Behavioral
Ecology
The official journal of the
ISBE
International Society for Behavioral Ecology
Behavioral Ecology (2014), 25(2), 409–414. doi:10.1093/beheco/art128
Original Article
Sexual signaling under predation: attractive
moths take the greater risks
Nils Cordes, Leif Engqvist, Tim Schmoll, and Klaus Reinhold
Department of Evolutionary Biology, Bielefeld University, Morgenbreede 45, 33615 Bielefeld, Germany
Received 12 August 2013; revised 9 December 2013; accepted 19 December 2013; Advance Access publication 29 January 2014.
Individuals often have to balance the costs of risky behavior against the potential benefits they may gain from it. This trade-off is especially obvious in the interplay between natural and sexual selection because traits for mate attraction may attract predators as well. In
the lesser wax moth Achroia grisella, ultrasonic sexual signaling comes with such a risk because calling for mates also attracts predatory bats. Attractive males should behave more cautiously than unattractive males under such circumstances, as they can expect
more future mating opportunities and therefore have more to lose (the so-called asset protection principle). Contrary to these predictions, we found that pulse pair rate and peak amplitude, 2 song components attractive to females, correlated negatively with the duration of a silence response which is displayed when courtship song is experimentally overlayed with the search signal of a predator,
the greater horseshoe bat. More attractive males thus recommence singing sooner than less attractive males. Although this is not in
line with the asset protection principle, we discuss 3 different ways in which these distinct behavioral differences might be explained:
1) attractive males are in better condition and thus can more easily evade predators, 2) attractive signals are costly in terms of reduced
life expectancy, and 3) risk-taking may in itself be a sexually selected trait and as such act as an honest signal of male quality.
Key words: Achroia grisella, asset protection, attractiveness, consistent behavioral differences, risk-taking.
Introduction
In most species exhibiting elaborate courtship and sexual signaling, males that are conspicuous enough to attract females often are
conspicuous enough to attract predators as well (Zuk and Kolluru
1998). Be it extravagant behavioral displays, bright colors, conspicuous odors, or attractive songs, sexual signaling often increases the
risk of being eaten (Ryan et al. 1982; Magnhagen 1991; Acharya
and McNeil 1998; Hedrick 2000; Kotiaho 2001), forcing individuals to trade current against future reproduction. Such is the case in
males of the lesser wax moth Achroia grisella (Lepidoptera; Pyralidae)
(Spangler 1984). Males gather in small leks on foliage or grass at
night and attract females with ultrasonic songs. Females choose the
individual with the most attractive song from the courting males
for mating (Spangler et al. 1984; Jang and Greenfield 1996). Song
attractiveness is determined by females based on 3 song traits:
pulse pair rate, peak song amplitude, and the asynchrony of wing
beat (Jang and Greenfield 1996, 1998). These courtship songs are,
however, conspicuous also to echolocating predatory bats, informing them of the presence and position of the prey (Schnitzler and
Ostwald 1983). The bat Rhinolophus ferrumequinum, for example,
has been shown to attack live wax moths as well as loudspeakers
broadcasting A. grisella courtship song (Alem et al. 2011). Male wax
Address correspondence to N. Cordes. E-mail: [email protected].
© The Author 2014. Published by Oxford University Press on behalf of
the International Society for Behavioral Ecology. All rights reserved. For
permissions, please e-mail: [email protected]
moths should therefore be selected to adjust their song in the presence of predators, for example, by ceasing acoustic signaling when
they are aware of adjacent bats.
Indeed, when signaling wax moths detect the search call of a
predatory bat, they respond with a pause from singing (henceforth
silence response) (Spangler 1984). Individuals differ repeatably in
the length of this silence response, which may last from a few hundred milliseconds to over a minute (Greenfield and Baker 2003;
Brunel-Pons et al. 2011; Cordes et al. 2013). There is evidence that
males recommence singing sooner if under competition for females
than if they sing alone (Brunel-Pons et al. 2011), suggesting that
individual males can adjust their silence response to different social
environments. The length of the silence response, however, can
be considered a measure of risk-taking: Moths which start singing
soon after the disturbance face greater risks of being detected by
the bat than moths which remain silent for longer (Greenfield and
Baker 2003).
Taking a risk means jeopardizing one’s assets (Clark 1994).
