Social context evokes rapid changes in bat song syntax
Kirsten M. Bohn
, Grace C. Smarsh, Michael Smotherman
Department of Biology, Texas A&M University, College Station, TX, U.S.A.
Received 11 December 2012
Initial acceptance 7 January 2013
Final acceptance 29 March 2013
Available online 3 May 2013
MS. number: A12-00940R
The capacity to modify vocal syntax to changes in social context is an important component of vocal
plasticity and complexity in adult vertebrates, especially in human speech. The ecological signiﬁcance of
this behaviour has been well established in some avian species but not in mammals where complex,
multisyllabic vocalizations are rare. The Brazilian free-tailed bat, Tadarida brasiliensis, is a mammal that
sings like a bird, producing hierarchically structured songs that vary in the order and number of phrases
(i.e. syntax) from one rendition to the next while simultaneously following speciﬁc organizational rules.
Here, we used playback experiments to examine the function of songs and tested whether song syntax is
correlated with social context. Free-tailed bats responded rapidly and robustly to echolocation calls that
mimicked a bat ﬂying past the roost but did not respond to conspeciﬁc song playbacks. We compared
songs that were directed at a passing bat with songs that were produced spontaneously and found that
bats produced longer songs with different phrase content and order when singing spontaneously than
when singing to bats approaching their roost. Thus, free-tailed bats quickly varied song composition to
meet the speciﬁc demands of different social functions. These distinct singing behaviours are similar to
those of some songbirds, suggesting that bats and birds have converged upon a similar set of production
modes that may reﬂect common neural mechanisms and ecological factors.
Ó2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
An important milestone in the evolution of animal communi-
cation is the transition from monosyllabic to polysyllabic vocali-
zations (Maynard Smith & Szathmáry 1995). While the information
in monosyllabic vocalizations is limited to changes in the acoustic
features of a syllable (phonology), multiple syllables add an entirely
new dimension of potential ﬂexibility and information, namely
syntax, or the way in which elements are ordered and combined.
The use of learned, multisyllabic vocalizations with ﬂexible syntax
is most widely seen in birds (Kroodsma & Miller 1996). Songbirds
are especially well known for the widespread use of multisyllabic
songs associated with mating and territorial defence (Marler &
Slabbekoorn 2004;Catchpole & Slater 2008). Syntax in birdsong
is salient (Balaban 1988), and social context may play a large role in
song structure and note use (Catchpole & Slater 2008;Byers &
In contrast, most mammals produce monosyllabic, ﬁxed signals
with much less ﬂexibility than birds (Hammerschmidt & Fischer
2008;Snowdon 2009; but see Arnold & Zuberbühler 2006;
Clarke et al. 2006;Ouattara et al. 2009). Although various examples
of singing have been documented in mammals (Payne & McVay
1971;Mitani & Marler 1989;Davidson & Wilkinson 2002;Behr &
von Helversen 2004;Holy & Guo 2005;Clarke et al. 2006;Bohn
et al. 2009), there is no evidence that mammals alter song
composition and structure in ways comparable to songbirds. This
key behavioural distinction has been attributed to the general
absence of a neural substrate supporting vocal plasticity in mam-
mals that is present in both songbirds and humans (Doupe & Kuhl
1999;Jarvis et al. 2005;Kao et al. 2005;Jürgens 2009).
Cetaceans and bats may be exceptional since both groups have
evolved a suite of neural adaptations to support laryngeal echolo-
cation. They are the only two groups of mammals that demonstrate
vocal learning (Boughman 1998;Janik 2000;Foote et al. 2006;
Knornschild et al. 2010), juvenile babbling (Knörnschild et al.
2006), regional dialects (Cerchio et al. 2001;Ouattara et al. 2009)
and cultural transmission of vocalizations (Deecke et al. 2000;
Garland et al. 2011). If these behaviours are indicative of a neural
substrate that supports vocal plasticity, it is likely that some ceta-
ceans and bats may also possess the capacity to rapidly vary vocal
syntax in response to social cues.
Brazilian free-tailed bats, Tadarida brasiliensis, produce songs that
are remarkably similar acoustically and behaviourally to birdsongs.
