Perceived Synchrony of Frog Multimodal Signal Components
Is Influenced by Content and Order
Ryan C. Taylor,
Rachel A. Page,
Barrett A. Klein,
Michael J. Ryan
and Kimberly L. Hunter
*Department of Biological Sciences, Salisbury University, 1101 Camden Avenue, Salisbury, MD 21801, USA;
Smithsonian Tropical Research Institute, Balboa Ancon, 56292 Panama, Republic of Panama;
Department of Biology,
University of Wisconsin—La Crosse, La Crosse, WI 54601, USA;
Department of Integrative Biology, University of Texas
at Austin, Austin, TX 12330, USA
From the symposium “Integrating Cognitive, Motivational and Sensory Biases Underlying Acoustic and Multimodal
Mate Choice” presented at the annual meeting of the Society for Integrative and Comparative Biology, January 4–8, 2017
at New Orleans, Louisiana.
Synopsis Multimodal signaling is common in communication systems. Depending on the species, individual signal
components may be produced synchronously as a result of physiological constraint (fixed) or each component may be
produced independently (fluid) in time. For animals that rely on fixed signals, a basic prediction is that asynchrony
between the components should degrade the perception of signal salience, reducing receiver response. Male t
frogs, Physalaemus pustulosus, produce a fixed multisensory courtship signal by vocalizing with two call components
(whines and chucks) and inflating a vocal sac (visual component). Using a robotic frog, we tested female responses to
variation in the temporal arrangement between acoustic and visual components. When the visual component lagged a
complex call (whine þchuck), females largely rejected this asynchronous multisensory signal in favor of the complex
call absent the visual cue. When the chuck component was removed from one call, but the robofrog inflation lagged
the complex call, females responded strongly to the asynchronous multimodal signal. When the chuck component was
removed from both calls, females reversed preference and responded positively to the asynchronous multisensory
signal. When the visual component preceded the call, females responded as often to the multimodal signal as to the
call alone. These data show that asynchrony of a normally fixed signal does reduce receiver responsiveness. The
magnitude and overall response, however, depend on specific temporal interactions between the acoustic and visual
components. The sensitivity of t
ungara frogs to lagging visual cues, but not leading ones, and the influence of acoustic
signal content on the perception of visual asynchrony is similar to those reported in human psychophysics literature.
Virtually all acoustically communicating animals must conduct auditory scene analyses and identify the source of
signals. Our data suggest that some basic audiovisual neural integration processes may be at work in the vertebrate
Animal signals are complex, often consisting of in-
dividual components transmitted and received
through multiple sensory channels (Hebets and
Papaj 2005;Higham and Hebets 2013;Hebets et al.
2016). Signal complexity has been an area of intense
research for more than 15 years (Partan and Marler
1999), yet we understand little about how a signal
component in one sensory channel influences the
perception and corresponding behavioral response
to a component in another channel. In animal court-
ship signals, for example, do individual components
in the auditory and visual channels combine to in-
crease female responses in an additive fashion?
Alternatively, does the addition of a visual compo-
nent induce an exponentially stronger response in
receivers or even reduce their response relative to
the acoustic signal alone? Some recent work in ani-
mal communication indicates that the perception
and subsequent behavioral response to multisensory
Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology 2017.
This work is written by US Government employees and is in the public domain in the US.
Integrative and Comparative Biology
Integrative and Comparative Biology, pp. 1–8
doi:10.1093/icb/icx027 Society for Integrative and Comparative Biology
signals is not additive or easily predicted (Taylor and
Ryan 2013;Rubi and Stephens 2016;Stange et al.
2016). To date, the most comprehensive work on
audiovisual integration and non-additive effects has
been done in cats and primates, including work in
human psychophysics (for review see Stein 2012).
The human psychophysical work has been critical
for informing us about how the senses are integrated
and how this integration allows individuals to make
sense of a complex world around them. In particular
the recruitment of additional senses, such as vision,
is one mechanism that humans use to locate and
recognize acoustic signals, increasing the efficacy of
our auditory scene analyses (Sumby and Pollack
1954). Psychophysical techniques have been applied
to a number of taxa, but frogs are especially amenable
to these methods, allowing us to address questions
about the perception of complex signals (Bee and
Micheyl 2008;Bee 2015). Male frogs produce stereo-
typed advertisement (mating) calls and their neural
systems are “tuned” to properties of these calls
(Ryan 2001). In most species, females search out call-
ing males and approach them to initiate mating. If
the mating signals deviate too far from their species-
specific properties, female receivers fail to perceive
them as appropriate mating signals (Phelps et al.