Often these assets can be an individual’s life or the chance for
future reproduction. The amount of risks taken should therefore
be higher for individuals with low compared with individuals with
high residual reproductive value, a concept termed the “asset protection principle” (Clark 1994), which has found support in a number of species (Candolin 1998; Kemp 2002), including A. grisella
(Lafaille et al. 2010). Wolf et al. (2007) showed that in theory such
differences in asset protection could lead to the evolution of distinct
Behavioral Ecology
410
behavioral phenotypes, i.e., animal personalities (Gosling and John
1999; Dall et al. 2004; Dingemanse and Wolf 2010). When courtship is performed by males in close vicinity, these males are in continuous competition with each other, so introducing a predator into
a system like a lek is a valuable opportunity to detect and investigate individual behavioral differences. When other males cease
courtship under the presence of a predator, opportunity arises for
those individuals which might otherwise not be noticeable or which
appear less attractive in direct comparison (Candolin 1997).
Surprisingly, there has been little research regarding attractiveness as a manifestation of asset value (Billing et al. 2007; see also
Schuett et al. 2010), but if attractive individuals can expect more
opportunities for future reproduction, less attractive individuals
should take higher risks as they have less to lose, in line with the
asset protection principle. To explore the interplay between sexual
signaling and predator avoidance, we tested for correlations between
risk-taking behavior and the expression of 2 song traits attractive to
females (pulse pair rate and peak amplitude). The analyzed data
were obtained from an experiment for which larvae were reared
under different densities to explore the effects of the larval environment on risk-taking behavior (cf. Cordes et al. 2013). We use this
to our advantage by probing how robust the correlations between
risk-taking and courtship display are across 3 different rearing environments. According to the asset protection principle, we expected
that males whose song trait values suggest high sexual attractiveness
would behave more cautiously than less attractive males when confronted with a predator search call. In pronounced contrast to this
prediction, we here show that more attractive males behave more,
not less risk-prone than their less attractive conspecifics.
Methods
We used a laboratory population of A. grisella, a short-lived, social
parasite of honey bees (Kunike 1930) and reared larvae from different families (i.e., full-sib offspring from 1 pair) at 3 different densities (1, 2, or 6 larvae per 30-ml plastic cup) with ad libitum food to
examine the effect of larval density on song quality and correlated
behaviors (see Cordes et al. 2013 for detailed information on study
populations, rearing methods, and diet). On eclosion, adult moths
were weighed, and body mass was measured to the nearest 0.01 mg
with a Kern 770 scale (Kern & Sohn GmbH, Balingen, Germany).
Experimental setup
Adult moths were 24–48 h old when measurements started. Before
acoustic measurements, we transferred all males individually into
gauze cages (3 cm diameter, 5 cm height) and kept them in shelves
covered with insulation foam. To prevent males from hearing the
ultrasonic songs of other males during the experiment, we placed
the gauze cages into insulated enclosures which opened only to the
front. Achroia grisella males sing continuously throughout the night
(scotophase). We therefore played back a recording of the search
signal of the greater horseshoe bat R. ferrumequinum, a potential
predator of A. grisella whose call specifically causes lesser wax moths
to intermit their behavior (Greenfield and Weber 2000; Alem et al.
2011; Cordes et al. 2013), for 3 trials: at the onset of scotophase,
again 1 h later and at the onset of the subsequent scotophase. For
this, we used the software Avisoft RECORDER USGH (Avisoft
Bioacoustics, Berlin, Germany). The search call was transferred
through an analogue/digital converter UltraSoundGate Player 116
(Avisoft Bioacoustics) and broadcasted by an ultrasonic speaker
ScanSpeak from Avisoft Bioacoustics (frequency range [±12 dB]:
60–110 kHz). Before and during each playback, we recorded the
courtship song of A. grisella using a condenser microphone (Brüel
& Kjær, Nærum, Denmark; Type 4939, frequency response [±2
dB]: 0.004–100 kHz) in a distance of 10 cm from the cage, and a
Nexus Conditioning Amplifier (Brüel & Kjær, Type 2690), which
transferred the sound through an analogue/digital converter
UltraSoundGate 116 (Avisoft Bioacoustics) to a computer. The
song was then recorded using the software Avisoft RECORDER
USG (Avisoft Bioacoustics).
