Extended amplification of acoustic signals by amphibian burrows

Abstract

Animals relying on acoustic signals for communication must cope with the constraints imposed by the environment for sound propagation. A resource to improve signal broadcast is the use of structures that favor the emission or the reception of sounds. We conducted playback experiments to assess the effect of the burrows occupied by the frogs Eupsophus emiliopugini and E. calcaratus on the amplitude of outgoing vocalizations. In addition, we evaluated the influence of these cavities on the reception of externally generated sounds potentially interfering with conspecific communication, namely, the vocalizations emitted by four syntopic species of anurans (E. emiliopugini, E. calcaratus, Batrachyla antartandica, and Pleurodema thaul) and the nocturnal owls Strix rufipes and Glaucidium nanum. Eupsophus advertisement calls emitted from within the burrows experienced average amplitude gains of 3–6 dB at 100 cm from the burrow openings. Likewise, the incoming vocalizations of amphibians and birds were amplified on average above 6 dB inside the cavities. The amplification of internally broadcast Eupsophus vocalizations favors signal detection by nearby conspecifics. Reciprocally, the amplification of incoming conspecific and heterospecific signals facilitates the detection of neighboring males and the monitoring of the levels of potentially interfering biotic noise by resident frogs, respectively.

Full text

1 3
J Comp Physiol A
DOI 10.1007/s00359-016-1093-0
ORIGINAL PAPER
Extended amplification of acoustic signals by amphibian burrows
Matías I. Muñoz1 · Mario Penna1
Received: 5 December 2015 / Revised: 9 May 2016 / Accepted: 12 May 2016
© Springer-Verlag Berlin Heidelberg 2016
Introduction
The ability of a receiver to detect a distantly emitted acous-
tic signal will depend on the interplay of diverse factors,
including the frequency of the sound emitted, the level of
interfering noise, the auditory sensitivity of the receiver,
and the amplitude of the acoustic signal at the receiver’s
position (Bradbury and Vehrencamp 2011). Together, these
variables determine the distance at which an acoustic signal
can be detected, namely, its active space.
In this framework, the amplitude of the signal at the
source strongly influences the effectiveness of commu-
nication. Sounds broadcast at relatively low amplitudes
are likely to propagate over shorter distances than louder
sounds, thus restricting their detection range. In addi-
tion, weak sounds are susceptible of acoustic masking by
moderate levels of background noise. To enhance the sali-
ence of their acoustic signals, various species of birds and
mammals have been shown to increase the amplitude of
their vocalizations in response to increasing levels of back-
ground noise, a behavior known as the Lombard effect
(reviewed in Brumm and Zollinger 2011). However, this
plastic response is not widespread among other soniferous
taxa, as to date there is no strong evidence of its occurrence
in insects (Römer 2013) or anurans (Schwartz and Bee
2013), although a recent study reported for the first time
evidence on the Lombard effect in one species of fish (Holt
and Johnston 2014).
Animals can also modify the amplitude of their emit-
ted signals by means of non-physiological mechanisms:
for instance, calling sites have been shown to amplify the
acoustic signals emitted from these posts. This effect has
been extensively investigated, and remarkable examples
have been reported in several taxonomic groups, including
insects (Prozesky-Schulze et al. 1975; Bennet-Clark 1987;
Abstract Animals relying on acoustic signals for com-
munication must cope with the constraints imposed by the
environment for sound propagation. A resource to improve
signal broadcast is the use of structures that favor the emis-
sion or the reception of sounds. We conducted playback
experiments to assess the effect of the burrows occupied
by the frogs Eupsophus emiliopugini and E. calcaratus
on the amplitude of outgoing vocalizations. In addition,
we evaluated the influence of these cavities on the recep-
tion of externally generated sounds potentially interfer-
ing with conspecific communication, namely, the vocali-
zations emitted by four syntopic species of anurans (E.
emiliopugini, E. calcaratus, Batrachyla antartandica, and
Pleurodema thaul) and the nocturnal owls Strix rufipes and
Glaucidium nanum. Eupsophus advertisement calls emit-
ted from within the burrows experienced average ampli-
tude gains of 3–6 dB at 100 cm from the burrow openings.
Likewise, the incoming vocalizations of amphibians and
birds were amplified on average above 6 dB inside the cavi-
ties. The amplification of internally broadcast Eupsophus
vocalizations favors signal detection by nearby conspecif-
ics. Reciprocally, the amplification of incoming conspecific
and heterospecific signals facilitates the detection of neigh-
boring males and the monitoring of the levels of potentially
interfering biotic noise by resident frogs, respectively.
Keywords Anura · Burrows · Sound amplification · Sound
emission · Sound reception
* Matías I. Muñoz
munozsandoval@ug.uchile.cl
1 Program of Physiology and Biophysics, Institute
of Biomedical Sciences, Faculty of Medicine, University
of Chile, Independencia, 838000 Santiago, Chile

J Comp Physiol A
1 3
Bailey et al. 2001), fishes (Lugli 2012, 2013; Kéver et al.
2014), mammals (Lange et al. 2007; Schleich and Ante-
nucci 2009; Chaverri and Gillam 2013), and amphibians
(Bailey and Roberts 1981; Lardner and bin Lakim 2002;
Tan et al. 2014). Among refugia used as calling posts, a
wide range of structures have been reported to amplify out-
going acoustic signals: tree-holes (Lardner and bin Lakim
2002), bivalve shells (Kéver et al. 2014), tubular leaves
(Chaverri and Gillam 2013), cavities excavated at ground
level (Bailey and Roberts, 1981) or underneath stones or
shells (Lugli 2012), and even human-made rain drains (Tan
et al. 2014).
These structures can also modify the amplitude of
incoming acoustic signals generated outside, such as the
signals emitted by nearby conspecifics. Playback experi-
ments have shown that the songs emitted by great tits
(Parus major) are attenuated inside the artificial nest boxes
used as shelters by fertile females of this species (Blumen-
rath et al. 2004). Another study in birds has shown that the
tree holes occupied by female black-capped chickadees
(Poecile atricapillus) have directional acoustic proper-
ties, so that songs broadcast in front of the entrance have
a larger amplitude than the incoming songs broadcast from
other directions (Mennill and Ratcliffe 2004). In contrast
to these studies reporting acoustic drawbacks of shelters
used by birds, studies in anurans of the genus Eupsophus
have shown that burrows occupied by male frogs amplify
calls emitted by neighboring conspecifics (Penna and Solís
1996, 1999; Penna 2004; Penna and Marquéz 2007). In
consonance with these findings in anurans, a recent study
has shown that tubular leaves used as roosting sites by bats
(Thyroptera tricolor) amplify incoming conspecific signals
(Chaverri and Gillam 2013).
The amplification of acoustic signals entering and
leaving these posts likely facilitates the maintenance of
vocal interactions among individuals of the same species.
However, acoustic signals emitted by many different spe-
cies are also main components of local soundscapes (e.g.,
Amézquita et al. 2011; Sueur 2002) and influence the
vocal activity of focal species. For example, neotropical
birds avoid overlaps in the timing of their vocal displays
with the calling activity of a species of cicada (Hart et al.
2015; Stanley et al. 2016). Similarly, the frog Limnody-
nastes convexiusculus reduces its vocal activity in the
presence of vocalizations of cane toads (Rhinella marina)
(Bleach et al. 2015). In addition, noises of abiotic ori-
gin, such as the natural noises of wind, waterfalls, and
rain, are ubiquitous and have been shown to influence
the vocal activity of different animals (Penna et al. 2005;
Penna and Hamilton-West 2007; Gough et al. 2014).
Yet, the influence of refugia on the reception of envi-
ronmental sounds other than conspecific signals remains
unexplored.
Male frogs Eupsophus emiliopugini and E. calcaratus
in the South American temperate forest emit advertise-
ment calls from inside partially flooded burrows excavated
among mosses and ferns (Fig. 1). Conspecific females are
attracted by these calls to the burrows inhabited by males,
where mating and oviposition take place. Females thereaf-
ter, leave the cavities and males stay providing egg attend-
ance (Úbeda and Nuñez 2006). These two species breed in
the same environment, overlapping partially their repro-
ductive periods during mid spring. Their advertisement
calls have contrasting structures: the call of E. calcaratus
is a single note having a harmonic structure, susceptible
to spectral degradation, whereas the call of E. emiliopug-
ini has a lower frequency spectrum and a pulsed structure
vulnerable to temporal degradation (Penna and Moreno-
Gómez 2015). The previous studies have also compared
the active acoustic space of both taxa, which is restricted to
distances below 2 m for E. calcaratus and extends beyond
8 m for E. emiliopugini (Penna et al. 2013; Penna and
Moreno-Gómez 2014).
In the current study, we evaluated the acoustic proper-
ties of the burrows occupied by these two species for sig-
nal broadcast and for reception of different sounds of biotic
origin. For this purpose, we carried out playback experi-
ments with small loudspeakers placed inside the burrows to
test the effects of these shelters on the amplitudes of male
outgoing calls and evaluated by means of playbacks the
effect of the burrows on the amplitudes of diverse external
sounds of biotic origin present in the breeding areas poten-
tially interfering conspecific communication. These meas-
urements were complemented with acoustic monitoring of
outgoing and incoming signals produced by animals vocal-
izing at the study site. The diverse measurements carried
Fig. 1 Resident Eupsophus calcaratus male inside a burrow (photo
courtesy of Daniela Díaz)