Free-tailed bat songs follow a hierarchical structure where three
types of phrases (chirps, trills and buzzes) are in turn composed of
four types of syllables (Chirp A, Chirp B, trill and buzz; Bohn et al.
2009;Fig. 1). Free-tailed bat songs are highly ﬂexible while
following a clear and consistent syntax. The number and order of
phrases dynamically vary from one rendition to the next, while
*Correspondence and present address: K. M. Bohn, School of Integrated Science
and Humanity, Florida International University, Miami, FL 33199, U.S.A.
E-mail address: kbohn@ﬁu.edu (K. M. Bohn).
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/locate/anbehav
0003-3472/$38.00 Ó2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 85 (2013) 1485e1491
simultaneously adhering to a speciﬁc set of rules (Bohn et al. 2009).
Although some “song types”are preferred over others (Fig. 1), the
number of repeats of trill and buzz phrases vary within each song
type. Thus, a few element types (in this case phrases) can be com-
bined to form a potentially enormous set of unique signals. This
‘combinatorial syntax’(Hailman & Ficken 1986) has been observed in
only a handful of avian species including chickadees (Hailman &
Ficken 1986;Hailman et al. 1987;Ficken et al. 1994) parrots
(Wright & Dahlin 2007) and hummingbirds (Rusch et al. 1996).
As in songbirds, free-tailed bat songs are produced exclusively
by males and singing is especially pronounced during the mating
season. During this time, males establish territories that they
aggressively defend against other males, but in which they allow
females to reside (Bohn et al. 2008). However, males also sing year
round and in all-male colonies, so songs may function in other
social contexts besides mating. This is remarkably similar to
chickadee calls, which have combinatorial syntax that varies with
social (Ficken et al. 1994) and environmental (Soard & Ritchison
2009) context. Free-tailed bats have a large vocal repertoire
(Bohn et al. 2008), roost in the millions and have all-male, all-fe-
male or mixed-sex colonies that ﬂuctuate across seasons. Thus, the
broad diversity of song types may reﬂect behavioural responses to a
highly variable social environment.
Due to the cryptic nature of bats, as well as the short duration,
fast tempo, and ultrasonic frequencies of their songs, it has been
difﬁcult to discern speciﬁc stimuli or behavioural contexts that
evoke singing. Since visual cues are generally unavailable to detect
conspeciﬁcs from within roost sites, we hypothesized that bats use
auditory stimuli to assess social context and that songs may be
primarily used to advertise males’hidden locations to ﬂying con-
speciﬁcs. In this study we used playbacks of echolocation passes
and conspeciﬁc songs to determine whether they evoke singing
and to test whether bats vary their songs with social context.
We conducted 10 playback experiments, each at a different
location at a natural roosting site located on the Texas A&M Uni-
versity campus (College Station, TX, U.S.A.). Atthis site, bats roost in
discrete locations in cracks that run between the concrete slabs of
the football stadium. We selected 10 roost sites that were outside of
the hearing range of each other and placed two speakers and a
microphone within 0.25 m of the crack. Each playback experiment
consisted of two types of playbacks, echolocation and song. In
addition, we conducted ﬁve echolocation playbacks within a vi-
varium using captive T. brasiliensis, where one male singer was
present at a time (N¼5 males). The Texas A&M Animal Care and
Use Committee approved all procedures and animal husbandry for
this research (protocol number number 2007-254). Stimuli were
played from a laptop computer running Avisoft Recorder
a PCMCIA card (NI DAQCard-6062E) to a Pioneer ribbon tweeter
(ART-55D), via a Sony power ampliﬁer (STR-DE598).
Figure 1. Spectrograms of songs produced by male free-tailed bats. (a) A bat produces a chirp-trill-buzz song type in response to an echolocation playback (Supplementary Audio
S1). (b) The same bat produces a chirp-trill-chirp song type spontaneously in the roost (Supplementary Audio S2). (c) A different bat produces a chirp-buzz song type
(Supplementary Audio S3). Upper case letters refer to the ﬁrst four Chirp A and Chirp B syllables. Note that for the song in (b), the complete phrase sequence is chirp-trill-trill-chirp,
but when deﬁning ‘song types’, sequential repetitions of phrases are omitted.
