Adjustment of vibratory signals to ambient temperature in a host‐searching parasitoid

Abstract

Abstract Certain ichneumonid parasitoids (Hymenoptera) use self-produced vibrations transmitted on plant substrate, so-called vibrational sounding, to locate their immobile concealed pupal hosts. An ambient temperature dependency with higher frequencies and intensities at higher temperatures is postulated because signals are of myogenic origin. Here, temperature influence on vibratory signals is analysed in the temperate parasitoid Pimpla turionellae under different thermal conditions using plant-stem models to elicit host-searching behaviour. Signals are measured with laser Doppler vibrometry and analysed for time parameters and frequency components applying fast-Fourier transformations. The results reveal an unexpected effect of ambient temperature on signals produced by the female wasps. Although average values of time parameters (pulse trains, pulse train periods, inter pulse duration) are unchanged by ambient temperature, the frequency parameters show an inverse thermal effect. Within the temperature range tested (8–26 °C), decreasing temperature leads to significantly higher frequency and intensity of the self-produced vibrations in the temperate species. This inverse thermal effect may be explained by a temperature-coupled signal production in the frequency domain to compensate negative low-temperature effects on the mechanoreceptors by increased muscle activity. The option of heterothermy to produce signals reliably during vibrational sounding under low temperature is also discussed.

Full text

Page 1

Physiological Entomology (2007) 32, 105–112 DOI: 10.1111/j.1365-3032.2006.00551.x
© 2006 The Authors
Journal compilation © 2006 The Royal Entomological Society 105
Introduction
The use of substrate-borne vibratory signals is widespread
throughout arthropods as an information source and a com-
munication pathway ( Lewis, 1983; Hill, 2001; Virant-Doberlet
& Cokl, 2004 ). Examples are related to sexual behaviour
( Roces & Manrique, 1996 ), aggregation and cooperative for-
aging behaviour ( Schneider et al. , 1986 ), alarm communica-
tion ( Wood, 1976 ) and, in the case of bending waves, to the
localization of the vibration source ( Miles et al. , 2001; Barth,
2002 ). Certain parasitoids locate concealed host insects by
host-specific or self-produced vibrations ( Meyhöfer et al. ,
1994; Fischer et al. , 2004 ) but information to potential hosts
on approaching natural enemies can also be vibration-mediated
( Bacher et al. , 1997; Meyhöfer et al. , 1997; Djemai et al. ,
2001 ). Vibrations are the least costly signals for communica-
tion in arthropods; the range can be up to 1000-fold greater
than the length of the body, propagation diffuses little, the
signal is confined within the substrate and is therefore easier
to locate and less likely to attract enemies ( Bennet-Clark,
1998; Barth, 2002; Virant-Doberlet & Cokl, 2004 ).
Nevertheless, little is known about the influence of envi-
ronmental factors on the sensory physiology of vibratory
systems and especially of temperature as one of the most
important abiotic factors. Temperature should be especially
influential because the physical activity of insects, as well as
the medium-specific attenuation of vibrational signals,
largely depends on thermal conditions ( Gogala, 1985;
Bennet-Clark, 1998; Greenfield, 2002 ). For example, in spi-
ders, higher ambient temperatures result in an increasing
frequency of vibratory pulses emitted during intraspecific
Adjustment of vibratory signals to ambient
temperature in a host-searching parasitoid
S T E F A N K R O D E R 1 , J Ö R G S A M I E T Z 1 , 2 , D AV I D S C H N E I D E R 1
and S I LV I A D O R N 1
1 Institute of Plant Sciences, Applied Entomology, ETH Zurich, Zurich and 2 Swiss Federal Research Station Agroscope
Changins-Wädenswil, Wädenswil, Switzerland.
Abstract. Certain ichneumonid parasitoids (Hymenoptera) use self-produced
vibrations transmitted on plant substrate, so-called vibrational sounding, to locate
their immobile concealed pupal hosts. An ambient temperature dependency with
higher frequencies and intensities at higher temperatures is postulated because sig-
nals are of myogenic origin. Here, temperature influence on vibratory signals is
analysed in the temperate parasitoid Pimpla turionellae under different thermal con-
ditions using plant-stem models to elicit host-searching behaviour. Signals are
measured with laser Doppler vibrometry and analysed for time parameters and fre-
quency components applying fast-Fourier transformations. The results reveal an un-
expected effect of ambient temperature on signals produced by the female wasps.
Although average values of time parameters (pulse trains, pulse train periods, inter
pulse duration) are unchanged by ambient temperature, the frequency parameters
show an inverse thermal effect. Within the temperature range tested (8 – 26 °C), de-
creasing temperature leads to significantly higher frequency and intensity of the
self-produced vibrations in the temperate species. This inverse thermal effect may
be explained by a temperature-coupled signal production in the frequency domain
to compensate negative low-temperature effects on the mechanoreceptors by
increased muscle activity. The option of heterothermy to produce signals reliably
during vibrational sounding under low temperature is also discussed.
Key words. Echolocation , host location , parasitoid , temperature , thermoregulation ,
vibrational sounding .
Correspondence: Dr Jörg Samietz, Swiss Federal Research Station
ACW, Schloss, PO Box 185, CH-8820 Wädenswil, Switzerland. Tel.:
+41 44 783 6193; fax: +41 44 783 6434; e-mail: joerg.samietz@acw.
admin.ch