J Comp Physiol A
1 3
out sought to provide an overall assessment of the rele-
vance of acoustic properties of these structures for acoustic
communication.
We expect that the burrows where Eupsophus males call
from amplify incoming conspecific vocalizations, as shown
in previous studies (Penna and Solís 1996, 1999; Penna
2004; Penna and Marquéz 2007). In addition, we predict
that these cavities increase the amplitude of outgoing Eup-
sophus vocalizations, as has been shown to occur for refu-
gia of other anurans (Bailey and Roberts 1981; Lardner and
bin Lakim 2002; Tan et al. 2014). Furthermore, we hypoth-
esize that incoming signals of other animals inhabiting the
temperate austral forest are also amplified inside the bur-
rows, as the resonant properties of the burrows found in the
previous studies (Penna and Solís 1996, 1999; Penna 2004;
Penna and Marquéz 2007) encompass the spectral range of
calls of other anurans and nocturnal birds present in Eupso-
phus breeding sites (see “Materials and methods”).
Materials and methods
Study site
The study was conducted in 2012 and 2013, during the
months of september, october (E. calcaratus burrows), and
november (E. emiliopugini burrows), at the locality of La
Picada (41°06′S, 72°30′W, 820 m.a.s.l.), within the Vicente
Pérez Rosales National Park in southern Chile. The study
site was a bog of volcanic substrate, where males of E. cal-
caratus and E. emiliopugini call from inside small burrows
along the borders of small streams or pools among vegeta-
tion composed mainly of mosses (Rhacomytrium), grasses
(Scyrpus and Myrteola), and ferns (Blechnum).
Experimental procedures
Playback emission experiments
For experiments in which sounds were broadcast from
inside the burrows, we created an audio file (44.1 kHz
and 16 bits) containing white noise, 38 pure tones of 1-s
duration, and advertisement calls of 11 individuals of E.
emiliopugini and 16 individuals of E. calcaratus recorded
previously. The series of tones consisted of 28 pure tones
between 0.3 and 3.0 kHz in 0.1 kHz steps and ten pure
tones between 3.2 and 5.0 kHz in 0.2 kHz steps. These pure
tones encompass most of the hearing range of both Eupho-
phus species (Penna et al. 2013; Penna and Moreno-Gómez
2014). Tones below 0.3 kHz were not tested, because they
are out of the frequency response range of the small loud-
speakers placed inside burrows (see below). For each spe-
cies of Eupsophus, five different calls from each individual
were included in the file, resulting in a total of 55 E. emili-
opugini and 80 E. calcaratus calls. Mean ± SD call dura-
tion and dominant frequency of E. emiliopugini calls were
265 ± 47 ms and 1036 ± 223 Hz, respectively. For E. cal-
caratus calls, mean ± SD call duration was 309 ± 41 ms,
and the frequency of the second and third harmonics, the
main spectral components of these signals, were 1395 ± 89
and 2074 ± 131 Hz, respectively.
The audio file was played back with an Ipod nano
(Apple Inc., Cupertino, CA, USA) connected to a custom
made amplifier based on an LM 2002 integrated circuit,
fed into a small loudspeaker having an ellipsoidal shape
(diameters 2.5 and 2.0 cm) housed in a 2-cm length rub-
ber cylinder. This enclosure was filled with mineral wool
to minimize internal resonances. The loudspeakers were
obtained from cellular telephones (Samsung SGH M310)
and were chosen because their frequency responses within
the range of the main components of Eupsophus vocali-
zations (1.0–2.5 kHz) were the flattest among other small
loudspeakers from the same model and from other brands
tested. During the experiments, we used two loudspeak-
ers of the same model, because the one used initially was
replaced after it plunged into the water of a burrow bottom.
The frequency responses for tone frequencies between 0.3
and 0.7 kHz were within ±10 and ±12 dB for the first and
second loudspeaker, respectively, and within ±7 dB for fre-
quencies between 0.8 and 5.0 kHz for both loudspeakers.
The loudspeaker used in an experiment was placed either
inside (Fig. 2a) or on the border (Fig. 2b) of a burrow in
which a resident male of any of the two species had been
observed calling and in some cases recorded on nights pre-
vious to the experiments. At the border position, the active
face of the loudspeaker was on the plane of the opening,
and for the inside position, it was 1–8 cm inside this level,
depending on the water level inside the cavity. Small tie-
clip microphones (Sennheiser MKE 2 with a K6 powering
module, Wedemark, Germany) were placed at distances of
0, 6.25, 12.5, 25, 50, and 100 cm in front of the burrow
opening. The microphone at the 0 position was 2–3 mm
apart from the loudspeaker when this transducer was placed
at the burrow opening. The sounds delivered from the loud-
speaker at the two positions tested for each burrow were
recorded simultaneously with the six microphones using a
six-channel digital recorder (Tascam DR-680, Montebello,
CA, USA).
Playback reception experiments
For the experiments in which sounds were broadcast from a
loudspeaker positioned outside the burrows, a second audio
file was created (44.1 kHz and 16 bits). This file contained
the same series of pure tones used in the emission experi-
ments (38 one-second duration tones from 0.3 to 5.0 kHz),

J Comp Physiol A
1 3
the calls from the same 16 individuals of E. calcaratus,
and the same 11 individuals of E. emiliopugini. In addi-
tion, calls from another four males of E. emiliopugini were
included, resulting in a total of 15 individuals of this spe-
cies. The audio file also included the vocalizations of ten
Pleurodema thaul males and 12 Batrachyla antartandica
males, two anuran species that breed at the study site. This
file contained five calls of each individual of E. calcara-
tus and E. emiliopugini and five pulses of one call of each
individual of B. antartandica and P. thaul. Therefore, the
number of signals analyzed was 80 for E. calcaratus, 75 for
E. emiliopugini, 50 for P. thaul, and 60 for B. antartandica.
The file also contained ten trains of ten pulses of one indi-
vidual of Glaucidium nanum and ten calls of one individual
of Strix rufipes, two owl species that call at the study site.
The dominant frequency and call duration of the vocaliza-
tions contained in the audio file used for playback reception
experiments are summarized in Table 1. An oscillogram
and a power spectrum of a representative call of each spe-
cies of anuran and bird are shown in Fig. 3.
The audio file was played back with an Ipod nano
(Apple Inc., Cupertino, CA, USA) connected to an attenu-
ator (Hewlett-Packard 355–3560) and an amplifier (Alpine
3540) connected to a 10-cm diameter loudspeaker (Ver-
satec, frequency response ±6 dB between 0.5 and 5.0 kHz)
placed 65–70 cm away from the burrow openings. Two small
microphones (RadioShack 33–3013) were placed inside and
outside a burrow. The inner microphone was placed at the
approximate location, where resident Eupsophus males typi-
cally call, and the outside microphone was located at 2–4 cm
away from the burrow opening with its axis oriented on the
same direction as the inner microphone. The disposition of
the loudspeaker and microphones is schematized in Fig. 2c.
Fig. 2 Position of the loud-
speaker a inside and b on the
border of a burrow, and disposi-
tion of six microphones for
playback emission experiments.
c Position of the loudspeaker
and two microphones for play-
back reception experiments
Table 1 Dominant frequency and call duration (mean ± SD) of the
frog and bird vocalizations included in the audio file used for play-
back reception experiments
For Eupsophus calcaratus calls, frequencies of the second and third
harmonics are given. The vocalizations of Glaucidium nanum were
long trills lasting between 18 and 49 s and containing between 66 and
169 pulses. From these long calls, ten trains of ten pulses each were
edited for the audio file, and therefore, call duration for this signal is
not listed
Species Number of indi-
viduals
Dominant
frequency
(Hz)
Call duration (ms)
Eupsophus cal-
caratus
16 1395 ± 89 309 ± 41
2074 ± 131
Eupsophus emili-
opugini
15 1059 ± 197 263 ± 48
Pleurodema thaul 10 1654 ± 112 79 ± 15
Batrachyla
antartandica
12 1919 ± 205 16 ± 7
Strix rufipes 1 2075 2960
Glaucidium
nanum
1 1385 –