Recordings started 20 s before the bat call playback and were
terminated 1 min after the bat signal was presented, even if moths
had not recommenced signaling. In total, 6 males did not resume
signaling in at least one of the playback trials (one did not recommence its song in all 3 trials); those measurements were considered as reflecting a 60-s duration of the silence response. For each
of the 3 recordings, we measured pulse pair rate (the number of
pulse pairs per second), peak amplitude (in V), and the duration
of the silence response to the bat call (with a precision of 10 ms).
Visualization of the song and measurements were performed with
the sound analysis application Avisoft SASLab Pro (Version 5.1.14,
Avisoft Bioacoustics). We also calculated a compound attractiveness
score, which was defined as the sum of normalized and standardized pulse pair rate and peak amplitude, based on linear selection
gradients for A. grisella, as determined by Jang and Greenfield (1998)
and following the procedure proposed by Brandt and Greenfield
(2004).
Statistical analysis
All statistical analyses were performed with the software package
R, version 2.13.0 (R Development Core Team 2011). Prior to the
analyses, we log transformed duration of the silence response (plus
1 ms) to approximate normality. In total, 171 individuals were used
for all analyses.
First, we separately tested for correlations between mean silence
response (per individual) and mean pulse pair rate, mean peak
amplitude, or mean attractiveness. We calculated Pearson’s product-moment correlation coefficients for all correlations for the full
sample and, separately, also for each of the 3 larval density treatments. We then fitted bivariate linear mixed effects models for each
correlation using Markov Chain Monte Carlo (MCMC) techniques
(implemented in the R package MCMCglmm, Hadfield 2010), which
allowed us to separate within- from between-individual effects for
each of the correlations (Dingemanse and Dochtermann 2013).
The MCMC chain was run for 1 300 000 iterations with a 300 000
iteration burn-in period and the remaining iterations sampled at
1000 iteration intervals. We estimated an unstructured two-by-two
variance–covariance matrix for 2 random effects (individual identity
and family identity) and the respective error variances. The priors
for the variances were set to half of the phenotypic variance, the
priors for the covariances were set to zero, and the degree of belief
parameter was set to nu = 1.002. We checked all models for convergence using the Gelman–Rubin diagnostic (Gelman and Rubin
1992). Potential scale reduction factors for all variables were below
1.05. Time series plots revealed no autocorrelation issues within
chains (Hadfield 2010). Repeatabilities for all 3 traits and correlation coefficients were calculated from the MCMC chains yielding
posterior distributions of parameter estimates and the 95% highest
posterior density (HPD) interval. We considered correlations significant when estimates’ 95% HPD intervals did not overlap with 0.
To determine whether larval density affected the strength of the
correlations, we fitted 3 separate models for each variable using the
Cordes et al. • Attractiveness and risk-taking in wax moths
same MCMC methods outlined above and compared betweenand within-individual correlation coefficients and 95% HPD intervals for all 3 treatment levels. To estimate the effect of larval density
on pulse pair rate, peak amplitude, attractiveness, and body mass,
we calculated separate linear mixed effects models (R function lmer
from the R package lme4, Bates 2007) with individual identity and
family identity as nested random effects and density treatment as
a fixed effect factor with 3 levels. We refrained from using adult
body mass as a covariate in addition to the treatment effect in the
linear mixed effects models for pulse pair rate, peak amplitude, and
attractiveness as body mass represents an intermediate outcome
variable which must not be used because it is impossible to separate
treatment effects from effects of the intermediate outcome variable
in this case (Gelman and Hill 2007, pp. 188–194). For post hoc
analysis of treatment level contrasts, we performed Tukey’s HSD
tests (R function glht from the multcomp package, Hothorn et al.
2008). Finally, to analyze the relationship between body mass and
song characteristics as well as risk-taking behavior, we calculated
Pearson’s correlations between these variables.
Results
Individual repeatability and treatment effects
The song traits pulse pair rate, peak amplitude, and attractiveness were significantly repeatable (Table 1), as was silence response
(R = 0.35, 95% HPD interval: 0.26, 0.44, n = 171). Larval rearing density had no effect on pulse pair rate (χ2 = 1.38, df = 2,
P = 0.5) but on peak amplitude (χ2 = 8.25, df = 2, P = 0.016) and,
as a consequence, on attractiveness (χ2 = 8.31, df = 2, P = 0.016).