2005). Because females readily respond to acoustic
playbacks of male signals, and engage in mate search-
ing behavior, they are easy to manipulate in behav-
ioral tests of signal perception. These perceptions are
directly relevant to understanding how communica-
tion signals evolve. In frogs, the sex ratio is typically
highly skewed and male reproductive success is like-
wise skewed. Therefore, female mate choice generates
strong selection on male signal evolution.
ungara frog, Physalaemus pustulosus,isa
small frog found from northern South America
through southern Mexico. Like many frog species,
they breed in ephemeral pools of water and males
produce a conspicuous acoustic signal, the advertise-
ment call. In t
ungara frogs, this advertisement call
consists of two components. The first is the “whine”
and the second is the “chuck.” The whine is necessary
and sufficient for mate attraction and males always
produce this component. The chuck is neither neces-
sary nor sufficient for mate attraction but males can
facultatively append up to seven chucks onto the end
of the whine (usually one to three). Chucks make the
whine more attractive to females, and always follow
the whine as a result of morphological constraint
(Ryan and Guerra 2014). The advertisement call is
also accompanied by the synchronous inflation of a
conspicuous vocal sac that has been shown to make
the call more attractive (Taylor et al. 2008).
Thus, females assess both the call and the vocal sac
inflation as part of a multimodal signal.
The visual cue of an inflating vocal sac increases
the attractiveness of a call when it is added, but its
effect can easily be overridden by an alternative call
that contains more attractive properties. Thus, the
acoustic signal component has primacy for female
mate choice. The temporal arrangement of the call
and vocal sac movement are critically important,
however. If the vocal sac inflation is delayed, such
that it lags the call in time, females strongly reject
this asynchronous multisensory signal (Taylor et al.
2011). Alternatively, temporally sandwiching the vo-
cal sac movement between the whine and chuck can
restore the saliency of the overall mating signal
(Taylor and Ryan 2013). For an individual male,
temporal delays between the call and vocal sac move-
ment are impossible due to morphological con-
straints. Our previous experimental data show that
females strongly attend to temporal synchrony of the
signal components, yet are flexible about how they
perceive and respond to temporal variation.
Our current understanding of the t
ungara frog sys-
tem suggests that two simple rules may govern female
choice for multisensory signals. First, if the vocal sac
inflates following a call, then reject the signal. Second,
if the whine and chuck “book end” the vocal sac, then
accept the signal. Despite these data, we still have a
largely incomplete understanding of how all three
components—whine, chuck, and vocal sac—interact
to influence perception and female mate choice.
In this study we further probed how females re-
spond to asynchronous signals. Specifically, we asked
two questions. First, we asked if acoustic content
matters. Do females find an asynchronous multi-
modal signal aversive, when one or more of the calls
lack a chuck? This question is important because it
helps to shed additional light on the cognitive/per-
ceptual system that governs how the frog audiovisual
system processes complex signals. Second, we asked
if there is a syntactical order effect. That is, does a
vocal sac that leads a call in time influence female
choice as it would if it lags the call? This question is
intriguing because for males, order of call compo-
nents is fixed; vocal sac inflations always coincide
with the call and chucks always follow whines.
Females, however, show permissiveness for temporal
arrangement of chuck placement in tests with no
visual cue (Wilczynski et al. 1999).
We collected mated pairs of t
ungara frogs at cho-
ruses within 4 h after sunset. The frogs were collected
2R. C. Taylor et al.
at breeding sites near the Smithsonian Tropical
Research Institute, Gamboa, Republic of Panama.
We placed individual frog pairs into plastic bags
and stored the frogs in a light-safe cooler (total dark-
ness) for a minimum of 1 h prior to testing. This
ensured that the frogs’ eyes were dark-adapted after
collection using flashlights. After testing, the frogs
were toe-clipped, following guidelines of the
American Association of Ichthyologists and
Herpetologists, which allowed us to avoid using re-
captures on subsequent nights. We released all col-
lected frogs at their sites of capture at the end of the
night, ensuring that they could breed in the wild. All
procedures were approved by STRI IACUC (2011-
0825-2014-02) and conducted with permits from
Panama’s ANAM permit No. SE/A-30-12. ANAM
is now the Ministry of the Environment,
We conducted phonotaxis experiments in a hemi-
anechoic chamber (Acoustic Systems, ETS-Lindgren,
Austin, TX, USA) measuring 2.7 m 1.8 m 2m.