J Comp Physiol A
1 3
To avoid acoustic interferences during the recordings,
playback emission and reception experiments were carried
out during the day, between 09:00 and 21:00 h, before the
beginning of the vocal activity of the native anuran species
and nocturnal birds. All the calls included in both audio
files were recorded in the previous years at the study site
at an approximate distance of 20–50 cm for anurans and
15–20 m for birds, using a directional microphone (Sen-
nheiser ME 66, Wedemark, Germany) and a digital recorder
(Tascam DR-100, Montebello, CA, USA).
Natural emission and reception recordings
For burrows of the two species (three E. emiliopugini and
five E. calcaratus burrows), where resident males called
from actively, calls produced were recorded at night with
two microphones simultaneously (Sennheiser MKE 2),
one placed at the burrow border and the other 25 cm in
front of the burrow. Ten vocalizations of each animal were
recorded, and the attenuation experienced by these natural
signals was compared to those measured in playback emis-
sion experiments conducted for these same burrows.
Eleven out of 12 E. emiliopugini and 10 out of 12 E. cal-
caratus burrows used for playback reception experiments
were used to record the calls emitted by nearby frogs. Two
to ten vocalizations of each neighboring individual were
recorded simultaneously with two small microphones
(RadioShack 33–3013) placed at the same positions as in
the playback reception experiments (see preceding sub-sec-
tion). These measurements were analyzed using the same
procedures as for the playback reception experiments. Both
measurements, natural emission and reception, were car-
ried out at night, between 21:00 and 24:00 h, when anurans
present at the study site were vocally active.
For the four types of experiments conducted in this
study, the microphones used were calibrated after each
recording session by recording the 93.8 dB SPL 1-kHz
pure tone of a portable calibrator (B&K 4231, Brüel &
Kjær, Nærum, Denmark) at the same recording level of the
digital recorder used for registering tones and vocalizations
during the experiments.
Burrow measurements
For each tested burrow, the diameter of the opening (cm),
length of the segment free from water (cm), and total length
(cm) were measured to the nearest centimeter with a flex-
ible tape, and the inclination of the burrows relative to the
horizontal was registered. The inner dimensions of the bur-
rows were confirmed with a micro-inspection video cam-
era fitted with flexible optic fibre (Ridgid SeeSnake micro,
Ridge Tool Company, Elyria, Ohio, USA).
−100
−50
0
−100
−50
0
−100
−50
0
Frequency (kHz)
Relative
amplitude (dB)
−100
−50
0
0.1 s
Eupsophus emiliopugini
Eupsophus calcaratus
Pleurodema thau
l
Batrachyla antartandica
F2 F3
a
−100
−50
0
Frequency (kHz)
−100
−50
0
012345
012345
012345
012345
0 1 2 3 4
5
01234
5
Strix rufipes
Glaucidium nanum
1
s
b
Relative
amplitude (dB)
Fig. 3 Oscillograms and power spectra of representative calls from
each species of a anurans (left columns) and b birds (right columns)
tested. The arrows in the power spectrum of E. calcaratus call show
the second (F2) and third (F3) harmonics of this signal. All the power
spectra were computed with a Hamming window of 3170 points (fre-
quency resolution 13.9 Hz)

J Comp Physiol A
1 3
Signal analysis
All the recordings were high-pass filtered at 0.2 kHz to
minimize low-frequency environmental noise. The root
mean square (RMS) amplitude of tones and calls was
measured with Raven Pro 1.4 (Cornell Lab of Ornithology,
Ithaca, NY, USA). For the playback emission and recep-
tion experiments, the RMS amplitudes of the five vocali-
zations of each individual of the four anuran species were
averaged. As bird vocalizations included in the audio file
used for reception experiments corresponded to a single
individual of each species, the calls from these individuals
were not averaged. The RMS amplitude values of calls and
tones were converted to sound pressure levels (SPLs; dB re.
20 μPa), after adjustment based on the recorded calibration
tone.
To measure the extent to which the burrows ampli-
fied sounds, for the playback emission experiments, we
computed the differences between the SPLs of the signals
(tones and vocalizations) broadcast from inside and from
the border of the burrows, so that positive amplitude gains
correspond to increase in amplitude for the inside relative
to the border loudspeaker position. For the playback recep-
tion experiments and natural reception experiments, we
used the same procedure to compare recordings obtained
with microphones placed inside and outside the burrows.
Statistical analysis
All statistical analyzes and figures were performed with R
(version 3.0.2, R Core Team 2013). Mean and SD of dB
values were obtained with the library “seewave” (version
1.7.3, Sueur et al. 2008).
The data of measurements performed in burrows of E.
emiliopugini and E. calcaratus were analyzed separately.
For each experiment, we fitted linear mixed models (LMM)
by maximum likelihood with the library “lme4” (version
1.1–7, Bates et al. 2004). Maximum likelihood was chosen
instead of restricted maximum likelihood to obtain P val-
ues of fixed effects by means of sequential likelihood ratio
tests (Bolker et al. 2009). All the fitted models included the
dependent variable (i.e., amplitude gain) in linear scale (N/
m2). Nevertheless, log transformation was performed to
attain the normality assumption, which together with homo-
scedasticity was evaluated by visually inspecting the resid-
uals. Models fitted to analyze the playback emission exper-
iments included the fixed effect of distance as an ordered
categorical variable. For playback emission and reception
experiments, the models fitted to analyze the tones included
the fixed effect of frequency as a numeric predictor, and the
models fitted to analyze the calls included the fixed effect
of species as a categorical variable.
Playback emission experiments
For the playback emission experiments, the recordings at
0 cm were excluded from the analysis, because when the
loudspeaker was placed at the burrow opening, the close
proximity between the microphone and the loudspeaker
constantly yielded negative amplitude gains. To analyze
the amplification of pure tones, we fitted an LMM that
included as fixed effects the distance (6.25, 12.5, 25, 50,
and 100 cm) and the frequency and a random slope of dis-
tance within each individual burrow as random effect. To
analyze the amplification of calls, we fitted an LMM that
included as fixed effects the distance (6.25, 12.5, 25, 50 and
100 cm) and the species (E. emiliopugini and E. calcara-
tus), and a random slope of distance for each individual
burrow and a crossed random intercept for each individual
frog as random effects.
Playback reception experiments
The analysis of the amplification of pure tones inside Eupso-
phus burrows was performed by fitting an LMM that included
the frequency as a fixed-effect. For the amplification of calls,
we fitted a second LMM that included the species (E. emili-
opugini, E. calcaratus, B. antartandica, P. thaul, G. nanum,
and S. rufipes) as a fixed effect. These two models incorpo-
rated a random intercept for each tested burrow.
For every model fitted, the significance of individual
fixed effects and the interactions between them were eval-
uated by means of sequential likelihood ratio tests (LRT)
performed with the library “afex” (version 0.13–145,
Singmann et al. 2015). Post hoc analyzes were performed
with the “multcomp” library (version 1.3.3, Hothorn et al.
2008). To perform the multiple comparisons, we fitted
reduced models, i.e., non-significant high-order interac-
tions among fixed effects were excluded. The P values of
all the comparisons among treatments were adjusted by
the false discovery rate (FDR). This procedure was cho-
sen instead of other possible corrections (e.g., Bonfer-
roni), because it allows to control the rate of type I error
minimizing the effect on the power of the analysis (Pike
2011).
To assess the relative influence of cavities on sound
emission and reception, we compared the amplitude gain
experienced by playback calls emitted and received from
inside the burrows for which emission and reception exper-
iments were carried out. For calls recorded during playback
emission experiments, only the vocalizations recorded at
100 cm from the burrow openings were used for this com-
parison. To make a precise comparison, the calls of four
E. emiliopugini males included in the audio file used for
playback receptions experiments but not in the audio file