Males raised under medium larval density performed louder songs
(Z = −2.96, P = 0.009) and had a higher attractiveness score
(Z = −2.97, P = 0.008) than males raised under high density.
Solitarily reared males did not differ from males reared at medium
411
(peak amplitude: Z = 1.37, P = 0.35, attractiveness: Z = 1.46,
P = 0.31) or high density (peak amplitude: Z = −1.09, P = 0.51,
attractiveness: Z = −1.03, P = 0.56).
The larval density treatment had no significant effect on male
body mass (χ2 = 2.39, df = 2, P = 0.3). Overall, body mass was
significantly positively correlated with peak amplitude (r = 0.40,
P < 0.001) and attractiveness (r = 0.39, P < 0.001), but not with
pulse pair rate (r = 0.09, P = 0.22) or silence response (r = −0.09,
P = 0.26).
Correlations between song characteristics and
silence response
We found a significant negative overall correlation between silence
response and pulse pair rate (Table 1). Individuals with high pulse
pair rates in courtship song showed shorter silence responses than
individuals with lower pulse pair rates (Fig. 1a). The bivariate
mixed model confirmed the significant negative between-individual
correlation between silence response and pulse pair rate (Table 1).
This correlation was consistent across treatment levels as reflected
by the same sign and similar strength of the correlations. A withinindividual correlation was only significant for males from the lowdensity treatment (Table 1).
We found a significant negative overall correlation between
silence response and peak amplitude (Table 1). Males with songs
of high amplitude had shorter silence responses than males singing
at lower amplitudes (Fig. 1b). The bivariate mixed model, however,
showed no significant between- or within-individual correlations
(Table 1). Although correlations were all consistently negative, 95%
HPD intervals for between- and within-correlation coefficients
overlapped with 0 for the full sample as well as for the treatmentlevel-specific subsamples (Table 1).
Across experimental treatment levels, pulse pair rate and peak
amplitude were significantly positively correlated (Pearson’s
Table 1
Correlations between the duration of the silence response in A. grisella courtship induced by playback of bat searching calls and 3
courtship song characteristics (pulse pair rate, peak amplitude, and a compound attractiveness score calculated from these 2 traits)
Pulse pair rate
Peak amplitude
Attractiveness
Pearson’s correlation
n
Estimate
P
Estimate
P
Estimate
P
rmeans
Low density
Medium density
High density
171
31
45
95
−0.47
−0.46
−0.65
−0.46
<0.001
0.009
<0.001
<0.002
−0.15
−0.15
−0.05
−0.15
0.048
0.43
0.76
0.15
−0.24
−0.32
−0.18
−0.23
0.002
0.08
0.23
0.03
Bivariate mixed model
n
Estimate
95% HPD interval
Estimate
95% HPD interval
Estimate
95% HPD interval
rbetween
Low density
Medium density
High density
rwithin
Low density
Medium density
High density
Repeatability
171
31
45
95
171
31
45
95
171
−0.49
−0.31
−0.65
−0.49
−0.12
−0.30
−0.04
−0.08
0.61
−0.66, −0.31
−0.69, 0.18
−0.88, −0.29
−0.71, −0.22
−0.22, −0.02
−0.52, −0.05
−0.25, 0.17
−0.22, 0.06
0.53, 0.68
−0.11
−0.12
−0.03
−0.16
−0.05
−0.23
0.02
0.05
0.61
−0.32, 0.09
−0.54, 0.36
−0.49, 0.51
−0.43, 0.14
−0.16, 0.05
−0.46, 0.01
−0.18, 0.22
−0.19, 0.09
0.53, 0.68
−0.21
−0.18
−0.10
−0.25
−0.08
−0.31
0.02
−0.07
0.63
−0.42, −0.003
−0.58, 0.30
−0.58, 0.39
−0.52, 0.03
−0.18, 0.02
−0.52, −0.08
−0.19, 0.23
−0.21, 0.07
0.56, 0.70
Correlations are presented separately for (sub) samples across and within 3 larval density treatment levels. rmeans is the Pearson correlation of means per individual
(from 3 subsequent measurements); rbetween represents between-individual correlations and rwithin represents within-individual correlations as determined by a
bivariate mixed model with individual identity and family identity as random effects and the overall mean as only fixed effect (see text for details). Repeatabilities
were estimated from the bivariate model. Bold face indicates significant effects (P < 0.05); estimates from the bivariate mixed model were considered significant if
an estimate’s 95% HPD interval did not overlap with zero.