Page 2

Temperature affects vibrational sounding 107
© 2006 The Authors
Journal compilation © 2006 The Royal Entomological Society, Physiological Entomology, 32, 105–112
(Mettler Toledo MT 5; Göttingen, Germany; accuracy ±
0.001 mg).
Laser Doppler vibrometry
Vibrations transmitted by the parasitoid during mechano-
sensory host location on the model were measured with laser
Doppler vibrometer (LDV; 2 mW He-Ne-Laser, type: 41 ×
62; Dantec, Denmark). The laser beam was focused onto a
dot of retroreflective tape (Scotchlite; 3M, Germany) to ob-
tain an optimal reflection. This tape was attached to the sur-
face of the plant-stem model at a distance of 22 mm from the
filter outside the cup.
The output of the LDV is a voltage signal directly propor-
tional to the measured velocity component. To digitize the
analogue signals, the recordings were first cleaned of fre-
quencies over 25 kHz with a low-pass filter (SR650; Stanford
Research Systems, Sunnyvale, California) and subsequently
sampled by a data acquisition system on an Apple Macintosh
computer (InstruNet Model 100B; GW Instruments,
Somerville, Massachusetts) with a sampling rate of 50000
points s – 1 . Thus, the measurements could be analysed in the
frequency domain up to 25 kHz. This broad range was cho-
sen to enable sampling of frequencies directly produced for
vibrational sounding as well as related eigenfrequencies of
the wasp’s body parts or percussion effects of the antennae.
Signal analysis
The terminology used for the time components of the vi-
bratory signals is according to Meyhöfer et al. (1994) and
Hrncir et al. (2004) . It is based on the measured velocity of
the vibrating substrate ( Fig. 1). The females elicit trains of
vibrations, composed of pulses, during vibrational sounding.
Only peaks in the time series of the vibrations with a veloc-
ity of more than 0.2 mm s – 1 and a hysteresis of 0.01 can
safely be taken to have been generated by the vibrating fe-
male and are referred to as single pulses. Consecutive pulses
form a pulse train (PT). A pulse train ends when a break of
at least 3 ms separates two successive pulses. Such breaks
between pulse trains are defined as inter pulse duration
(IPD). An upper threshold for IPD in these analyses was set
at 200 ms, although the maximum value reached was only 83
ms. Longer gaps are considered to be breaks in the vibra-
tional sounding of the female. The duration from the begin-
ning of a pulse train to the beginning of the next pulse train
represents the pulse train period (PTP).
Data were analysed both in the time and frequency do-
main. The software SoundScope 16 3.0 (GW Instruments)
was used for analysis on an Apple Macintosh computer (G3,
350 MHz). The algorithms contained in SoundScope were
programmed to meet the special requirements covered in
measuring vibrational sounding ( Weinreb & McCabe, 1996 ).
The measurements were high-pass filtered at a threshold of
0.5 kHz to eliminate disturbing low-frequency oscillations.
The peaks of pulses served as points of measurements both
in the time and frequency domain.
The frequency domain of vibrations was analysed by fast-
Fourier transformation (FFT) subdividing the vibrations into
frequency components containing a primary oscillation with
respective harmonics ( Bracewell, 1989 ). The result of a FFT
is a frequency spectrum, referred to as periodogram (i.e. the
intensity of the harmonics displayed as a function of the fre-
quency). The frequency with the highest intensity was de-
fined as the carrier frequency. Fast-Fourier transformation
was performed over a range of 5.12 ms (256 points) around
each pulse of the time series. Thus, the FFT-result function
consisted of 128 points. Subsequently, the frequency compo-
nents and the corresponding intensities of all detected pulses
were averaged. A smoothed periodogram between 0 – 25 kHz
with a carrier frequency was obtained in each recording. The
resolution was 195 Hz corresponding to 128 points extended
over 25000 Hz ( Fig. 2).
The most feasible description of the intensity level (IL) is
by definition a logarithmic function with the unit decibel
(dB) related to the parameters of acoustic sounds. The for-
mula to compute the IL was:
?
?
?
?
?
?
?
?
?
???
22
2
10 lg
–4
10
realimag
IL
FFTpoints
In this equation, FFT points is the number of points included
in one FFT, ‘real’ is the real point value and ‘imag’ is the
imaginary point value of the complex pairs in the FFT result-
ing wave ( Weinreb & McCabe, 1996 ).
For each single recording, the average duration of PT, PTP
and IPD were calculated. In addition, the carrier frequency
and the intensity at the carrier frequency of the smoothed
periodogram were recorded for further analysis of the data.
Data analyses
The vibratory signals of 20 females per temperature treat-
ment were measured. Thus, a total number of 80 females in
four treatments were tested. The number of recordings per fe-
male depends on individual activity and reaches from 1 to a
maximum of 16, with the number of measurements made be-
ing: 96 (8 °C), 139 (14 °C), 140 (20 °C) and 178 (26 °C). The
recordings taken from single females have durations of 2 – 35 s
and comprise a number of peaks between 76 and 19 177 with
a higher velocity than 0.2 mm s – 1 and a hysteresis of 0.01.
The influence of temperature on the frequency parameters
of the vibrations (carrier frequency and intensity) was exam-
ined using analyses of covariance ( ancova ) with subsequent
multiple post-hoc comparisons by the Sidak-test. The covari-
ate bodyweight was included in the frequency and intensity
analysis because female size is known to influence the inten-
sity of vibrations ( Otten et al. , 2001 ). Because the number of
records for each individual was different, analyses of covari-
ance were also carried out with mean values weighted by the
individual number of records using weighted least squares
regression (WLS). The influence of temperature on the time
parameters of the vibratory signals (PT, PTP, IPD) were