J Comp Physiol A
1 3
used for emission experiments (see “Materials and meth-
ods”), and were not considered in this analysis. The ampli-
tude gains experienced by outgoing and incoming calls of
conspecific individual frogs were averaged for each burrow
and compared with two-tailed paired t tests. The amplitude
decrements of playback calls broadcast from inside bur-
rows and calls emitted by resident males inside the same
burrows were compared with two-tailed paired t tests.
Amplitude gains experienced inside burrows by incoming
playback calls and calls emitted by neighbours were also
compared with two-tailed paired t tests.
The resonant frequencies (i.e., the frequencies for which
amplification was maximum) expected for one-end-open
cylinders having lengths equal to the segment free from
water of the burrows were computed using the equation:
resonant frequency (kHz) = speed of sound in the air
(cm/s)/4 × cylinder length (cm) (Kinsler et al. 1982). For
playback reception experiments, paired t tests were used to
compare the observed resonant frequency of the burrows
and the expected resonances. The significance level used
was α = 0.05.
Results
Playback emission experiments
Burrow dimensions
Emission experiments were carried out for 22 burrows
of E. emiliopugini and 21 burrows of E. calcaratus. The
lengths of the segments free from water were (mean ± SD),
8.32 ± 3.30 cm for E. emiliopugini, and 6.86 ± 2.92 cm
for E. calcaratus. The diameters of the openings were
3.33 ± 0.77 cm for E. emiliopugini and 3.47 ± 0.77 cm
for E. calcaratus. The total lengths of the burrows were
14.14 ± 4.03 cm for E. emiliopugini and 11.43 ± 2.54 cm
for E. calcaratus. Most of E. emiliopugini (15 out of 22)
and E. calcaratus (15 out of 21) burrows had inclinations
above 45o relative to the horizontal plane.
Amplification of pure tones
Pure tones broadcast from inside E. emiliopugini and E.
calcaratus burrows showed similar amplification pat-
terns: on average, frequencies between 1.0–2.5 kHz expe-
rienced higher amplifications relative to frequencies out-
side this range at all recording distances (Fig. 4a, b). The
mean ± SD resonant frequencies of E. emiliopugini bur-
rows were 1.8 ± 1.0, 1.8 ± 1.0, 1.9 ± 0.8, 2.0 ± 1.0,
and 1.6 ± 0.6 kHz recorded at 6.25, 12.5, 25, 50, and
100 cm from the burrow openings, respectively. The
mean ± SD resonant frequencies of E. calcaratus burrows
were 1.9 ± 0.9, 2.0 ± 0.8, 2.1 ± 0.8, 2.2 ± 0.9, and
2.0 ± 0.9 kHz measured at 6.25, 12.5, 25, 50, and 100 cm
from the burrow openings, respectively.
The effects of frequency and recording distance on
the amplification attained by pure tones broadcast from
the burrows of E. emiliopugini were statistically signifi-
cant (χ2 = 619.18, df = 1, P < 0.0001 and χ2 = 21.56,
df = 4, P = 0.0002, respectively). Similar results were
obtained for the burrows of E. calcaratus (χ2 = 469.80,
df = 1, P < 0.0001 and χ2 = 15.88, df = 4, P = 0.0032,
respectively). The interaction between the two effects
was not significant in burrows of either species (E. emili-
opugini χ2 = 3.57, df = 4, P = 0.4671, and E. calcara-
tus χ2 = 6.71, df = 4, P = 0.1521). Pairwise comparisons
showed that when broadcast from inside E. emiliopugini
and E. calcaratus burrows, the overall amplification expe-
rienced by tones at 6.25 cm was lower relative to farther
distances (P < 0.001, for all comparisons). In addition,
for E. emiliopugini burrows, amplification at 12.5 cm was
lower than at 50 cm (z = 3.21, P = 0.0022) and 100 cm
(z = 3.57, P = 0.0007). The rest of the pairwise compari-
sons among distances were not significant in both the types
of burrows.
The length of the segment free from water of E. emili-
opugini burrows was not related to their resonant frequency
(linear regression, P > 0.2 for all recording distances).
However, the length of the segment free from water and
resonant frequency of E. calcaratus burrows were nega-
tively associated, although only for tones measured at 6.25,
12.5, and 100 cm (Fig. 5).
Amplification of calls
The calls of both species of Eupsophus broadcast from
inside the burrows were amplified on average about
3–6 dB, at all recording distances (Fig. 4c, d).
When broadcast from E. emiliopugini burrows, the
amplification of calls was affected by the interaction
between species and recording distance (χ2 = 29.91,
df = 4, P < 0.0001). Planned comparisons revealed that E.
calcaratus calls measured at 6.25 cm experienced lower
amplification relative to vocalizations recorded at far-
ther distances (P < 0.001, for all comparisons) and calls
recorded at 12.5 cm experienced lower amplification than
calls measured at farther distances (P < 0.05, for all com-
parisons). Similarly, E. emiliopugini vocalizations recorded
at 6.25 cm were significantly less amplified than at 12.5
(z = −2.83, P = 0.0115), 25 (z = −2.66, P = 0.0139),
and 50 cm (z = −2.60, P = 0.0157). Comparisons between
the calls of both the species of Eupsophus showed that E.
emiliopugini calls were significantly more amplified than
E. calcaratus calls at all the distances measured (P < 0.05,
for all comparisons).

J Comp Physiol A
1 3
When calls were emitted from inside E. calcaratus bur-
rows, the vocalizations of E. emiliopugini were amplified
to a larger extent than E. calcaratus calls, and the dif-
ference had a marginal significance (χ2 = 3.85, df = 1,
P = 0.0498). The amplification attained by calls also dif-
fered among recording distances (χ2 = 15.68, df = 4,
P = 0.0035). Calls measured at 6.25 cm from the bur-
row openings experienced lower amplification than calls
recorded at farther distances (P < 0.05, for all compari-
sons) and calls recorded at 25 and 50 cm experienced lower
amplification than at 100 cm (z = 2.22, P = 0.0475 and
z = 2.71, P = 0.0226, respectively).
The length of the segment free from water of E. emili-
opugini burrows was not related to the amplification
attained by E. emiliopugini calls (linear regression, P > 0.1
for all recording distances). For E. emiliopugini burrows,
the amplification experienced by E. calcaratus calls was
inversely related to the length of the segment free from
water of these cavities, only for vocalizations recorded
at 6.25, 12.5, and 50 cm (Fig. 6a). In E. calcaratus bur-
rows, the calls of E. calcaratus and E. emiliopugini expe-
rienced larger amplification inside cavities with intermedi-
ate lengths of the segment free from water (Fig. 6b, c). All
other regressions were not significant.
Natural emission recordings
Calls emitted by resident males from burrows for which
playback emission experiments were carried out experi-
enced amplitude losses similar to those measured experi-
mentally. The advertisement calls emitted by three E. emili-
opugini individuals experienced a mean ± SD amplitude
drop of 14.25 ± 1.84 dB between 0 and 25 cm, while E.
emiliopugini calls played back from inside the same three
−15
−10
−5
0
5
10
15
012345
Frequency (kHz)
Amplitude gain (dB)
E. emiliopugini burrows
−15
−10
−5
0
5
10
15
012345
Frequency (kHz)
E. calcaratus burrows
−15
−10
−5
0
5
10
15
6.25 12.52550 100
Distance (cm)
Amplitude gain (dB)
E. emiliopugini burrows
−15
−10
−5
0
5
10
15
6.25 12.5 25 50 100
Distance (cm)
E. calcaratus burrows
Distance
6.25 cm
12.5 cm
25 cm
50 cm
100 cm
Species
E. calcaratus
E. emiliopugini
ba
dc
Fig. 4 Mean amplitude gains for pure tones emitted from the bur-
rows of a E. emiliopugini (N = 22) and b E. calcaratus (N = 21) and
recorded at 6.25, 12.5, 25, 50, and 100 cm from the burrow open-
ings. Amplitude gains (mean ± SD) experienced by Eupsophus calls
emitted from the burrows of c E. emiliopugini and d E. calcaratus
recorded at different distances from the burrows openings