For the behavioral tests, we used a restraining funnel
placed in the center of the chamber. The funnel kept
the females equidistant (80 cm) from the two speak-
ers (Mirage Nanosat, Klipsch Audio, Indianapolis,
IN, USA) used to broadcast the male calls (Fig.
1a). Each speaker was separated by 80 cm and
formed a triangle with ca. 60separation relative
to the female’s release point. To generate a multisen-
sory signal, we placed a robotic frog (robofrog) with
an inflatable vocal sac in front of one speaker. We
inflated the vocal sac of the robotic frog remotely
using a pneumatic pump that was triggered by the
computer producing the acoustic stimulus. By using
a sound file to trigger the robofrog vocal sac infla-
tion, we were able to precisely control the timing of
the robofrog’s inflation/deflation sequence relative to
the calls produced at the speaker. Because the
speaker broadcast the call from the same location
as the robofrog, this closely matched the spatial lo-
cation of the natural visual and acoustic signal com-
ponents (Taylor et al. 2008;Klein et al. 2012).
We illuminated the test chamber with a single GE
nightlight (ca. 2.27 10
, model no.
55507; Fairfield, CT, USA). The spectrum and inten-
sity of light at nocturnal breeding sites varies tre-
mendously with location (forest cover vs. open),
moon phase, and cloud cover. The light environment
we provided was well within the range of what frogs
naturally experience (Cummings et al. 2008;Taylor
et al. 2008). For each trial, we placed a female under
the funnel and broadcast digitally synthesized male
vocalizations (see Ryan et al.  for details on
call synthesis). The robofrog vocal sac was also
activated to inflate/deflate asynchronously with the
call broadcast at the speaker (for more details see
Taylor et al. 2011). These playbacks were broadcast
for 2 min, which allowed the female to acclimatize to
the playbacks while under the funnel. For all exper-
iments, we used a synthetic, simple (whine), or com-
plex (whine þone chuck) call broadcast at 82 dB SPL
(re. 20 mPa; RMS, fast, C weighting) measured at
the point of release for the females. We used Adobe
Audition software (ver. 3.0) for playbacks and each
call was played once every 3 s.
After the acclimation period, we lifted the funnel
so the female was free to move around the test arena.
We recorded a choice when a female approached to
within 5 cm of a speaker or speaker/robofrog com-
bination and remained there for 5 s. The 5 s rule
avoided false positives or negatives caused by females
simply walking by a stimulus. To control for side
bias, we systematically alternated the sides on which
the robofrog and calls were presented between trials.
If a female did not move for 2 min after the funnel
was raised or failed to enter a choice zone within
10 min, we discarded the trial from the data set
due to a lack of motivation. Response rates by fe-
males were typically around 65% each night. We
recorded female behavior using an infrared sensitive
camera (Everfocus EHD500IR, Everfocus Electronics,
Duarte, CA, USA) mounted on the ceiling of the
chamber. A video feed allowed us to view the fe-
male’s behavior in real time from outside the sound
chamber, while simultaneously recording video
Following these general procedures, we conducted
three experiments. In Experiment 1, we presented
females with a complex call (whine plus one chuck,
hereafter “WC”) versus a simple call (the whine
alone, hereafter “W”). The WC had the visual com-
ponent of a robofrog with inflating vocal sac added,
but the vocal sac inflation lagged the call. The call
and robofrog inflation were 100% out of phase such
that the inflation began immediately following the
terminus of the call (Fig. 1b). The temporal sequence
of this stimulus was: whine, then chuck, then vocal
sac inflation, hereafter abbreviated as (WC-robo). In
Experiment 2, we presented females with the identi-
cal W call at each speaker. To one speaker we also
added a robofrog with inflation following the whine,
hereafter abbreviated as (W-robo). Here also, the
inflation occurred 100% out of phase, immediately
following the call (Fig. 1b). In the third experiment,
we presented females with two identical WC calls,
but one speaker again had a robofrog added. The
robofrog vocal sac inflation preceded the call yielding
a temporal sequence of: vocal sac inflation, then
Synchrony of frog multimodal signals 3
whine, then chuck, hereafter abbreviated as (robo-
WC). Although the inflation preceded the call, the
inflation still occurred 100% out of phase; the call
began immediately following the deflation of the
robofrog vocal sac (Fig. 1b).