Page 3

Temperature affects vibrational sounding 109
© 2006 The Authors
Journal compilation © 2006 The Royal Entomological Society, Physiological Entomology, 32, 105–112
between 14 – 26 °C (post-hoc Sidak-tests, WLS weighted
mean values: 8 – 20 °C, P = 0.036; 8 – 26 °C, P = 0.003; 14 –
20 °C, P = 0.024; 14 – 26 °C: P = 0.001).
The intensity at the carrier frequency was also influenced
by temperature ( ancova factor: F 3,79 = 5.09, P < 0.001)
( Fig. 4). There was no significant effect of bodyweight on the
intensity ( ancova covariate: F 1,79 = 1.06, P = 0.169). The
highest value (mean ± SE) was found at 8 °C (29.40 ± 0.19
dB) and the lowest at 26 °C (28.36 ± 0.15 dB). Significant
differences were found in the subsequent multiple compari-
sons between 8 – 20 °C and between 8 – 26 °C (post-hoc Sidak-
tests; 8 – 20 °C, P < 0.001; 8 – 26 °C, P < 0.001).
In the comparison of the mean values weighted by the in-
dividual number of records, both the effect of temperature
and the effect of bodyweight were significant ( ancova , WLS
weighted mean values: factor: F 3,79 = 10.917, P < 0.001;
covariate: F 1,79 = 5.891, P = 0.018). The weighted mean
value ranged from 28.37 dB at 26 °C to 29.28 dB at 8 °C.
The differences between 8 – 20 °C, 8 – 26 °C, 14 – 20 °C and
14 – 26 °C were significant in the multiple comparisons (post-
hoc Sidak-tests, WLS weighted mean values: 8 – 20 °C, P <
0.001; 8 – 26 °C, P < 0.001; 14 – 20 °C, P = 0.007; 14 – 26 °C:
P = 0.002).
Time domain of vibratory signals
Similar average lengths of the parameters PT, PTP and
IPD were detected in all four temperature treatments ( Fig. 5).
Average length (mean ± SE) was 8.64 ± 0.85 ms for a PT,
25.1 ± 2.5 ms for a PTP and 16.4 ± 2.5 ms for an IPD over
all treatments. There was no significant difference between
Table 1. Frequency parameters of vibrational sounding in Pimpla turionellae.
Temperature
Unweighted mean values WLS – weighted
Freq ± SEM (Hz)
1391 ± 28
1410 ± 30
1333 ± 30
1290 ± 21
Int ± SEM (dB)
29.40 ± 0.19
28.91 ± 0.17
28.41 ± 0.15
28.36 ± 0.15
Freq ± SEM (Hz)
1419 ± 24
1418 ± 25
1324 ± 33
1296 ± 23
Int ± SEM (dB)
29.28 ± 0.16
29.03 ± 0.15
28.42 ± 0.17
28.37 ± 0.16
8 °C
14 °C
20 °C
26 °C
Mean values of dominant frequency (Freq) and intensity at dominant frequency (Int) with standard error of mean (SEM) at 8, 14, 20 and 26 °C. Mean
values are presented as unweighted data (left) and weighted data by individual number of records using weighted least squares regression (WLS) (right).
Fig. 3. Carrier frequency (unweighted data) of Pimpla turionellae
at 8, 14, 20 and 26 °C. Bold midlines indicate medians (values
shown); dashed indicate means. Boxes range from the 25th to 75th
percentile. Error lines extend from the 10th to 90th percentile. Sig-
nificant differences between temperatures are indicated by different
letters (analysis of covariance, post-hoc Sidak-test).
Fig. 4. Intensity at carrier frequency (unweighted data) of Pimpla
turionellae at 8, 14, 20 and 26 °C. Bold midlines indicate medians
(values shown); dashed lines indicate means. Boxes range from the
25th to 75th percentile. Error lines extend from the 10th to 90th per-
centile. Significant differences between temperatures are indicated
by different letters (analysis of covariance, post-hoc Sidak-test).