J Comp Physiol A
1 3
burrows experienced a mean ± SD amplitude drop of
12.59 ± 3.54 dB. These values did not differ statistically
(paired t test, t = 1.56, df = 2, P = 0.2583). Likewise, the
advertisement calls emitted by five individuals of E. calcar-
atus from inside their corresponding burrows experienced
a mean ± SD amplitude drop of 17.20 ± 3.10 dB between
0 and 25 cm, while the amplitude of E. calcaratus calls
played back and broadcast with the loudspeaker placed
inside the same five burrows dropped 20.81 ± 3.88 dB.
These values did not differ statistically (paired t test,
t = −2.03, df = 4, P = 0.1121).
Playback reception experiments
From the total of the burrows for which emission playback
experiments were carried out, a subset of 12 burrows of
each species of Eupsophus was used to perform playback
reception experiments.
Amplification of pure tones
Pure tones broadcast from a loudspeaker located 60–70 cm
away from the burrows showed amplification patterns
similar to those found in emission experiments; on aver-
age frequencies between 1.0 and 2.5 kHz experienced
0
1
2
3
4
5
24681012
Length of the segment free from water (cm)
Resonant frequency (kHz)
6.25 cm 12.5 cm 100 cm
E. calcaratus burrows
Fig. 5 Negative association between the length of the segments free
from water and the resonant frequency of E. calcaratus burrows for
tones measured at 6.25 (r2 = 0.26, P = 0.0188), 12.5 (r2 = 0.27,
P = 0.0169), and 100 cm (r2 = 0.35, P = 0.0050). Different colors
depict different recording distances: 6.25 (red), 12.5 (yellow), and
100 cm (purple). Only significant regression lines are shown
−10
−5
0
5
10
15
246810 12 14
Length of the segment free from water (cm)
Amplitude gain (dB)
6.25 cm 12.5 cm 50 cm
E. emiliopugini burrows
E. calcaratus calls
−10
−5
0
5
10
15
24681012
Length of the segment free from water (cm)
Amplitude gain (dB)
6.25 cm 12.5 cm 25 cm 50 cm 100 cm
E. calcaratus burrows
E. calcaratus calls
−10
−5
0
5
10
15
246810 12
Length of the segment free from water (cm)
Amplitude gain (dB)
6.25 cm 12.5 cm 25 cm
E. calcaratus burrows
E. emiliopugini calls
abc
Fig. 6 a Negative association between the length of the segment
free from water and the amplitude gains experienced by E. calcar-
atus calls broadcast from E. emiliopugini burrows and measured
at 6.25 (r2 = 0.25, P = 0.0184), 12.5 (r2 = 0.28, P = 0.0119), and
50 cm (r2 = 0.18, P = 0.0491). b Quadratic association between
the length of the segment free from water and the amplitude gains
experienced by E. calcaratus calls broadcast from E. calcaratus bur-
rows and measured at 6.25 (r2 = 0.35, P = 0.0210), 12.5 (r2 = 0.38,
P = 0.0128), 25 (r2 = 0.47, P = 0.0034), 50 (r2 = 0.31, P = 0.0374),
and 100 cm (r2 = 0.41, P = 0.0083). c Quadratic association between
the length of the segment free from water and the amplitude gains
experienced by E. emiliopugini calls broadcast from E. calcara-
tus burrows and measured at 6.25 (r2 = 0.44, P = 0.0057), 12.5
(r2 = 0.40, P = 0.0104), and 25 cm (r2 = 0.32, P = 0.0324). Regres-
sions were fitted for the average amplification of the calls of each
species in each individual burrow and at each recording distance.
Different colors depict different recording distances: 6.25 (red), 12.5
(yellow), 25 (green), 50 (blue), and 100 cm (purple). Only significant
regression lines are shown

J Comp Physiol A
1 3
larger amplifications relative to frequencies out of this
range (Fig. 7a, b). The cavities that sheltered E. emili-
opugini males had a mean ± SD resonant frequency of
1.5 ± 0.4 kHz. This value did not differ from the reso-
nant frequencies estimated for one-end-open pipes hav-
ing the same length as the cavities (expected resonant fre-
quency = 1.6 ± 1.1 kHz; paired t test, t = 0.27, df = 11,
P = 0.7914). Similarly, for E. calcaratus burrows, the
observed (1.9 ± 0.7 kHz) and expected (1.8 ± 1.0 kHz)
resonant frequencies did not differ (paired t test, t = 0.66,
df = 11, P = 0.5233). The effect of frequency was statisti-
cally significant on the amplification experienced by tones
broadcast from the burrows of E. emiliopugini (χ2 = 66.28,
df = 1, P < 0.0001) and E. calcaratus (χ2 = 32.16, df = 1,
P < 0.0001).
The length of the segment free from water of the cavities
occupied by the two species of Eupsophus was not asso-
ciated with their resonant frequency (E. emiliopugini bur-
rows, P > 0.6 and E. calcaratus burrows, P > 0.1 for all
regressions).
Amplification of calls
When recorded from inside the burrows of both species of
Eupsophus, average amplitude gains experienced by the
calls of the four species of amphibians and the two spe-
cies of owls were between about 6–11 dB (Fig. 7c, d).
The amplification of calls differed among the six species
when broadcast to the burrows occupied by E. emiliopu-
gini (χ2 = 86.32, df = 5, P < 0.0001) and E. calcaratus
(χ2 = 19.05, df = 5, P = 0.0019). In E. emiliopugini bur-
rows, the amplification experienced by conspecific calls
was significantly lower than the calls of E. calcaratus
(z = −2.88, P = 0.0067), P. thaul (z = −3.18, P = 0.0036),
and G. nanum (z = −7.20, P < 0.0001), and no differences
occurred between the calls of E. emiliopugni and the calls
of B. antartandica (z = 0.33, P = 0.7407) and S. rufipes
(z = 1.72, P = 0.1077). Inside E. calcaratus burrows, the
amplification of conspecific calls was similar to calls of
B. antartandica (z = −1.21, P = 0.3758), E. emiliopugini
(z = 2.12, P = 0.1200), P. thaul (z = −1.98, P = 0.1200),
Fig. 7 Amplitude gains
(mean ± SD) for pure tones
broadcast from an external
loudspeaker and recorded from
inside the burrows of a E.
emiliopugini (N = 12) and b E.
calcaratus (N = 12). Amplitude
gains (mean ± SD) experienced
by the calls anuran and bird spe-
cies played back via an external
loudspeaker and recorded from
inside the burrows of c E. emili-
opugini and d E. calcaratus.
Ee, Eupsophus emiliopugini;
Ec, Eupsophus calcaratus; Ba,
Batrachyla antartandica; Pt,
Pleurodema thaul; Gn, Glaucid-
ium nanum; Sr, Strix rufipes
−40
−20
0
20
40
01234
Frequency (kHz)
Amplitude gain (dB)
E. emiliopugini burrows
−40
−20
0
20
40
5012345
Frequency (kHz)
E. calcaratus burrows
−25
−15
−5
5
15
25
Ee Ec Ba Pt Gn Sr
Species
Amplitude gain (dB)
E. emiliopugini burrows
−25
−15
−5
5
15
25
Ee Ec Ba Pt Gn Sr
Species
E. calcaratus burrows
dc
ab

