53 – Skroblin & Blows (2006)
Cockroach
Nauphoeta cinerea Male 0.17–0.76 0.32–0.72 Moore (1997)
Chemical communication in mate choice 275
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
promoting the evolution of a signal and, perhaps more
importantly, the mechanisms that maintain a signal used in
mate assessment (Andersson, 1994). In this section we
review the current evidence for pheromones as direct or
indirect signals used in mate assessment (see Table 2). Here,
we concentrate on research that has tested not only the
physiological, and/or neurological response of a species to
a particular chemical signal, but also the consequences of
pheromone-based choice.
For a sexual signal to be adaptive, it must convey a benefit
to the receiver (Andersson, 1994). In mating systems where
individuals obtain immediate benefits from their choice of
mate (for example nuptial gifts, territories or oviposition
sites), the benefit of a particular preference is relatively
intuitive. However, in species where choosy individuals gain
nothing more than the genetic material for the production
of offspring, such as in extra-pair copulations or in lek
mating systems, the reasons for choice are less obvious
(Höglund & Alatalo, 1995). Traditionally, models proposed
to explain the evolution of mate choice explore the
consequences of mate preferences through the direct or
indirect gains that an individual accrues (Andersson, 1994).
While the predictions of benefit models are mutually non-
exclusive (Kokko et al., 2003), here we consider the evidence
for each model in turn and indicate where additional
benefits may be accrued.
(1) Pheromones as direct benefits or indicators
thereof
Direct benefits models predict that choosy individuals gain
immediate benefits from their choice of mate through
increased survival and/or fertilisation success (Kirkpatrick
& Ryan, 1991; Price, Schluter & Heckman, 1993). As with
other sensory modalities, such as vision or sound, odours
may be correlated with the ability of an individual to
provision their mate with a resource such as a territory or
oviposition site. Alternatively they may indicate the fertility
of a potential mate. In addition, a unique feature of
pheromones is that, because they consist of matter (rather
than photons or sound waves), they may also act directly as
the resource.
(a ) Pheromones as a resource
Many insects that rely on plant-derived substances as
a source of protection against predators use these chemicals,
or their derivatives, as pheromones in mate attraction
(Gullan & Cranston, 1994; Landolt & Phillips, 1997).
Pyrrolizidine alkaloids, an intensely bitter-tasting group of
chemicals found in plant families such as Asteracae and
Fabacae, are utilised by a wide range of lepidopterans such
as danaid butterflies (Dussourd et al., 1989), arctiid moths
(Dussourd et al., 1991; Vonnickischrosenegk & Wink, 1994)
and possibly ithomiine butterflies (Boppré, 1990). Resource-
based pheromones are also found in the pyrochroid beetle
Neopyrochroa flabellata (Eisner et al., 1996) and the chrysomelid
beetle Diabrotica undecimpunctata (Tallamy, Gorski & Burzon,
2000). In the above species, males incorporate sequestered
chemicals into their spermatophores, but also use a pro-
portion of the sequestered chemicals as pheromones for
mate attraction. Females discriminate between males based
on the pheromones advertised and preferentially mate with
males emitting the largest chemical signal. Females gain
a direct benefit from their choice as the vast majority of the
protective chemicals that are transmitted from the male to
the female via the spermatophore during mating are either
utilised by the female or incorporated into the fertilised eggs
reducing their risk of predation. On average, female D.
undecimpunctata beetles sequester approximately 10% of the
cucurbitacins for themselves; incorporate 80% into their
eggs and excrete the final 10% (Tallamy et al.,2000). The
benefits of choice are such that female D. undecimpunctata
beetles mate just once, but may sample the pheromone of
up to 15 males prior to accepting a cucurbitacin-loaded
spermatophore. Similar systems seem to operate in all
insects using resource-based pheromones, the most well
investigated being the arctiid moth Utetheisa ornatrix that has
been subject to detailed studies for over two decades (e.g.
Conner et al., 1981, 1990; Eisner et al., 2000).