Behavioral Ecology
412
Mean duration of
silence response (log10 (ms))
a
5
4
3
2
1
0
50
Mean duration of
silence response (log10 (ms))
b
60
70
80
90
100
Mean pulse pair rate (pulse pairs s-1)
5
4
3
2
1
0
0
Mean duration of
silence response (log10 (ms))
c
0.2
0.4
0.6
0.8
Mean peak amplitude (V)
1.0
5
4
3
2
1
0
-1.0
-0.5
0
0.5
1.0
Attractiveness score
1.5
2.0
Figure 1
Correlations between the duration of the silence response and (a) pulse
pair rate, (b) peak amplitude, and (c) compound attractiveness of Achroia
grisella courtship songs (n = 171). Symbols indicate means of 3 subsequent
measurements per individual for males from low (circle), medium (square),
and high (triangle) larval density treatment levels.
correlation: r = 0.28, P < 0.001, n = 171). Furthermore, the compound attractiveness score was overall significantly negatively
correlated with silence response (Table 1). Individuals with high
attractiveness scores showed shorter silence responses than individuals with lower attractiveness scores (Fig. 1c). The bivariate mixed
model confirmed a significant negative between-individual correlation for silence response and attractiveness (Table 1). This correlation was also consistent across treatment levels as reflected by the
same sign and similar strength of the correlations. However, 95%
HPD intervals overlapped with 0 for the treatment-level-specific
subsamples, and within-individual correlation was only significant
for males from the low-density treatment (Table 1).
Discussion
We hypothesized that individual attractiveness would covary negatively with risk-taking behavior in the lesser wax moth. Specifically,
we predicted a positive correlation between 2 sexually selected
traits and the duration of a silence response in the presence of bats
as this would suggest that more attractive males take fewer risks.
Hence, individuals with low pulse pair rate and low peak amplitude
should have shorter silence responses than moths with high pulse
pair rates and peak amplitudes because they should expect fewer
future matings due to their lower attractiveness to females (Jang
and Greenfield 1996) and should benefit more from recommencing
courtship sooner under the perceived threat of predation. However,
contrary to our expectations, we found consistently negative correlations between the variables. Individuals with higher pulse pair
rate and peak amplitude were more likely to recommence their
courtship signaling soon after a confrontation with a bat signal.
This correlation appears robust as effect sizes for between-individual correlations were similar across all larval densities. Furthermore,
the signs of the correlations were the same for pulse pair rate, peak
amplitude, and the compound attractiveness score in both Pearson’s
correlations of means per individual and the bivariate mixed model
analysis. This means that males with short silence responses always
had higher values for the attractiveness traits.
The observation that more attractive individuals seem to take
greater risks than less attractive ones is quite surprising and rare
(Schuett et al. 2010; Wilson et al. 2010). In fact, according to the asset
protection principle, there are good reasons for attractive individuals
to take fewer risks. Considering that risky behavior means jeopardizing one’s assets, individuals with much to lose, i.e., attractive virgin
males with an expectation of multiple future mating opportunities,
should behave more cautiously than individuals for whom future
copulations are unlikely (Clark 1994). Expanded into a theory for the
evolution of distinct behavioral correlations, Wolf et al. (2007) used
the residual reproductive value as a crucial component in modeling
the evolution of separate strategies within 1 population, resulting
in risk-prone and risk-averse individuals. Although this explanation
may apply to behavioral differences in lesser wax moths of different
age, younger wax moths take fewer risks than old ones (Lafaille et al.
2010), it does not seem to fit with the observed between-male differences in relation to acoustic attractiveness. However, considering
that costs are involved in producing high pulse pair rate (Reinhold
et al. 1998), attractive individuals may exhaust their resources sooner
and have shorter life expectancies than less attractive individuals (cf.