Page 4

110 S. Kroder et al.
© 2006 The Authors
Journal compilation © 2006 The Royal Entomological Society, Physiological Entomology, 32, 105–112
the temperature treatments in these parameters ( anova ; PT:
F 3,79 = 1.31, P = 0.279; PTP: F 3,79 = 0.122, P = 0.947;
IPD: F 3,79 = 0.215, P = 0.886). A trend without significance is
noticeable in the average duration of a PT, with a maximum
at 8 °C (9.24 ± 0.52 ms) and a minimum at 26 °C (8.22 ±
0.46 ms). By contrast, variability of the parameters charac-
terized by the coefficient of variation (CV) differed signifi-
cantly (CV- ? 2 test; PT: ? 2 = 13.2, d.f. = 3, 76, P = 0.001;
PTP: ? 2 = 12.5, d.f. = 3, 76, P = 0.009; IDP: ? 2 = 10.9,
d.f. = 3, 76, P = 0.022). In pairwise comparisons, both PT
and PTP variations differed significantly between all tem-
peratures used in the experiments except for the difference of
PT variation between the extreme temperatures of 8 and 26 °C
where variations were significantly higher than in the inter-
mediate temperatures (CV pairwise tests: PT, 8 – 14 °C: P <
0.001; PT, 8 – 20 °C: P < 0.001; PT, 14 – 20 °C: P < 0.001;
PT, 14 – 26 °C: P < 0.001; PT, 20 – 26 °C: P < 0.001; PTP,
8 – 14 °C: P < 0.001; PTP, 8 – 20 °C: P < 0.001; PTP, 8 – 26
°C: P < 0.001; PTP, 14 – 20 °C: P < 0.001; PTP, 14 – 26 °C:
P = 0.002; PTP, 20 – 26 °C: P < 0.001). Furthermore, the
variability of IPD was significant higher at 8 °C than at all
other temperatures used in the experiments (CV pairwise
test: IDP, 8 – 14 °C: P < 0.001; IDP, 8 – 20 °C: P < 0.001;
IDP, 8 – 26 °C: P < 0.001).
Discussion
The present study shows an unexpected effect of ambient tem-
perature on the vibratory signals produced by the female wasps
during host location. Although vibratory or acoustic signals in
ectotherms are usually slower and less intense at low tempera-
tures, here the time parameters are unchanged by ambient
temperature for their duration and are only affected with re-
spect to their variation. Moreover, the frequency parameters
even show an inverse thermal effect during vibrational sound-
ing. Decreasing ambient temperature leads to a higher fre-
quency and intensity of the self-produced vibrations.
Reports on vibratory or acoustical signals in other ecto-
therms document a positive relationship of frequency and in-
tensity to temperature. Higher body temperatures elevate the
muscle contraction rates and allow faster and stronger oscil-
lation of the muscles that generate the vibrations and hence
produce higher frequencies with a higher intensity
( Greenfield, 2002 ). Such a correlation between temperature
and sound power is well known in cicada calling songs where
acoustic output and song intensity are highly dependent on
body temperature ( Sueur & Sanborn, 2003 ). The fact that vi-
bratory signals of P. turionellae do not show an according
effect indicates an independence of signal production from
ambient temperature. Such independence could be achieved