Protective chemicals that are utilised by both males and
females represent a valuable resource that is desirable for
both sexes. If the signaller does not transfer what is
advertised, this creates the potential for considerable
conflict. In the case of the arctiid moth, the quantity of
pheromone produced by the male is proportional to the
amount of alkaloids transmitted to the female during
mating, therefore serving as an honest signal that can be
used to discriminate between males (Eisner & Meinwald,
1987, 1995). However, males may use the pheromone signal
to exploit a pre-existing sensory bias of the female. In the
German cockroach, Blatella germanica, males lure females
into a pre-copulatory position by exploiting a feeding
response cued by male glandular secretions (Nojima et al.,
1999a, b). These glandular secretions act as strong phago-
stimulants in cockroaches of both sexes, as well as in the
immature nymphs, suggesting that males are exploiting the
innate feeding response of females. Resolution of the con-
flict is achieved because while mating always seems to be
preceded by female feeding on the male secretions, the
feeding response of females is not always followed by
mating.
(b ) Pheromones as indicators of a resource
For the majority of species, the chemical signal is merely an
indicator of a potential resource. Pheromones that advertise
an individual’s dominance may be indicators of the
potential for that individual to protect or provide for the
choosing sex. However, there are few explicit tests of this
idea. One possible example where a male pheromone may
be used as an indicator of parental ability is in the black
goby, Gobius niger. In this fish, males that build and guard
nests produce more of a pheromone attractive to females than
males adopting a ‘‘sneaking’’ mating strategy (Locatello,
Mazzoldi & Rasotto, 2002). An individual that uses che-
mical signals to discriminate between dominant and sub-
ordinate males may also accrue direct benefits through
reduced harassment. A number of empirical studies on
Björn G. Johansson and Therésa M. Jones276
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
vertebrates and invertebrates provide support for this idea.
In the house mouse, Mus domesticus, females prefer territories
with well-protected nest areas, but also use scent marks as
cues that signal the ability of the male to defend his nest site
from rivals (Rich & Hurst, 1998, 1999). Similar patterns are
found in invertebrates. In the American lobster, Homarus
americanus, males fight for access to large, established holes
that are preferred by females (Atema, 1986, 1995). Females
prefer the chemical signals of the territory-holding males
who are larger and better able to defend their territory.
Finally, in the cockroach Nauphoeta cinerea male pheromones
signal social status to both males and females (Moore, 1988;
Moore & Breed, 1986), but the pheromone blends preferred
by females are different to those indicating the socially
dominant male (Moore et al., 2001). Choosy females use
chemical signals to avoid mating with aggressive, manipu-
lative males and benefit from their choice through increased
survival and offspring production (Moore, Gowaty &
Moore, 2003). In many of these species, the choosy sex
may also gain indirect genetic benefits from their choice, or
accrue direct fertility benefits.
(c ) Pheromones as indicators of fertility
The number of gametes provided by a mating partner can
limit the reproductive potential of females (e.g. Matthews,
Evans & Magurran, 1997; Royer & McNeil, 1993; Sakurai,
1996; Sheldon, 1994) as well as males (Côté & Hunte, 1989;
Olsson, 1993; Verrell, 1995). Functional fertility may vary
with environmental factors such as food availability
(Kvarnemo & Simmons, 1999; McLain, 1998) or disease
(Dufva, 1996; Gustafsson et al., 1994) and such variation is
likely to be reflected in sexually selected phenotypic
characters (Sheldon, 1994). A direct link between phero-
mone productivity and fertility is difficult to demonstrate.
However, mate assessment pheromones do correlate with
fertility-related characters such as body size, age or
reproductive status.
The relationship between body size and fertility is likely
to be stronger for females than for males. This is based on
the idea that female reproductive success is limited, in part,
by the number of gametes she is able to produce, which is
often tightly correlated with body size (Trivers, 1972). In the
salamander species Plethodon vehiculum and P. dunni, clutch
size is related to female body size, and males choose large
females based on odour cues alone (Marco et al., 1998).
Females of P. vehiculum are also able to distinguish between
the chemistry of males of different sizes, although the
benefits of such discrimination are less clear (Marco et al.,
1998). A similar pattern is found in the Iberian rock lizard,
Lacerta monticola. In this species, males use chemical signals
to distinguish between females, preferentially mating with
larger females. Females also use chemical signals to
discriminate against younger males in preference for older
individuals even though males also vary visually with age
(López, Aragon & Martı́n, 2003). The authors suggest that,
because there is a certain degree of overlap in coloration
between male age classes, pheromones provide a more
accurate cue for male age, a trait that may relate to male
size and possibly dominance. Evidence from other taxa is
suggestive rather than tested. In the winter moth, Operophtera
brumata, males preferentially mate with larger females. The
laboratory experiments did not measure female pheromone
emissions, but as males do not respond to visual and tactile
female stimuli, the authors suggest that male choice is most
likely based on either the quality or quantity of the
pheromones produced by the females (Van Dongen et al.,
1998). However, as several studies have failed to find any
correlation between female body mass and pheromonal
characteristics in moths (Svensson, 1996) this assumption
requires further investigation.