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491148 6
To create echolocation stimuli, echolocation passes were
recorded from ﬁve males, one at a time, that were free ﬂying in a
623m ﬂight-room lined with acoustic foam. The micro-
phone (Bruel & Kjær, type 4939, ¼ inch) was placed in the centre
of the room where echolocation passes were recorded as the bats
ﬂew over the microphone, resulting in an amplitude pattern that
mimicked a wild bat’s approach during playbacks (Fig. 2).
Although the echolocation passes recorded from captive bats
differed somewhat from those of wild bats, they were produced in
similar highly cluttered environments (at the wild colony, ca. 2 m
of space in front of each roosting site is surrounded by cement).
Echolocation passes were digitized with a National Instruments
DAQmx card, NI PCI-6251 (300 kHz, 16-bit sample rate) and Avi-
soft software (v.2.97, Avisoft Bioacoustics, Berlin). Song stimuli
were obtained from previous recordings we had performed in
captive colonies where songs were produced spontaneously in the
roosts (Bohn et al. 2009;Fig. 2). Three songs with high signal-to-
noise ratios were selected from each of ﬁve males. For echoloca-
tion passes and songs, intervals between syllables were replaced
by silence, and a small (0.5 ms) taper was added to the beginning
and end of each syllable using Signal 4.0 (Engineering Design,
Belmont, MA, U.S.A.). Files were then band-pass ﬁltered between
5 kHz and 80 kHz and peak-to-peak amplitudes were normalized
to 1.5 V.
One bout of three echolocation passes and one bout of three
songs were created for each caller (N¼5 for echolocation, N¼5 for
song) using random intervals from recordings (between 3.5 and
9.5 s for echolocation and between 0.3 and 5 s for song; Fig. 2). Each
echolocation playback and each song playback consisted of alter-
nating 270 s of background noise with one bout from each of the
ﬁve callers in random order. Echolocation bout durations ranged
from 15 to 19 s, while song bout durations ranged from 6 to 13 s.
The different durations for echolocation and song stimuli resulted
in each echolocation playback lasting about 24 min and each song
playback lasting about 21 min. For all analyses we corrected for this
by examining the number of responses per second. In the labora-
tory, only echolocation playbacks were conducted. In the ﬁeld, the
entire echolocation playback and the entire song playback were
conducted in alternating order at each site.
Vocalizations were recorded and digitized using an Avisoft
CM16 condenser microphone and Avisoft Ultrasoundgate
–1.50 2 4 6 8 10 12 14 16 18 20
0 0.2 0.4 0.6 0.8 1 1.2
Bout 1 Bout 2 Bout 3 Bout 4 Bout 5
270 s 270 s 270 s 270 s 270 s 20 s
Figure 2. Examples of playbacks. (a) Diagram of an entire playback of echolocation passes or songs. Each bout is from a different individual. (b) A bout of echolocation passes from
one bat. (c) A bout of songs from one bat. (d) Spectrograms of the boxed portion of (a). (e) Spectrogram of the boxed portion of (b).
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491 148 7
(300 kHz/16 bit). Recordings were divided into four contiguous
periods: (1) 20 s pre, immediately preceding stimuli, (2) echo or
song, beginning with and encompassing stimuli, (3) 20 s post,
immediately following stimuli and (4) 230 s control. For responses,
songs were identiﬁed from other vocalizations by their unique
alternating temporal pattern of Chirp A and Chirp B syllables (Fig.1;
also see Bohn et al. 2009). We visually inspected all recordings and
noted the duration and phrase sequence (each chirp, trill and buzz
phrase) of every song. Additionally, because sequences were so
variable, we also categorized songs into ‘song types’by excluding
repeats of trills and buzz phrases (Fig. 1).
We examined the number of song responses in two ways. First,
we pooled responses across all experiments for each stimulus type
and for each playback caller (N¼5 bats’echolocation passes, N¼5
bats’songs) and used a randomized block design ANOVA with
playback caller as a block, and period (control, pre, echo/song, post)
and stimulus type (echo or song) as factors. To verify whether our
ﬁndings were consistent across experiments and not due to one
exaggerated experiment, we performed an ANOVA as above, but
instead of pooling all responses, we kept each experiment separate
as a block (N¼10 sites in the wild). Statistics were conducted
identically for the captive colony experiments except they lacked
the song/echo effect (only echo playbacks were performed) and
there were only ﬁve experiments, each on a separate subject.