by thermoregulation of the body in a temperature range in
which signal production is not or negligibly affected. Many
poikilotherms, and in particular several hymenopteran spe-
cies, raise their body temperature above the environmental
temperature by temporary endothermy ( Heinrich & Kammer,
1973; Heinrich, 1993; Willmer et al. , 2000; Stabentheiner,
2001 ). This so-called heterothermy ( Heinrich, 1993 ) is ex-
emplified in bumblebees ( Bombus vosnesenskii ), which sta-
bilize their thoracic temperature at 33 – 36 °C over ambient
temperatures in the range 5 – 30 °C ( Heinrich & Kammer,
1973 ). Similarly, the cicada Tibicen winnemanna is able to
maintain tymbal muscle temperature 13 °C above ambient
temperature by a warm-up buzz previous to full song produc-
tion ( Sanborn, 2001 ).
In addition, none of the time parameters during vibrational
sounding in P. turionellae is thermally influenced with re-
spect to the average values. This provides further evidence
for a regulation of body temperature. Other arthropods show a
strong temperature dependence of temporal signal structures.
For example, temporal signal patterns of acoustic communi-
cation in crickets are strongly linked to temperature ( Pires &
Hoy, 1992 ). Likewise, the duration of most signal parameters
during vibratory courtship communication of the spider
Cupiennius salei are thermally affected, although there are
also temperature-invariant parameters, such as the duty cycle
( Shimizu & Barth, 1996 ). Whereas average values of time pa-
rameters in P. turionellae are unchanged by ambient tempera-
ture, their variation changes and is much stronger at extreme
than at moderate temperatures. A higher individual variability
to deal with suboptimal conditions might underlie this find-
ing. The wasp population may consist of more or less capable
regulators on the extremes. Furthermore, this finding corre-
sponds to the strong thermal dependence of responsiveness,
activity and precision of mechanosensory host location in this
species ( Samietz et al. , 2006 ), which could be also attributed
to individuals with different thermoregulatory capabilities.
The question arises as to why signals are not characterized
by a constant high intensity and frequency over a broad range
of temperature under a postulated thermoregulation. Whereas
the former focuses on the production side of the signals, the
Fig. 5. Durations of pulse trains (PT), pulse train periods (PTP)
and inter pulse durations (IPD) of Pimpla turionellae at 8, 14, 20 and
26 °C. Midlines indicate medians; dashed lines indicate means.
Boxes range from the 25th to 75th percentile. Error lines extend
from the 10th to 90th percentile. Multiple comparisons (CV) refer to
significant different coefficients of variation between temperatures
and are indicated by different letters (pairwise Z -test with adjusted
P -values by sequential Bonferroni adjustment).