All experiments consisted of a two-choice test, where
females had the option of responding to a unimodal
call (speaker only) or a multimodal signal (speaker
plus the visual cue of a robofrog). The data were
analyzed using a binomial exact test and the mid-P
value (Agresti 2001). We previously showed that
when the robofrog’s vocal sac inflation temporally
lagged the complex WC call by either 50% or
100%, females chose the multisensory signal only
25% of the time (Taylor et al. 2011). The timing
of the lagging vocal sac in the current study matched
the timing of the 100% from previous experiments.
These experiments were later repeated (unpublished
data), confirming the results. Given the repeatable
and robust nature of the female preference function
for a lagging visual component, we set our a priori
expected binomial response to this asynchronous
multisensory signal at 0.25 (Experiments 1 and 2).
In Experiment 3, where the vocal sac inflation led
the call, we had no prior data to suggest how females
would respond to this particular temporal arrange-
ment. Therefore, we set our a priori expected re-
sponse rate at random choice ¼0.5.
In our first experiment, we presented females with a
WC versus W call, but the robofrog was added to
the speaker playing the WC and the vocal sac was
inflated asynchronously following the call (WC-
robo). Females chose the WC-robo in 75% of trials
(n¼24; binomial test, expected ¼0.25; P<0.0001;
Fig. 2). This reversed the general avoidance of the
asynchronous multisensory signal when the calls
broadcast from alternative speakers were held con-
stant (both WC). This distribution is similar to fe-
male behavior in a standard WC versus W
experiment when no robofrog is present (Gridi-
Papp et al. 2006). Thus, the presence of the chuck
at one call was enough to overcome the unattractive-
ness of the asynchronous signal when the alternative
call was just the whine.
In the second experiment, we presented females
with two identical calls consisting of the whine
only. The speaker with the robofrog lagged the call
(W-robo). Here, females also did not exhibit an
overall aversion to the asynchronous multisensory
Fig. 1 (A) Diagram of two-choice test arena. Females could
choose between two stimuli, a call only or a call with the asyn-
chronously inflating robofrog placed in front of the speaker. (B)
Detail of female choice tests. The asynchronous multimodal signals
are depicted on the left side; the calls only are depicted by the
sonograms on the right. In Experiment 1, the robofrog vocal sac
inflation lagged the call (depicted in the timeframe above the
whine–chuck sonogram). The alternative was a whine only. In
Experiment 2, the robofrog vocal sac inflation lagged the call
(depicted in the timeframe above the whine only sonogram). The
alternative was also a whine only. In Experiment 3, the robofrog
vocal sac inflation led the call (depicted in the timeframe above the
whine–chuck sonogram). The alternative was also a whine chuck.
4R. C. Taylor et al.
signal. They chose it 60% of the time, significantly
more often than expected (n¼40; binomial test, ex-
pected ¼0.25; P<0.0001; Fig. 2).
In the final experiment, we presented females
again with two identical calls consisting of a WC.
This time, the speaker with the robofrog inflated
before the call (robo-WC). Females chose the asyn-
chronous signal 40% of the time (n¼40; binomial
test, expected ¼0.5; P¼0.21; Fig. 2). Thus, females
did not choose either the visually leading asynchro-
nous multimodal signal or the unimodal call more
often that expected from random chance.
All else being equal, the presence of a synchronously
inflating vocal sac makes a male’s call more attractive
to females (Taylor et al. 2008). Further, females tend
to reject an asynchronous signal when the vocal sac
inflation lags the call (Taylor et al. 2011). Male
ungara frogs often call in dense choruses and due
to physiological constraints cannot alter the timing
of vocal sac inflation and call production. Taylor
et al. (2011) suggested that the assessment of the
vocal sac by females may provide a means of iden-
tifying individual callers with a chorus, much like a
human reads lips at noisy parties (Sumby and
In this study, we show that the acoustic and visual
signal components of the t
ungara frog’s mating sig-
nal interact in complex ways to influence female
choice. Since males cannot alter the timing of their
audiovisual signals in nature, it seems intuitive that
females would recognize any incongruency and
adopt a simple rule that rejects any combination
that does not match the natural template.