Using the experiments conducted at the natural colony, we
compared song structure during two distinct contexts: songs pro-
duced during an echolocation pass (see Results,Fig. 3) and songs
produced spontaneously in the roost during control and pre pe-
riods of all playbacks. For this analysis, we did not use post periods
because song production was slightly elevated and may have been
in response to stimuli. We also did not use responses to song
stimuli because there were almost no songs produced during these
periods. Although we were unaware what speciﬁcally triggered
spontaneous songs, we knew it was not a passing bat since no bats
were ﬂying during these experiments.
First, we used a chi-square contingency test to compare the
frequency of the four most common song types (phrase order with
buzz and trill repeats ignored) and complete phrase sequences
(including repeats) between echolocation playback periods (echo)
and control/pre periods (spontaneous). To test whether individuals
modiﬁed songs between contexts, we identiﬁed individuals using
their Chirp B syllables, which are highly stereotyped within
individuals but vary among individuals (Bohn et al. 2008). We
identiﬁed eight bats for which we had at least 10 songs and per-
formed a logistic regression using context as a factor and subject as
a block on the same song type and phrase order categories as the
chi-square contingency test. In addition, we identiﬁed 21 bats for
which we had at least ﬁve songs in both contexts and used paired t
tests to compare the duration of songs, the frequency of songs with
buzzes and the frequency of songs with trills between the two
contexts. Finally, for bats that produced trills or buzzes in at least
ﬁve songs in both contexts (N¼15 and 18, respectively), we used t
tests to determine whether the number of repeats varied between
During 10 playback experiments, we recorded songs at a rate of
more than 100 songs/h, with a total of 699 songs from 39 different
males (equipment failed during one song playback). Although the
number of singing males varied between roost sites, the density of
singers at a given site did not affect individual singing rates
(regression of average number of songs per male versus number
of males in a site: F
¼0.14, P¼0.7). Bats responded to echolo-
cation passes but not to conspeciﬁc songs (ANOVA: song/echo:
¼20.6, P<0.0001; control/pre/stimulus/post: F
P<0.0001; interaction: F
¼24.8, P<0.0001; Figs 3,4a). The
only signiﬁcant difference following Tukey tests (alpha ¼0.05) was
between echolocation periods and all others. These effects were
robust in that the same results were found when data were ana-
lysed relative to each experiment (song/echo: F
P<0.0001; control/pre/stimulus/post: F
Responses to echolocation passes were both rapid and robust.
Out of 50 sets of echolocation passes (N¼5 different bats’passes
per playback, N¼10 playbacks), bats responded 74% of the time
with a total of 215 songs. During the interval immediately pre-
ceding stimuli, we recorded songs only 6% of the time across
all experiments (three songs). Echolocation-provoked singing
occurred very rapidly with onset latencies ranging from 200 to
500 ms (Fig. 3). We also ran echolocation playbacks on isolated
males in our captive colony and found similar results: singing was
rapidly and robustly provoked via playbacks of echolocation passes
¼16.05, P¼0.0002; mean SE number of songs/20 s: echo
stimulus ¼5.8 1.2; pre ¼0.13 0.08; post ¼0.93 0.46; con-
trol ¼0.14 0.07; analysed relative to subject: F
P<0.0001; Fig. 3;Supplementary Video S1,Audio S4).
Number of songs
Time from stimulus onset (ms)
Figure 3. Number of songs produced by bats at 100 ms intervals relative to the start of stimulus onset of (a) songs (N¼135 stimulus songs, N¼10 response songs) or (b)
echolocation passes (N¼135 stimulus echolocation passes, N¼92 response songs). In (b), grey bars represent wild bats’responses; black bars are superimposed and represent
captive bats’responses (N¼28 songs, N¼45 echolocation passes).