J Comp Physiol A
1 3
G. nanum (z = 0.88, P = 0.4712), and S. rufipes (z = 0.61,
P = 0.5447).
The length of the segment free from water of the cavities
occupied by the two species of Eupsophus was not associ-
ated with the amplification experienced by the calls of any
of the species tested (E. emiliopugini burrows, P > 0.3 and
E. calcaratus burrows, P > 0.1 for all regressions).
Natural reception recordings
In 11 burrows occupied by E. emiliopugini males, calls
emitted by one to three conspecific nearby individuals were
recorded and vocalizations of B. antartandica males were
recorded in two of these cavities. Natural calls of E. emili-
opugini experienced similar amplitude gains as compared
to values found by means of playback reception experi-
ments conducted for the same eleven burrows (mean ± SD,
natural 7.0 ± 2.8 dB, playback 8.8 ± 5.2 dB; paired t test,
t = −1.16, df = 10, P = 0.2748). The mean ± SD ampli-
tude gains of natural and pre-recorded B. antartandica calls
were 8.8 ± 1.6 and 10.4 ± 1.6 dB, respectively.
In ten burrows occupied by E. calcaratus males, calls
emitted by one to three conspecific nearby individuals
were recorded. Calls of E. emiliopugini, B. antartandica,
and P. thaul were recorded in four, two, and one of these
cavities, respectively. Natural calls of E. calcaratus experi-
enced similar amplitude gains as compared to values found
by means of playback reception experiments conducted for
the same ten burrows (mean ± SD, natural 9.9 ± 9.1 dB,
playback 8.1 ± 4.6 dB; paired t test, t = 0.27, df = 9,
P = 0.7969). Similarly, amplitude gains of E. emiliopug-
ini natural and played back calls did not differ statistically
(mean ± SD, natural 5.9 ± 3.6 dB, playback 7.1 ± 4.0 dB;
paired t test, t = 0.98, df = 3, P = 0.3997). For B.
antartandica, natural and previously recorded calls expe-
rienced a mean ± SD amplitude gain of 10.9 ± 8.7 and
16.6 ± 9.1 dB, respectively. For P. thaul, natural and previ-
ously recorded calls experienced a mean amplitude gain of
6.5 and 9.6 dB, respectively.
Comparison between playback emission and playback
reception experiments
The amplitude gains of E. emiliopugini calls in conspe-
cific burrows (N = 12) were (mean ± SD) 7.09 ± 4.64 and
8.41 ± 5.11 dB for playback emission and reception exper-
iments, respectively. These values did not differ statistically
(paired t test, t = −1.90, df = 11, P = 0.0842). For E. cal-
caratus calls in conspecific burrows (N = 12), mean ± SD
amplitude gains were 3.44 ± 4.41 and 7.45 ± 5.07 dB for
playback emission and reception experiments, respectively.
These values differed statistically (paired t test, t = −2.33,
df = 11, P = 0.0400).
Discussion
Effect of burrows on sound emission
Measurements of amplitudes of calls produced by males of
both Eupsophus species calling from inside their burrows
yielded amplitude losses with distance similar to those
experienced by pre-recorded vocalizations broadcast from
small loudspeakers placed inside the cavities. This corre-
spondence indicates that our playback emission experi-
ments emulate closely natural signal broadcasts.
Our results show that the acoustic properties of the
cavities inhabited by two species of the genus Eupso-
phus are favorable for the emission of their advertisement
calls. The playback emission experiments show that the
burrows amplified mainly the tones of frequencies that
are within the ranges of the spectral components of the
calls of E. emiliopugini and E. calcaratus. In correspond-
ence with this pure tone amplification pattern, previously
recorded Eupsophus advertisement calls experienced
average amplitude gains of up to 6 dB. However, the
amplification attained by tones and calls was dissimi-
lar across recording distances. In general, sounds meas-
ured closer to the burrows experienced lower amplifica-
tion than the signals recorded at farther distances, which
probably resulted from the short distances between the
loudspeaker placed at the burrow openings and the clos-
est microphones, distant only a few centimeters from the
cavity borders. Overall, the maximum amplifications of
tones and calls occurred at the longest distances of 50
and 100 cm from the burrows openings. This result indi-
cates that such amplification is likely to prevail at longer
distances at which neighboring males and females posi-
tion themselves in calling assemblages, rendering the
increase in signal amplitude an alteration of communica-
tive significance.
The amplitude increases of about 3–6 dB on average
experienced by Eupsophus calls broadcast from inside bur-
rows are comparable to the values reported for other verte-
brates. For instance, the calls emitted by Spix’s disc-winged
bats (Thyroptera tricolor) from inside tubular leaves expe-
rience a 1–6 dB amplitude gain at 30 cm from the leaf tip
(Chaverri and Gillam 2013), and the advertisement calls
emitted by male Mientien tree frogs (Kurixalus idiotoocus)
from inside storm drains are amplified about 5 dB at 70 cm
(Tan et al. 2014; personal communication). Other exam-
ples of the amplification of outgoing sounds correspond to
the crickets Scapteriscus acletus and Rufocephalus sp., for
which amplitude gains of up to 24 and 20 dB have been
reported, respectively (Bennet-Clark 1987; Bailey et al.
2001). The burrows built by these insects are highly effec-
tive for signal broadcasting as compared to Eupsophus
cavities.

J Comp Physiol A
1 3
The adaptive value of the amplitude enhancement
attained by Eupsophus vocalizations emitted from inside
burrows is dependent on two counteracting selective pres-
sures. The reproductive performance of males is likely to
be improved by broadcasting signals at a higher amplitude,
increasing their detectability by conspecific females, which
typically show phonotactic responses to conspecific adver-
tisement calls (e.g., Moreno-Gómez et al. 2015). On the
other hand, the amplification effect can also facilitate the
detection of signals by unintended receivers, such as preda-
tors or parasites exploiting acoustic signals to locate poten-
tial preys or hosts (reviewed in Zuk and Kolluru 1998),
impairing the survival of calling males.
There is no evidence as to whether Eupsophus frogs call
from inside burrows because of the favorable acoustic prop-
erties of these structures, or if this acoustic phenomenon is
a by-product of the utilization of these structures as a safe
refuge for mating and subsequent parental care of the eggs.
Our measurements revealed that not all the individuals
vocalize from inside acoustically favorable cavities, sug-
gesting that the acoustic properties of these shelters are not
an indispensable feature. Other anurans are known to select
oviposition sites based on water availability, for instance,
phytotelm-breeding species choose sites with large water-
holding capacity (Lin et al. 2008) and large water volumes
(von May et al. 2009).
Effect of burrows on sound reception
The amplification experienced by Eupsophus calls emitted
during natural interactions is in agreement with the ampli-
fication obtained in playback reception experiments in E.
emiliopugini and E. calcaratus burrows. For E. calcaratus
burrows, natural calls of B. antartandica and P. thaul expe-
rienced lower amplitude gains relative to the values found
in playback reception experiments, a disparity that could
be due to the small number of natural calls of these spe-
cies recorded from inside these burrows. However, overall,
the recording of natural vocalizations validates the utiliza-
tion of pre-recorded calls as a reliable method to assess the
amplification of acoustic signals inside Eupsophus burrows.
Our measurements show that the cavities occupied by
both species of Eupsophus generate a mean amplitude
gain of at least 6 dB (i.e., a twofold amplitude increase)
for biotic sounds generated outside these structures. These
results corroborate the amplification of conspecific adver-
tisement calls inside the burrows of E. emiliopugini (Penna
and Solís 1996, 1999) and E. calcaratus (Penna 2004).
Interestingly, the conspecific signals were amplified to
a lower or similar extent than the signals of two syntopic
anuran species and two species of birds, indicating that the
cavities are not particularly tuned to conspecific communi-
cation signals.
Males of E. emiliopugini and E. calcaratus are particu-
larly sensitive to sounds of frequencies between 1.0 and
2.0 kHz (Penna et al. 2013; Penna and Moreno-Gómez
2014). This range of enhanced sensitivity matches the main
spectral components of the vocalizations emitted by the six
species tested in playback reception experiments, indicat-
ing that Eupsophus frogs are capable of detecting all these
signals. Therefore, males of E. emiliopugini and E. calcar-
atus from inside their burrows can effectively monitor the
biotic soundscape, comprising conspecific and heterospe-
cific vocalizations. Different studies have shown that ani-
mals respond to the intrusion of heterospecific sounds by
either increasing (Schwartz and Wells 1985; Phelps et al.
2007) or reducing (Wong et al. 2009; Penna and Meier
2011; Penna and Velásquez 2011) their vocal output. Stud-
ies on the responses to heterospecific acoustic signals by
Eupsophus frogs would allow for the assessment of the sig-
nificance of an enhanced reception of these sounds.
The increase in mean amplitude of the vocalizations of
both the species of Eupsophus is about 8 dB when meas-
ured inside their conspecific burrows, as revealed by play-
back reception experiments and monitoring of natural sig-
nals. This amplitude enhancement is likely to increase the
active space of these signals, which depends on the receiv-
ers’ auditory sensitivity. Advertisement calls emitted by E.
calcaratus males are soft sounds, having average ampli-
tudes of 57 and 50 dB SPL at 2 and 4 m from the signaler,
respectively (Penna and Moreno-Gómez 2015). These
amplitudes fall below the 58 dB SPL auditory threshold
measured for males of this species (Penna et al. 2013), and
therefore, acoustic communication between males is esti-
mated to be restricted to distances below 2 m (Penna et al.
2013; Penna and Moreno-Gómez 2015). However, the 8 dB
amplification of calls produced by the burrows results in a
twofold expansion of the active space, reaching up to 4 m.
The calls of E. emiliopugini males are loud relative to the
vocalizations of E. calcaratus, reaching 50 dB SPL at 8 m
from the signaler (Penna and Moreno-Gómez 2015). This
amplitude value is above the 44 dB SPL auditory thresh-
old measured for males of this species (Penna and Moreno-
Gómez 2014), and therefore, acoustic interactions are esti-
mated to occur at distances beyond 8 m. At the study site,
excess attenuation (i.e., attenuation in excess to the one
expected due to spherical spreading) of E. emiliopugini
calls increases linearly with distance (0.6 dB/m; Penna and
Moreno-Gómez 2015). This attenuation rate and the ampli-
tude boost produced by the cavities imply that these calls
could be detected by conspecific males at distances well
beyond the 8 m previously reported, up to 19 m.
The effect of burrows on signal reception is larger rela-
tive to their effect on broadcast, as shown by significant
and non-significant differences between amplification val-
ues for both experiments conducted with conspecific calls