Age-dependent chemical signals have the potential to be
used as indicators of an individual’s fertility. To our
knowledge no study has tested this explicitly, however
several lines of evidence suggest that this may be the case. In
the oblique-banded leafroller, Choristoneura rosaceana, female
mating success declines linearly with age while male mating
success increases for the first few days following adult
emergence and then declines (Delisle & Royer, 1994). There
is a negative correlation between the quantity of active
components in pheromone emissions and age in females
(Delisle & Royer, 1994). Older females compensate for the
decrease in attractiveness by calling earlier and for longer
than younger females (Delisle, 1995). Unfortunately, the
relationship between pheromone emissions and fertility was
not measured in this study. Evidence for age-related
declines in male fertilisation is increasing (for a recent
review see Brooks & Kemp, 2001). In the lekking sandfly
Lutzomyia longipalpis females prefer to mate with intermediate-
aged males and accrue direct fertilisation benefits from their
choice (Jones, Balmford & Quinnell, 2000). In the hide
beetle, Dermestes maculatus, intermediate-aged males have
higher mating success than young or old males and are
more fertile (Jones & Elgar, 2004). While neither study has
tested explicitly whether females use chemicals to discrim-
inate between differentmale age-classes, several lines of
evidence suggest that this is the case: in the hide beetles,
males produce a pheromone from an abdominal gland that
is known to elicit a response in females (Abdel-Kader &
Barak, 1979); in L. longipalpis, there is a strong correlation
between female mating preferences and the amount of
pheromone contained in a male’s tergal gland (Jones &
Hamilton, 1998). As all reproducing individuals have been
young at some time, and there consequently can be no
individual variation to select in this particular character,
many of these pheromones are probably best understood as
mate recognition signals. However, mate assessment signals
are likely to be susceptible to senescence on account of their
condition dependency, making it at least possible that they
could act as indicators of age.
The fertility benefits of pheromone-based mate choice
are not always clear. In the Hawaiian fruit-fly Drosophila
grimshawi females are attracted to leks by male pheromones
that are deposited on leaves or twigs. Males that invest most
in condition-dependent traits such as pheromone pro-
duction and courtship intensity have the highest mating
success (Droney & Hock, 1998). Recent evidence suggests
that the relationship between signal rate and fertility is
negative rather than positive. Females mating with males
who court at the highest rate produce fewer offspring than
Chemical communication in mate choice 277
Biological Reviews 82 (2007) 265–289 � 2007 The Authors Journal compilation � 2007 Cambridge Philosophical Society
females whose mates courted at a lower rate (Droney, 2003).
Pheromone production is correlated with courtship activity
but was not measured in this study. One possible
explanation for this result is that courtship activity is an
indicator of indirect rather than direct benefits (Droney,
2003). Alternatively, increased egg production may be costly
(Chapman et al., 1995), thus the observed variation in
female fecundity may in fact be the result of conflict
between the sexes with respect to oviposition rates. If this
holds, female D. grimshawi may gain survival benefits by
mating with males that invest most in courtship and
pheromone production as they inflict least harm on females.
Discriminating between these ideas requires knowledge of
fitness parameters such as survival and lifetime fecundity.
(2) Pheromones as indicators of indirect
benefits
Indirect benefits models assume that discriminating indi-
viduals benefit from their choice of mate through the
increased fitness of their offspring. The models require that
the preference of the choosing individual and the preferred
trait of the chosen individual are heritable, and that they
covary across generations (Andersson, 1994; Lande, 1981).