Interestingly, there does not appear to be a set “rule”
that governs a simple template recognition of signal
synchrony by females (Taylor and Ryan 2013).
In our first experiment, where we played an asyn-
chronous multimodal WC versus a unimodal W, fe-
males showed virtually no aversion to the
asynchronous signal and responded to the WC al-
most as strongly as the same experiment, absent the
visual component (85% preference Gridi-Papp et al.
2006; 75% this study). This indicates that although
the asynchronous audio-visual signal is generally
aversive, if one call contains a chuck, the asynchrony
is still more attractive than an isolated whine.
In nature, chucks always follow whines.
Wilczynski et al. (1999), however, showed that fe-
ungara frogs are permissive to the temporal
order of whines and chucks. In particular, they
found that in stimuli where a chuck artificially
preceded a whine, females found this as attractive
as one that followed in the natural position. Given
the difficult task that females have assigning calls to
their source when many males are calling within a
small area, one prediction might be that females use
the chucks to determine when a call is finished. This
should improve a female’s ability to assign calls to
their source. The data from Wilczynski et al. (1999)
suggest that this is not true, at least when a female is
presented with only two, spatially separated calling
males. So for the acoustic component of the signal,
syntax for female receivers is flexible. Farris and
Ryan (2011,2017) also demonstrated that female
ungara frogs make relative comparisons when iden-
tifying callers acoustically. In a series of experiments,
they showed that females perceptually group whines
and chucks that are temporally and spatially sepa-
rated, effectively responding as if the disparate com-
ponents belong to the same source. Here again, the
females show permissiveness for signal variation in
time and space. They showed that females more
readily group calls that have a smaller spatial sepa-
ration and non-natural sequence relative to calls with
a greater spatial separation but natural sequence
(Farris and Ryan 2017). Although females perceptu-
ally weight spatial cues more, when multiple cues
become available, females integrate these into their
perceptual and decision making processes (Farris and
When the visual component is added to the signal,
syntax becomes more important. In our second ex-
periment where we removed the chucks altogether
and just presented females with whines in the acous-
tic domain, the asynchronous multisensory signal
(W-robo) was no longer aversive, and females chose
this signal more often than expected. In the absence
of the chuck, females are less likely to be influenced
by the incongruency. This may indicate that when
females are simultaneously evaluating acoustic and
visual components, the chuck indicates that the call
is finished, and any vocal sac inflation following this
call does not belong. Thus like relative comparisons
within the auditory domain (Farris and Ryan 2011,
2017), female t
ungara frogs also appear to make rel-
ative comparisons when integrating visual and
acoustic cues (for other cross-modal comparisons,
see also Halfwerk et al. 2014).
In our final experiment, we presented females with
a pair of identical WCs, but at one speaker, the
robofrog inflation preceded the call (robo-WC).
Females responded to the asynchronous signal statis-
tically as often as the unimodal call only. This sug-
gests that females do not recognize the temporal
asynchrony or that their perception of the leading
Synchrony of frog multimodal signals 5
visual signal is less aversive than when it lags the call.
Interestingly, this behavior coincides with audiovi-
sual discrepancy detection in human listeners.
Human listeners, like many vertebrates, integrate au-
ditory and visual signals and generate perceptions of
synchrony as part of their overall auditory scene anal-
ysis (Stein 2012;Farris and Ryan 2017). Humans
more easily detect asynchrony when a visual cue lags
an acoustic signal versus one that leads (Dixon and
Spitz 1980). Given that light travels dramatically faster
than sound, audiovisual discrepancies occur in nature
with increasing communication distances. Specifically,
since sound naturally lags a visual stimulus, it might
be expected that receivers, humans or otherwise, are
somewhat permissive of lagging sound. For example,
Navarra et al. (2009) showed that human listeners
increased reaction times to audio signals that lagged
the visual cue, but were unable to do this for lagging
visual signals. They suggested this effect may result
from auditory processing plasticity that can compen-
sate for the normal temporal lag that occurs in nature,
thereby improving the ability of the brain to bind
relevant audiovisual cues into a coherent stimulus
(also see Sugita and Suzuki 2003). Given stimulus
transmission and neural transduction speeds, commu-
nication distances need to exceed about 10 m before
audio signals begin to perceptually lag visual signals
(Po¨ppel and Artin 1988) and human listeners remain
unaware of asynchronies until the audio stimulus lags
the visual by about 250 ms (Dixon and Spitz 1980).