Page 5

Temperature affects vibrational sounding 111
© 2006 The Authors
Journal compilation © 2006 The Royal Entomological Society, Physiological Entomology, 32, 105–112
observed inverse pattern of carrier frequency and intensity of
vibrations in this parasitoid can be related to the receiving
side of the sensory system. The subgenual organ by which
the host-searching female receives the self-produced vibra-
tions is located in the tibiae ( Otten et al. , 2002 ), and the pro-
posed heterothermy would hardly affect the temperature
there. Due to the high surface-volume ratio in the small and
elongated tibiae, an endothermic effect can be excluded
( Heinrich, 1993 ). Heated haemolymph would be cooled
down to operative environmental temperature as it flows into
the legs. However, in mechanoreceptors of arthropods, sub-
stantial thermal effects on mechanotransduction, action po-
tential encoding, and action potential conduction are well
known ( French, 1985; Coro & Perez, 1990; Franz &
Ronacher, 2002 ). Decreasing temperature leads to higher re-
sponse thresholds and a lower sensitivity to physical stimuli
( Coro & Perez, 1990; Franz & Ronacher, 2002 ). Therefore, a
higher amplitude of the vibratory signal has a quantitative
impact on displacement, velocity and acceleration of the vi-
bration receptor in insects and results in stronger stimulation
of the linked sensory nerves ( Greenfield, 2002 ). With higher
intensity and frequency of vibrations, P. turionellae produces
more reliable signals as stronger intensity leads to higher
precision of the host-searching activity ( Otten et al. , 2001 ).
In acoustic communication of crickets, temperature-
dependent variability in calling song production and accord-
ingly adjusted phonotactic preference of the receiving females
is referred to as ‘temperature coupling’ ( Pires & Hoy, 1992 ).
In the case of vibrational sounding in P. turionellae , the tem-
perature dependent signal production and signal detection
would represent a further kind of temperature coupling. The
parameters of the self-produced vibrations in the present study
are most likely adjusted to the shift of the response threshold
in the mechanoreceptors of the parasitoid at low temperatures.
Based on the present results, it is postulated that female
wasps compensate for ambient temperature changes with their
muscle activity to achieve an independence of ambient tem-
perature during vibrational sounding. By means of hetero-
thermy, wasps can control their body temperature and further
adjust the vibratory signals to an optimum. The hypothesis of
heterothermy during vibrational sounding is supported by fe-
males being able to elevate vibratory signals at low tempera-
tures. Such increased intensity and frequency result in more
reliable signals that balance negative thermal influences on
the mechanoreceptors. A temperature-coupled signal produc-
tion can be considered for such an adjustment of the fre-
quency domain of the vibrations. As a consequence, the
wasps would be able to use vibrational sounding successfully
in a broad range of ambient temperature, as demonstrated in a
recent study ( Samietz et al. , 2006 ). Such temperature-
coupling would have to be justified with electrophysiological
studies during the host-location process, which would be
challenging in freely moving wasps of this size. Whether or
not P. turionellae is able to maintain and regulate body tem-
perature during vibrational sounding can be assessed in rela-