Höglund and Sheldon 1998; Hunt et al. 2004). If this was indeed the
case, attractive individuals rather employ something akin to a “live
fast, die young” strategy (Bonduriansky et al. 2008) and therefore
may actually benefit more from taking greater risks than unattractive
males. Thus, what we perceive as a moth with high residual reproductive value (and therefore assume should be a more cautious one)
may in fact be the one which—in concordance with the asset protection principle—should behave less cautiously.
It is also possible that mechanical constraints may contribute to the explanation of the observed correlations. For instance,
both pulse pair rate and silence response might be affected by the
dynamics of muscular tension in the thorax, as pulse pair rate is
simply the wing beat frequency during singing (Spangler et al.
1984), and an interaction between muscular tension and proprioceptors might influence the recommencement of acoustic signaling
after a pause. In principle, body mass and wing size may also have
an influence on pulse pair rate and peak amplitude (Brandt and
Greenfield 2004), but little is known about their effect on the silence
response (Cordes et al. 2013). In any case, as we did not find a correlation between silence response and body mass, it is unlikely that
body mass or size should be able to explain our results.
The argument that attractive males take higher risks by reducing the duration of their silence responses implicitly assumes that
Cordes et al. • Attractiveness and risk-taking in wax moths
the extent to which sexual advertisement increases predation risk
is uncorrelated with male attractiveness. However, actual predation risk for attractive and unattractive A. grisella males has not been
studied and remains to be tested. Thus, attractive males could, for
example, be more viable and better at actively avoiding predation.
In this case, attractiveness would signify a reduction in the degree
of risk for an individual male (Lindström et al. 2006; Sih and Bell
2008)—a type of positive feedback between asset and behavior
contrary to the negative feedback suggested by Wolf et al. (2007).
Risk-taking behavior could then be perceived as an honest signal
of male quality. Honest sexual signals must be costly and condition dependent (Kotiaho 2000, 2001). There is as yet no evidence
that attractive wax moths should be generally better at avoiding
predation, but because individuals need more energy to produce a
high pulse pair rate (Reinhold et al. 1998) and high peak amplitude
reflects a larger body size, attractive—and, based on this study, the
more risky—individuals may actually have and broadcast an overall
higher phenotypic and possibly genotypic quality.
Finally, the bivariate mixed model analyses allowed us to determine the contribution of within-individual effects on these correlations. The presence of within-individual correlations here means
that individuals show correlated changes in song characteristics and
risk-taking behavior over subsequent observations. Curiously, withinindividual correlations between silence response and pulse pair rate
as well as attractiveness were largest and significantly different from
0 for solitarily reared males. This suggests that within-individual
behavioral changes are stronger in solitarily reared males than those
reared in groups. Such individual behavioral adjustments may be
expected for pulse pair rate, which males are known to adjust slightly,
albeit only under high-competition conditions (Jia et al. 2001). In
general, the within-individual correlation between silence response
and the song traits could be explained by an age effect. The experiment was performed over the course of 25 h, and although this is a
relatively short period, age is known to affect both song (Jang et al.
1997) and risk-taking behavior (Lafaille et al. 2010) in A. grisella. Age
may therefore have contributed similarly to a change in pulse pair
rate and the attractiveness score within individuals, leading to the
detected within-individual correlation between the traits.
We do not know whether the among-individual correlations we
found on the phenotypic level reflect genetically correlated traits.
Interestingly, there is some indication that silence response has
a heritable component (Cordes et al. 2013), and both pulse pair
rate and peak amplitude have actually been shown to be heritable
(Collins et al. 1999). This implies that investigating genetic correlations along these lines may provide a valuable direction for future
research. We are currently working on a theoretical framework in
order to understand under which circumstances it would be expected
that more attractive individuals should take greater risks. Irrespective
of whether the observed correlation is the result of differential risktaking in response to expected future reproductive output, or risk-taking behavior in A. grisella actually represents an honest signal of male
quality, our results illustrate the importance to understand in detail
the causes of repeatable behavioral differences among individuals.
Funding
German Research Foundation (DFG) (project number FOR 1232).
We thank M. Greenfield for providing us with recordings of R. ferrumequinum, R. Feist, M. Hausmann, P. Erkeling, D. Kluge, and J. Gumpert for the
support in rearing of wax moth populations, and H. Schielzeth for help
413
with the analysis and comments on an earlier version of this manuscript.