Our results have important implications for our
understanding of sensory ecology, perception, and
multimodal signal evolution. First, for nocturnally
communicating frogs that use multimodal signals,
the evaluation distance is nearly always less than 10
m (personal observation). Thus, a female receiver is
unlikely to experience a noticeable audiovisual asyn-
chrony produced by a particular calling male. In
light of this, there is no ecological reason why female
frogs should be more sensitive to a lagging visual
signal versus a lagging audio signal. Our data show
that they are, however. One explanation may be that
neural integration of auditory and visual signals, par-
ticularly the perception of synchrony, is a conserved
process across many vertebrate taxa. In particular, if
the vertebrate auditory processing is more plastic
than the visual system (Navarra et al. 2009), then
this may constrain receivers to be permissive of lag-
ging audio signals, irrespective of whether they ex-
perience them in nature.
The second implication of our results is that con-
textual aspects of audiovisual integration may be as
important as temporal structure per se. For t
frogs, the chuck component of the call must be ac-
companied by the whine in order for females to even
recognize it as a salient signal (Ryan 1985). Even so,
once the context is set (e.g., the presence of the
whine), the chuck strongly modulates female
attraction, making the complex call five times more
attractive as the whine only (Gridi-Papp et al. 2006).
Fig. 2 Proportion of females choosing an asynchronous multimodal signal (audioþvisual) versus an alternative unimodal signal (call
only). The far left experiment separated by a vertical line is from Taylor et al. (2011) and was used to set prior expectation of
asynchrony response at 0.25 (horizontal line). For Experiment 3 on the far right, the expected response was set at 0.5 (horizontal line).
The x-axis legends refer to the temporal sequence of the stimuli. WC-robo ¼whine, then chuck, then robotic frog inflation.
W-robo ¼whine, then robotic frog inflation. Robo-WC ¼robotic frog inflation, then whine, then chuck. The graphic of the timing
of the robofrog inflation/sonogram follows from Fig. 1b.
6R. C. Taylor et al.
The presence of the chuck also overrides the aversive
nature of the lagging visual signal. Likewise, when the
chucks are removed completely, females are no longer
averse to the temporal asynchrony. In sum, females
are permissive to variation in call syntax when pre-
sented with a call only (e.g., chuck precedes whine)
and they are permissive of multisensory asynchrony
when chucks are absent. The presence of the chuck,
however, alerts females to the asynchrony of the mul-
tisensory signal (when the visual cue lags a standard
complex call), and modulates their behavior.
We suggest that future studies of multimodal sig-
naling should include experiments that are not only
signal isolation tests (sensu Partan and Marler 2005),
but also explore how different arrangements of both
context and timing influence receiver behavior.
Doing so is likely to reveal the full range of multi-
sensory space over which receivers recognize and re-
spond to conspecific signals (Smith and Evans 2013),
including variations that don’t naturally occur. This
will provide insights into how neural integration and
sensory perception can promote or constrain the
evolution of complex signal design.
Joey Stein and Moey Inc. developed the robotic frog
control system. We thank Nic Stange, Kyle Wilhite,
and Kelsey Mitchell for help with data collection. We
are grateful to the Smithsonian Tropical Research
Institute for logistical support. Constructive criticism
from one anonymous reviewer improved the quality
of the manuscript. The work was conducted under
STRI IACUC protocol No. 2011-0825-2014-02 and
collecting permit from Panama’s Autoridad
Nacional del Ambiente (ANAM).
This work was supported by a National Science
Foundation grant [IOS 1120031 to R.C.T., M.J.R.,
Agresti A. 2001. Exact inference for categorical data: recent ad-
vances and continuing controversies. Stat Med 20:2709–22.
Bee MA. 2015. Treefrogs as animal models for research on
auditory scene analysis and the cocktail party problem. Int
J Psychophysiol 95:216–37.
Bee MA, Micheyl C. 2008. The cocktail party problem: what
is it? How can it be solved? And why should animal be-
haviorists study it? J Comp Psychol 122:235–251.
Cummings ME, Bernal XE, Reynaga R, Rand AS, Ryan MJ.
2008. Visual sensitivity to a conspicuous male cue varies by
reproductive state in Physalaemus pustulosus females. J Exp
Dixon NF, Spitz L. 1980. The detection of auditory visual
desynchrony. Perception 9:719–21.