tively easy behavioural experiments with plant-stem models,
including thermographic real-time measurements of body
temperature during the host-location process.
Acknowledgements
We thank Kathrin Tschudi-Rein, Jim Hardie and two anony-
mous reviewers for valuable comments on earlier drafts of
the paper. The study was supported by a grant of the ETH
Department of Agriculture and Food Sciences to Silvia Dorn.
References
Bacher , S . , Casas , J . , Wäckers , F . & Dorn , S . ( 1997 ) Substrate vibra-
tions elicit defensive behaviour in leafminer pupae . Journal of
Insect Physiology , 43 , 945 – 952 .
Barth , F . G . ( 2002 ) Spider senses: technical perfection and biology .
Zoology , 105 , 271 – 285 .
Bennet-Clark , H . C . ( 1998 ) Size and scale effects as constraints in
insect sound communication . Philosophical Transactions of the
Royal Society of London Series B , 353 , 407 – 419 .
Bracewell , R . N . ( 1989 ) Die Fourier-transformation . Spektrum der
Wissenschaft , 6 , 90 – 99 .
Broad , G . R . & Quicke , D . L . J . ( 2000 ) The adaptive significance of
host location by vibrational sounding in parasitoid wasps . Pro-
ceedings of the Royal Society of London Series B, Biological Sci-
ences , 267 , 2403 – 2409 .
Coro , F . & Perez , M . ( 1990 ) Temperature affects auditory receptor
response in an arctiid moth . Naturwissenschaften , 77 , 445 – 447 .
Djemai , I . , Casas , J . & Magal , C . ( 2001 ) Matching host reactions to
parasitoid wasp vibrations . Proceedings of the Royal Society of
London Series B, Biological Sciences , 268 , 2403 – 2408 .
Fischer , S . , Samietz , J . & Dorn , S . ( 2003 ) Efficiency of vibrational
sounding in parasitoid host location depends on substrate density .
Journal of Comparative Physiology A , 189 , 723 – 730 .
Fischer , S . , Samietz , J . & Dorn , S . ( 2004 ) Host location of a pupal par-
asitoid in a tritrophic system compared to a model offering mecha-
nosensory cues only . Journal of Insect Behavior , 17 , 191 – 199 .
Franz , A . & Ronacher , B . ( 2002 ) Temperature dependence of tempo-
ral resolution in an insect nervous system . Journal of Compara-
tive Physiology A , 188 , 261 – 271 .
French , A . S . ( 1985 ) The effects of temperature on action-potential
encoding in the cockroach tactile spine . Journal of Comparative
Physiology A , 156 , 817 – 821 .
Gogala , M . ( 1985 ) Vibrational communication in insects (biophysi-
cal and behavioural aspects) . Acoustic and Vibrational Communi-
cation ( ed . by K . Kalmring and N . Elsner ), pp . 117 – 134 . Paul
Parey , Germany .
Greenfield , M . D . ( 2002 ) Signalers and Receivers – Mechanisms and
Evolution of Arthropod Communication . Oxford University Press ,
U.K .
Heinrich , B . ( 1993 ) The Hot-Blooded Insects – Strategies and Mech-
anisms of Thermoregulation . Springer , Germany .
Heinrich , B . & Kammer , A . E . ( 1973 ) Activation of the fibrillar mus-
cles in the bumblebee during warm-up, stabilization of thoracic
temperature and flight . Journal of Experimental Biology , 58 ,
677 – 688 .
Henaut , A . ( 1990 ) Study of the sound produced by Pimpla instigator
(Hymenoptera, Ichneumonidae) during host selection . Entomoph-
aga , 35 , 127 – 139 .
Hill , P . S . M . ( 2001 ) Vibration and animal communication: a review .
American Zoologist , 41 , 1135 – 1142 .
Hrncir , M . , Jarau , S . , Zucchi , R . & Barth , F . G . ( 2004 ) Thorax vi-
brations of a stingless bee ( Melipona seminigra ) . I. No influence
of visual flow. Journal of Comparative Physiology A , 190 ,
539 – 548 .