Special thanks to G. Machado and two anonymous reviewers for critical
discussion and valuable suggestions.
Handling editor: Glauco Machado
References
Acharya L, McNeil JN. 1998. Predation risk and mating behavior: the
responses of moths to bat-like ultrasound. Behav Ecol. 9:552–558.
Alem S, Koselj K, Siemers BM, Greenfield MD. 2011. Bat predation
and the evolution of leks in acoustic moths. Behav Ecol Sociobiol.
65:2105–2116.
Bates D. 2007. lme4: linear mixed-effects models using S4 classes. R package, version 0.99875–9.
Billing AM, Rosenqvist G, Berglund A. 2007. No terminal investment in
pipefish males: only young males exhibit risk-prone courtship behavior.
Behav Ecol. 18:535–540.
Bonduriansky R, Maklakov A, Zajitschek F, Brooks R. 2008. Sexual selection, sexual conflict and the evolution of ageing and life span. Funct Ecol.
22:443–453.
Brandt LS, Greenfield MD. 2004. Condition-dependent traits and the
capture of genetic variance in male advertisement song. J Evol Biol.
17:821–828.
Brunel-Pons O, Alem S, Greenfield MD. 2011. The complex auditory scene
at leks: balancing antipredator behaviour and competitive signalling in an
acoustic moth. Anim Behav. 81:231–239.
Candolin U. 1997. Predation risk affects courtship and attractiveness of
competing threespine stickleback males. Behav Ecol Sociobiol. 41:81–87.
Candolin U. 1998. Reproduction under predation risk and the trade-off
between current and future reproduction in the threespine stickleback.
Proc R Soc B. 265:1171–1175.
Clark CW. 1994. Antipredator behavior and the asset-protection principle.
Behav Ecol. 5:159–170.
Collins RD, Jang Y, Reinhold K, Greenfield MD. 1999. Quantitative genetics of ultrasonic advertisement signalling in the lesser waxmoth Achroia
grisella (Lepidoptera: Pyralidae). Heredity. 83:644–651.
Cordes N, Schmoll T, Reinhold K. 2013. Risk-taking behavior in the lesser
wax moth: disentangling within- and between-individual variation. Behav
Ecol Sociobiol. 67:257–264.
Dall SRX, Houston AI, McNamara JM. 2004. The behavioural ecology of
personality: consistent individual differences from an adaptive perspective. Ecol Lett. 7:734–739.
Dingemanse NJ, Dochtermann NA. 2013. Quantifying individual variation
in behaviour: mixed-effect modelling approaches. J Anim Ecol. 82:39–54.
Dingemanse NJ, Wolf M. 2010. Recent models for adaptive personality differences: a review. Philos Trans R Soc Lond B Biol Sci. 365:3947–3958.
Gelman A, Hill J. 2007. Data analysis using regression and multilevel/hierarchical models. Cambridge (MA): Cambridge University Press.
Gelman A, Rubin DB. 1992. Inference from iterative simulation using multiple sequences. Stat Sci. 7:457–472.
Gosling SD, John OP. 1999. Personality dimensions in nonhuman animals:
a cross-species review. Curr Dir Psychol. 8:69–75.
Greenfield MD, Baker M. 2003. Bat avoidance in non-aerial insects:
the silence response of signaling males in an acoustic moth. Ethology.
109:427–442.
Greenfield MD, Weber T. 2000. Evolution of ultrasonic signalling in wax
moths: discrimination of ultrasonic mating calls from bat echolocation
signals and the exploitation of an anti-predator receiver bias by sexual
advertisement. Ethol Ecol Evol. 12:259–279.
Hadfield JD. 2010. MCMC methods for multi-response generalized linear
mixed models: the MCMCglmm R Package. J Stat Softw. 33:1–22.
Hedrick AV. 2000. Crickets with extravagant mating songs compensate for
predation risk with extra caution. Proc Biol Sci. 267:671–675.
Höglund J, Sheldon BC. 1998. The cost of reproduction and sexual selection. Oikos. 83:478–483.
Hothorn T, Bretz F, Westfall P. 2008. Simultaneous inference in general
parametric models. Biom J. 50:346–363.