J Comp Physiol A
1 3
in E. calcaratus and E. emiliopugini burrows, respectively.
As discussed above, increasing the amplitude of the adver-
tisement calls emitted from inside the burrows increases
the probability of detection by nearby conspecifics, but
also enhances the risk of detection by unintended receiv-
ers. Contrastingly, receiving incoming conspecific sounds
with increased amplitude has the advantage of favoring
intraspecific acoustic communication without facilitating
eavesdropping, as potential unintended receivers exploiting
propagating signals would not benefit from the amplitude
boost. Eavesdropping by predators, parasites, and parasi-
toids has been shown to be a strong selective force shap-
ing acoustic communication systems across various taxa
(e.g., Zuk et al. 2006; Schmidt and Belinsky 2013; Page
et al. 2014) Thus, the larger amplification of incoming rela-
tive to outgoing Eupsophus calls likely implies advantages
for males’ performance, favoring the maintenance of vocal
interactions among neighboring frogs without increasing
the risk of attracting predators or parasites. The poten-
tial costs and benefits associated with the amplification of
sounds incoming and outgoing from these posts requires
further formal testing.
The limited amplification of outgoing relative to incom-
ing Eupsophus calls indicates that the cavities occupied by
these frogs are not particularly suited for sound broadcast,
probably due to the relatively simple cylinder-like struc-
ture of these shelters. In contrast, mole crickets (Scapter-
iscus acletus) dig complex horn-shaped cavities that match
the impedance of the small sound-producing wings to the
air around the entrance of the cavity, resulting in highly
effective sound radiation (Bennet-Clark 1987). Similar to
Eupsophus frogs, vocal responses emitted by Thyroptera
tricolor bats from inside tubular leaves are amplified to a
lower extent than incoming inquiry calls (Chaverri and Gil-
lam 2013). In contrast to the impedance matching effect of
the horn-shaped cavities of mole crickets, the leaves used
by Spix’s disc-winged bats apparently amplify the high-
frequency vocalizations by improving the directionality of
outgoing sounds (Chaverri and Gillam 2013). In frogs Eup-
sophus, directional effects of burrows are apparently not
relevant, since the vocal responses of E. calcaratus males
are independent from the direction from which external
stimuli are broadcast (Penna and Quispe 2007). In addition,
a large proportion of these refugia has vertical orientation,
not improving the beaming of broadcast sounds to conspe-
cific individuals dwelling in the nearby soil substrate.
To summarize, the present study expands the evidence
on the effect of animal refugia on the amplitude of outgo-
ing acoustic signals and shows that this phenomenon is
concurrent with the increase in reception of incoming con-
specific signals described previously for frogs Eupsophus.
In addition, our results show that these shelters enhance
the reception of diverse sounds composing the biotic
soundscape, thus facilitating the detection of potentially
interfering acoustic signals. Further research is needed to
assess how the burrow amplification effect on outgoing
and incoming sounds may affect male vocal interactions
and female phonotactical responses in these frogs and the
relative importance of these complementary phenomena for
acoustic communication. In addition, the effect of burrows
on natural noises of abiotic origin remains to be explored.
Acknowledgments Daniel Opazo provided fundamental directions
for the experimental design. Reinaldo Marfull, Daniel Opazo, Jessica
Toloza and José Serrano helped with the field measurements. Felipe
Moreno-Gómez helped with the experimental design and provided
valuable statistical advice. Two anonymous reviewers contributed
valuable comments on the manuscript. Research supported by FON-
DECYT Grant 1110939. M.I.M. received financial support from The
Guillermo Puelma Foundation for the Neurosciences. This study
implied minimal animal handling. The presence of resident frogs
inside the cavities interfered with the experimental procedures, and
therefore these individuals were gently removed and returned to their
corresponding burrows after the experiments were completed.
References
Amézquita A, Flechas SV, Lima AP, Gasser H, Hödl W (2011) Acous-
tic interference and recognition space within a complex assem-
blage of dendrobatid frogs. Proc Natl Acad Sci 108:17058–
17063. doi:10.1073/pnas.1104773108
Bailey WJ, Roberts JD (1981) The bioacoustics of the burrowing
frog Heleioporus (Leptodactylidae). J Nat Hist 15:693–702.
doi:10.1080/00222938100770491
Bailey WJ, Bennet-Clark HC, Fletcher NH (2001) Acoustics of a
small Australian burrowing cricket: the control of low-frequency
pure-tone songs. J Exp Biol 204:2827–2841
Bates D, Maechler M, Bolker B, Walker S (2004) lme4: linear mixed
effects models using Eigen and S4. R package version 1.1–7.
http://CRAN.R-project.org/package=lme4
Bennet-Clark HC (1987) The tuned singing burrow of mole crickets. J
Exp Biol 128:383–409
Bleach IT, Beckmann C, Both C, Brown GP, Shine R (2015) Noisy
neighbours at the frog pond: effects of invasive cane toads on the
calling behaviour of native Australian frogs. Behav Ecol Socio-
biol 69:675–683. doi:10.1007/s00265-015-1879-z
Blumenrath SH, Dabelsteen T, Pedersen SB (2004) Being inside nest
boxes: does it complicate the receiving conditions for great tit
parus major females? Bioacoustics 14:209–223. doi:10.1080/09
524622.2004.9753526
Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens
MHH, White JSS (2009) Generalized linear mixed models: a
practical guide for ecology and evolution. Trends Ecol Evol
24:127–135. doi:10.1016/j.tree.2008.10.008
Bradbury JW, Vehrencamp SL (2011) Principles of animal communi-
cation. Sinauer Associates, Sunderland
Brumm H, Zollinger SA (2011) The evolution of the Lombard effect:
100 years of psychoacoustic research. Behaviour 148:1173–
1198. doi:10.1163/000579511X605759
Chaverri G, Gillam EH (2013) Sound amplification by means of a
horn-like roosting structure in Spix’s disc-winged bat. Proc R
Soc B 280:20132362. doi:10.1098/rspb.2013.2362
Gough DC, Mennill DJ, Nol E (2014) Singing seaside: pacific wrens
(Troglodytes pacificus) change their songs in the presence of
natural and anthropogenic noise. Wilson J Ornithol 126:269–278