Increasing evidence, from a range of taxa, suggests that
ornamental traits may be used as condition-dependent
signals by females to assess the potential quality of mates
and thereby gain genetic benefits (Andersson, 1994). These
‘‘good genes’’ models assume that a particular trait reflects
an innate quality of the male and on average all females
should prefer the same male. However, females may mate
for reasons other than to obtain ‘‘good genes’’ for their
offspring; they may mate to obtain the most compatible
genes (Trivers, 1972), or to increase the genetic diversity of
their offspring (Tregenza & Wedell, 2000). Compatibility
models assume that to increase the fitness of their offspring,
females choose mates that are optimally genetically
dissimilar to them rather than males that bear the largest
or most elaborate ornament (Mays & Hill, 2004; Tregenza
& Wedell, 2000). Alternatively, they may mate with the most
heterozygous males because these increase the genetic
diversity of their offspring. However, as males are unable to
pass their heterozygosity per se onto their offspring, the male
trait is non-heritable (Tregenza & Wedell, 2000). From the
point of sexual selection, this might be viewed as an extreme
form of frequency-dependent mate choice. Females may
also increase the genetic diversity of their offspring by
mating multiply thus ensuring many sires. Empirical
support for indirect benefits models has increased steadily
over the past three decades, and recent evidence suggests
that pheromones may provide a more reliable assessment of
genotype prior to mating than either visual or acoustic
signals (Tregenza & Wedell, 2000).
(a ) Pheromones as indicators of ‘‘good genes’’
There are a number of studies indicating good genes
benefits accrued through pheromone-based mate assess-
ment (Table 2). In the cockroach Nauphoeta cinerea, females
use chemical signals to select their mates (Moore & Breed,
1986). Females gain indirect benefits by mating with the
male of their choice through the production of offspring
that reach sexual maturity more rapidly (Moore, 1994). In
addition, discriminating N. cinerea females gain direct
benefits from their choice as the time period between
successive egg clutches is also reduced. In the tiger moth,
Utetheisa ornatrix, choosy females gain direct benefits through
the acquisition of resource-based pheromones (see above),
but they also gain genetic benefits from their choice because
large tiger moth males release more pheromone than small
males and size is heritable in both sexes. Choosy females
accrue good-genes benefits through the production of larger
daughters that are themselves more fecund and Fisherian
benefits (see Section IV.3 below) through the production of
larger sons with increased mating success (Iyengar & Eisner,
1999a, b). Similar indirect benefits promote mate choice in
the tobacco moth, Ephestia elutella, where females preferably
mate with large males that release more pheromone,
although the pheromone in this case is not resource based
(Phelan & Baker, 1986). (For more details on the Fisherian
benefits accrued in these moth species, see Section IV.3). In
the fruit fly Drosophila serrata, female preference for male
CHCs has been shown to be genetically correlated with
offspring fitness, indicating genetic benefits from female
mate choice (Hine et al., 2002). A good genes mechanism
is also proposed for the evolution of pheromone expression
in the unicellar brewers yeast, Saccharomyces cerevisiae (Nahon
et al., 1995). Mating in S. cerevisiae is mediated by a com-
bination of short peptides, which are relatively cheap to
produce, and large proteins that are suggested to have
a mate assessment function. The large proteins are thought
to act as honest indicators of quality because they provide
information about a cells’ ability to invest in resources and
utilise different biochemical pathways (Nahon et al., 1995).
However, an alternative view is that mate choice in S.
cerevisiae is based on the relative symmetry between the
quantities of chemicals released by both the two mates
(Pagel, 1993).
Pheromones may also reflect an individual’s symmetry,
a suggested (although somewhat controversial) indicator
of an individual’s quality (for a debate, see the Journal of
Evolutionary Biology 1997, vol 10, pp1-76). This idea is based
on the premise that symmetry is a reflection of an
individual’s ability to withstand environmental stress and
perturbation during development and is thus a potential
indicator of genetic quality (Jennions & Oakes, 1994;
Palmer & Strobeck, 1986). Several studies report a correla-
tion between the symmetry of a particular trait and
chemical-based mate choice, and argue that pheromones
act as honest signals of phenotypic and/or genetic quality.
In the Japanese scorpionfly Panorpa japonica, females prefer
males with symmetrical fore wings, a choice that apparently
is guided by pheromones alone (Thornhill, 1992). Similarly,
female Iberian rock lizards, Lacerta monticola, prefer the scent
of males with a symmetrical (and higher) number of femoral
pores on the hind limbs (Martı́n & López, 2000). In
humans, Homo sapiens, females prefer the scent of males with
the highest degree of facial symmetry, and express the
strongest preference