Hunt J, Brooks R, Jennions MD, Smith MJ, Bentsen CL, Bussière LF. 2004.
High-quality male field crickets invest heavily in sexual display but die
young. Nature. 432:1024–1027.
Jang Y, Collins RD, Greenfield MD. 1997. Variation and repeatability of
ultrasonic sexual advertisement signals in Achroia grisella (Lepidoptera:
Pyralidae). J Ins Behav. 9:87–98.
414
Jang Y, Greenfield MD. 1996. Ultrasonic communication and sexual selection in wax moths: female choice based on energy and asynchrony of
male signals. Anim Behav. 51:1095–1106.
Jang Y, Greenfield MD. 1998. Absolute versus relative measurements of
sexual selection: assessing the contributions of ultrasonic signal characters to mate attraction in lesser wax moths, Achroia grisella. Evolution.
52:1383–1393.
Jia F-Y, Greenfield MD, Collins RD. 2001. Ultrasonic signal competition
between male wax moths. J Ins Behav. 14:19–33.
Kemp DJ. 2002. Sexual selection constrained by life history in a butterfly.
Proc R Soc B. 269:1341–1345.
Kotiaho, JS. 2000. Testing the assumptions of conditional handicap theory:
costs and condition dependence of a sexually selected trait. Behav Ecol
Sociobiol 48:188–194.
Kotiaho JS. 2001. Costs of sexual traits: a mismatch between theoretical considerations and empirical evidence. Biol Rev Camb Philos Soc.
76:365–376.
Kunike G. 1930. Zur Biologie der kleinen Wachsmotte, Achroea grisella Fabr.
Z angew Ent. 16:304–356.
Lafaille M, Bimbard G, Greenfield MD. 2010. Risk trading in mating
behavior: forgoing anti-predator responses reduces the likelihood of missing terminal mating opportunities. Behav Ecol Sociobiol. 64:1485–1494.
Lindström L, Ahtiainen JJ, Mappes J, Kotiaho JS, Lyytinen A, Alatalo RV.
2006. Negatively condition dependent predation cost of a positively condition dependent sexual signalling. J Evol Biol. 19:649–656.
Magnhagen C. 1991. Predation risk as a cost of reproduction. Trends Ecol
Evol. 6:183–186.
Behavioral Ecology
R Development Core Team. 2011. R: a language and environment for
statistical computing. Vienna (Austria): R Foundation for Statistical
Computing. Available from: http://www.R-project.org.
Reinhold K, Greenfield MD, Jang Y, Broce A. 1998. Energetic cost of sexual
attractiveness: ultrasonic advertisement in wax moths. Anim Behav. 55:905–913.
Ryan MJ, Tuttle MD, Rand AS. 1982. Bat predation and sexual advertisement in a neotropical Anuran. Am Nat. 119:136–139.
Schnitzler H-U, Ostwald J. 1983. Adaptations for the detection of fluttering
insects by echolocation in horseshoe bats. In: Ewert J-P, Capranica RR,
Ingle DJ, editors. Advances in vertebrate neuroethology. New York (NY):
Plenum Press. p. 801–827.
Schuett W, Tregenza T, Dall SR. 2010. Sexual selection and animal personality. Biol Rev Camb Philos Soc. 85:217–246.
Sih A, Bell AM. 2008. Insights for behavioral ecology from behavioral syndromes. Adv Stud Behav. 38:227–281.
Spangler HG. 1984. Silence as a defense against predatory bats in two species of calling insects. Southwest Nat. 29:481–488.
Spangler HG, Greenfield MD, Takessian A. 1984. Ultrasonic mate calling
in the lesser wax moth. Physiol Entomol. 9:87–95.
Wilson ADM, Whattam EM, Bennett R, Visanuvimol L, Lauzon C,
Bertram, SM. 2010. Behavioral correlations across activity, mating,
exploration, aggression, and antipredator contexts in the European house
cricket, Acheta domesticus. Behav Ecol Sociobiol. 64:703–715.
Wolf M, van Doorn GS, Leimar O, Weissing FJ. 2007. Life-history tradeoffs favour the evolution of animal personalities. Nature. 447:581–584.
Zuk M, Kolluru GR. 1998. Exploitation of sexual signals by predators and
parasitoids. Q Rev Biol. 73:415–438.