J Comp Physiol A
1 3
Hart PJ, Hall R, Ray W, Beck A, Zook J (2015) Cicadas impact
bird communication in a noisy tropical rainforest. Behav Ecol
26:839–842. doi:10.1093/beheco/arv018
Holt DE, Johnston CE (2014) Evidence of the Lombard effect in
fishes. Behav Ecol 25:819–826. doi:10.1093/beheco/aru028
Hothorn T, Bretz F, Westfall P (2008) Simultaneous inference in
general parametric models. Biom J 50:346–363. doi:10.1002/
bimj.200810425
Kéver L, Colleye O, Lugli M, Lecchini D, Lerouvreur F, Herrel A,
Parmentier E (2014) Sound production in Onuxodon fowleri
(Carapidae) and its amplification by the host shell. J Exp Biol
217:4283–4294. doi:10.1242/jeb.109363
Kinsler LE, Frey AR, Coppens AB, Sanders JV (1982) Fundamentals
of acoustics. Wiley, New York
Lange S, Burda H, Wegner RE, Dammann P, Begall S, Kawalika M
(2007) Living in a “stethoscope”: burrow-acoustics promote
auditory specializations in subterranean rodents. Naturwissen-
schaften 94:134–138. doi:10.1007/s00114-006-0168-0
Lardner B, bin Lakim M (2002) Tree-hole frogs exploit resonance
effects. Nature 420:475. doi:10.1038/420475a
Lin YS, Lehtinen RM, Kam YC (2008) Time- and context-dependent
oviposition site selection of a phytotelm-breeding frog in relation
to habitat characteristics and conspecific cues. Herpetologica
64:413–421. doi:10.1655/08-020R1.1
Lugli M (2012) Acoustics of fish shelters: frequency response
and gain properties. J Acoust Soc Am 132:3512–3524.
doi:10.1121/1.4754581
Lugli M (2013) Sand pile above the nest amplifies the sound emit-
ted by the male sand goby. Environ Biol Fish 96:1003–1012.
doi:10.1007/s10641-012-0097-z
Mennill DJ, Ratcliffe LM (2004) Nest cavity orientation in
black-capped chickadees Poecile atricapillus: do the acous-
tic properties of cavities influence sound reception in the
nest and extra-pair matings? J Avian Biol 35:477–482.
doi:10.1111/j.0908-8857.2004.03351.x
Moreno-Gómez FN, Bacigalupe LD, Silva-Escobar AA, Soto-
Gamboa M (2015) Female and male phonotactic responses
and the potential effect of sexual selection on the advertise-
ment calls of a frog. Anim Behav 104:79–86. doi:10.1016/j.
anbehav.2015.03.010
Page RA, Ryan MJ, Bernal XE (2014) Be loved, be prey, be eaten.
In: Yasukawa K (ed) Animal behavior. Case studies: integration
and application of animal behavior, vol 3. Praeger, New York, pp
123–154
Penna M (2004) Amplification and spectral shifts of vocalizations
inside burrows of the frog Eupsophus calcaratus (Leptodactyli-
dae). J Acoust Soc Am 116:1254–1260. doi:10.1121/1.1768257
Penna M, Hamilton-West C (2007) Susceptibility of evoked vocal
responses to noise exposure in a frog of the temperate austral for-
est. Anim Behav 74:45–56. doi:10.1016/j.anbehav.2006.11.010
Penna M, Marquéz R (2007) Amplification and spectral modification
of incoming vocalizations inside burrows of the frog Eupsophus
roseus (Leptodactylidae). Bioacoustics 16:245–259
Penna M, Meier A (2011) Vocal strategies in confronting inter-
fering sounds by a frog from the southern temperate for-
est, Batrachyla antartandica. Ethology 117:1147–1157.
doi:10.1111/j.1439-0310.2011.01973.x
Penna M, Moreno-Gómez FN (2014) Ample active acoustic space of
a frog from the South American temperate forest. J Comp Phys-
iol A 200:171–181. doi:10.1007/s00359-013-0875-x
Penna M, Moreno-Gómez FN (2015) Contrasting propagation of nat-
ural calls of two anuran species from the South American tem-
perate forest. PLoS One 10(7):e0134498. doi:10.1371/journal.
pone.0134498
Penna M, Quispe M (2007) Independence of evoked vocal
responses from stimulus direction in burrowing frogs
Eupsophus (Leptodactylidae). Ethology 113:313–323.
doi:10.1111/j.1439-0310.2007.01339.x
Penna M, Solís R (1996) Influence of burrow acoustics on sound
reception by frogs Eupsophus (Leptodactylidae). Anim Behav
51:255–263. doi:10.1006/anbe.1996.0026
Penna M, Solís R (1999) Extent and variation of sound enhance-
ment inside burrows of the frog Eupsophus emiliopugini (Lep-
todactylidae). Behav Ecol Sociobiol 47:94–103. doi:10.1007/
s002650050654
Penna M, Velásquez N (2011) Heterospecific vocal interactions in a
frog from the southern temperate forest, Batrachyla taeniata.
Ethology 117:63–71. doi:10.1111/j.1439-0310.2010.01847.x
Penna M, Pottstock H, Velásquez N (2005) Effect of natural and
synthetic noise on evoked vocal responses in a frog of the tem-
perate austral forest. Anim Behav 70:639–651. doi:10.1016/j.
anbehav.2004.11.022
Penna M, Plaza A, Moreno-Gómez FN (2013) Severe constraints for
sound communication in a frog from the South American tem-
perate forest. J Comp Physiol A 199:723–733. doi:10.1007/
s00359-013-0831-9
Phelps SM, Rand A, Ryan MJ (2007) The mixed-species chorus as
public information: tungara frogs eavesdrop on a heterospecific.
Behav Ecol 18:108–114. doi:10.1093/beheco/ar1063
Pike N (2011) Using false discovery rates for multiple compari-
sons in ecology and evolution. Methods Ecol Evol 2:278–282.
doi:10.1111/j.2041-210X.2010.00061.x
Prozesky-Schulze L, Prozesky OPM, Anderson F, var der Merwe GJJ
(1975) Use of a self-made sound baffle by a tree cricket. Nature
255:142–143. doi:10.1038/255142a0
R Core Team (2013) R: a language and environment for statisti-
cal computing. R foundation for statistical computing, Vienna.
http://www.R-project.org/
Römer H (2013) Masking by noise in acoustic insects: problems and
solutions. In: Brumm H (ed) Animal communication and noise.
Springer, Berlin, pp 33–63. doi:10.1007/978-3-642-41494-7_3
Schleich CE, Antenucci DC (2009) Sound transmission and bur-
row characteristics of the subterranean rodent Ctenomys
talarum (Rodentia: Ctenomyidae). Acta Theriol 54:165–170.
doi:10.1007/BF03193172
Schmidt KA, Belinsky KL (2013) Voices in the dark: predation risk
by owls influences dusk singing in a diurnal passerine. Behav
Ecol Sociobiol 67:1837–1843. doi:10.1007/s00265-013-1593-7
Schwartz JJ, Bee MA (2013) Anuran acoustic signal produc-
tion in noisy environments. In: Brumm H (ed) Animal
communication and noise, Springer, Berlin, pp 91–132.
doi:10.1007/978-3-642-41494-7_5
Schwartz JJ, Wells KD (1985) Intra- and interspecific vocal behavior
of the neotropical treefrog Hyla microcephala. Copeia 1:27–38.
doi:10.2307/1444787
Singmann H, Bolker B, Westfall J (2015) Afex: analysis of factorial
experiments. R packages version 0.13–145. http://CRAN.R-pro-
ject.org/package=afex
Stanley CQ, Walter MH, Venkatraman MX, Wilkinson GS (2016)
Insect noise avoidance in the dawn chorus of neotropical birds.
Anim Behav 112:255–265. doi:10.1016/j.anbehay.2015:12.003
Sueur J (2002) Cicada acoustic communication: potential sound par-
titioning in a multispecies community from Mexico (Hemip-
tera: Cicadomorpha: Cicadidae). Biol J Linn Soc 75:379–394.
doi:10.1111/j.1095-8312.2002.tb02079.x
Sueur J, Aubin T, Simonis C (2008) Seewave: a free modular tool for
sound analysis and synthesis. Bioacoustics 18:213–226
Tan WH, Tsai CG, Lin C, Lin YK (2014) Urban canyon effect: storm
drains enhance call characteristics of the Mientien tree frog. J
Zool 294:77–84. doi:10.1111/jzo.12154
Úbeda CA, Nuñez JJ (2006) New parental care behaviours in two tel-
matobiine genera from temperate Patagonian forests: Batrachyla

J Comp Physiol A
1 3
and Eupsophus (Anura: Leptodactylidae). Amphibia–Reptilia
27:441–444. doi:10.1163/156853806778190015
von May R, Medina-Muller M, Donnelly MA, Summers K (2009)
Breeding-site selection by the poison frog Ranitomeya biolat in
Amazonian bamboo forests: an experimental approach. Can J
Zool 87:453–463. doi:10.1139/Z09-026
Wong S, Parada H, Narins PM (2009) Heterospecific acoustic interfer-
ence: effects on calling in the frog Oophaga pumilio in Nicaragua.
Biotropica 41:74–80. doi:10.1111/j.1744-7429.2008.00452.x
Zuk M, Kolluru GR (1998) Exploitation of sexual signals by predators
and parasitoids. Q Rev Biol 73:415–438. doi:10.1086/420412
Zuk M, Rotenberry JT, Tinghitella RM (2006) Silent night: adaptive
disappearance of a sexual signal in a parasitized population of
field crickets. Biol Lett 2:521–524. doi:10.1098/rsbl.2006.0539