Farris HE, Ryan MJ. 2011. Relative comparisons of call param-
eters enable auditory grouping in frogs. Nat Commun 2:410.
Farris HE, Ryan MJ. 2017. Schema vs. primitive perceptual
grouping: the relative weighting of sequential vs. spatial
cues during an auditory grouping task in frogs. J Comp
Physiol A 203:175–82.
Gridi-Papp M, Rand AS, Ryan MJ. 2006. Animal communi-
cation: complex call production in the t
ungara frog. Nature
Halfwerk W, Page RA, Taylor RC, Wilson PS, Ryan MJ. 2014.
Crossmodal comparisons of signal components allow for
relative-distance assessment. Curr Biol 24:1751–5.
Hebets EA, Barron AB, Balakrishnan CN, Hauber ME, Mason
PH, Hoke KL. 2016. A systems approach to animal com-
munication. Proc R Soc B Biol Sci 283:20152889.
Hebets EA, Papaj DR. 2005. Complex signal function: devel-
oping a framework of testable hypotheses. Behav Ecol
Higham JP, Hebets EA. 2013. An introduction to multimodal
communication. Behav Ecol Sociobiol 67:1381–8.
Klein BA, Stein J, Taylor RC. 2012. Robots in the service of
animal behavior. Commun Integr Biol 5:466–72.
Navarra J, Hartcher-O’Brien J, Piazza E, Spence C. 2009.
Adaptation to audiovisual asynchrony modulates the speeded
detection of sound. Proc Natl Acad Sci U S A 106:9169–73.
Partan S, Marler P. 1999. Behavior–communication goes mul-
timodal. Science 283:1272–3.
Partan SR, Marler P. 2005. Issues in the classification of mul-
timodal communication signals. Am Nat 166:231–45.
Phelps SM, Rand AS, Ryan MJ. 2005. A cognitive framework
for mate choice and species recognition. Am Nat 167:28–42.
Po¨ ppel E, Artin TT. 1988. Mindworks: time and conscious
experience. Harcourt Brace Jovanovich.
Rubi TL, Stephens DW. 2016. Why complex signals matter,
sometimes. In: Bee MA, Miller C, editors. Psychological
mechanisms in animal communication. New York (NY):
Springer. p. 119–35.
Ryan MJ. 1985. The t
ungara frog: a study in sexual selection
and communication. Chicago: University of Chicago Press.
Ryan MJ. 2001. Anuran communication. Washington, DC:
Smithsonian Institution Press.
Ryan MJ, Guerra MA. 2014. The mechanism of sound pro-
duction in tungara frogs and its role in sexual selection and
speciation. Curr Opin Neurobiol 28:54–59.
Ryan MJ, Rand W, Hurd PL, Phelps SM, Rand AS. 2003.
Generalization in response to mate recognition signals.
Am Nat 161:380–94.
Smith CL, Evans CS. 2013. A new heuristic for capturing the
complexity of multimodal signals. Behav Ecol Sociobiol
Stange N, Page RA, Ryan MJ, Taylor RC. 2016. Interactions
between complex multisensory signal components result in
unexpected mate choice responses. Anim Behav 116:83–7.
Stein BE. 2012. The new handbook of multisensory process-
ing. Cambridge (MA): MIT Press.
Sugita Y, Suzuki Y. 2003. Audiovisual perception: implicit
estimation of sound-arrival time. Nature 421:911.
Sumby WH, Pollack I. 1954. Visual contribution to speech
intelligibility in noise. J Acoust Soc Am 26:212–5.
Synchrony of frog multimodal signals 7
Taylor RC, Klein BA, Stein J, Ryan MJ. 2008. Faux frogs:
multimodal signalling and the value of robotics in animal
behaviour. Anim Behav 76:1089–97.
Taylor RC, Klein BA, Stein J, Ryan MJ. 2011. Multimodal
signal variation in space and time: how important is
matching a signal with its signaler? J Exp Biol 214:815–20.
Taylor RC, Ryan MJ. 2013. Interactions of multisensory com-
ponents perceptually rescue t
ungara frog mating signals.
Wilczynski W, Rand AS, Ryan MJ. 1999. Female preferences
for temporal order of call components in the tungara frog:
a Bayesian analysis. Anim Behav 58:841–51.
8R. C. Taylor et al.