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491148 8
For the natural colony experiments, we tested whether the
composition of songs produced during the stimulus periods of
echolocation playbacks differed from those produced spontane-
ously in the roost. First, we examined the distribution of phrase
sequences and song types. There were a total of 69 different phrase
sequences, of which 52% were the four most common sequences:
chirp, chirp-buzz, chirp-trill-buzz and chirp-trill-trill-buzz). We
found no correlation between phrase sequences and context using
the four most common sequences and pooling the remainder into
an ‘other’category (
¼7:0, N¼587 songs, P¼0.14). There was
also no difference between contexts for eight bats that produced
at least 10 songs in each context (logistic regression:
P¼0.28). There were 26 song types (phrase sequences excluding
trill and buzz repetitions (Fig. 1,Table 1), of which 84% were the four
most common types: chirp-trill-buzz, chirp-buzz, chirp-trill and
chirp. Comparing the relative frequency of these song types and a
ﬁfth ‘other’category between echolocation and spontaneous con-
texts revealed that more songs during echolocation passes were
chirp-trill-buzz than expected (38% of 213 echolocation response
songs versus 28% of 385 spontaneous songs), whereas more
spontaneous songs were chirp-trill (13% versus 7%) or chirp (22%
versus 15%) than expected (
¼13:0, P¼0.01). There was no
signiﬁcant difference in the frequency of chirp-buzz song types
(24% and 21% for echo and spontaneous, respectively) or in the
frequency of ‘other’songs (16% in both contexts). We used a logistic
regression on the same song types with bat as a block and context
as a factor for eight bats that had at least 10 songs in each context
and found similar results (
¼22:0, P¼0.0002; mean percentage
of song types during echo and spontaneous contexts: chirp: 16%
and 26%; chirp-trill: 3% and 18%; chirp-buzz: 27% and 15%; chirp-
trill-buzz: 40% and 27%; other: 18% and 13%).
Next, we examined four song parameters for individuals that
produced at least ﬁve songs in the two contexts: duration, number
of phrases, whether or not songs contained buzzes and whether or
not songs contained trills. We found that songs produced during
echolocation stimuli were signiﬁcantly shorter but also more
frequently included buzz phrases than spontaneously produced
songs (paired ttests: duration: t
¼2.8, P¼0.01, Fig. 4b; buzz
¼2.7, P¼0.01, Fig. 4c). There was no difference between
songs produced in the two contexts with respect tothe frequency of
¼0.64, P¼0.53) or number of phrases (t
P¼0.84). Finally, we tested whether the number of phrase repeats
differed between contexts for those songs that contained buzzes or
trills. We found no difference between contexts for the number of
buzz repeats (t
¼0.81, P¼0.43) or trill repeats (t
Playbacks mimicking the echolocation call sequences of passing
bats rapidly triggered stimulus-speciﬁc singing. The rapid response
and tight temporal correlation between the echolocation pulse
stimuli and the onset of the bats’singing leaves little doubt that the
evoked songs were directed to the perceived presence of a passing
bat. Although our experiments were conducted during the day,
when few bats normally ﬂy (except between roosts), we obtained
the same results in the laboratory and when bats were actually
ﬂying (Video S1). Thus, we are conﬁdent that bat song is a robust
and natural response to approaching conspeciﬁcs. Conspeciﬁc song,
however, did not evoke singing, even though the spectrotemporal
acoustic structure of some song syllables are very similar to
Distribution of song types produced by Brazilian free-tailed bats at the wild colony
during playback experiments
Song type Nsongs Percentage
chirp-trill-buzz 243 34.8
chirp-buzz 148 21.2
chirp 134 19.2
chirp-trill 65 9.3
chirp-trill-chirp 27 3.9
chirp-trill-chirp-buzz 27 3.9
chirp-trill-chirp-trill-buzz 17 2.4
chirp-trill-chirp-trill 5 0.7
Other (N¼18 song types) 33 4.7
26 song types 699
% with buzz phrase
Pre Stimulus Post Control
Number of songs/20 s
Figure 4. (a) Songs produced by bats before (pre), during (stimulus) and after (post)
playback of echolocation and song stimuli, and during the control period (control). (b)
Song duration and (c) percentage of songs with buzzes produced in response to
echolocation stimuli and produced spontaneously (control and pre periods of (a)).
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491 148 9
echolocation pulses (Bohn et al. 2008). This indicates that the
overlying temporal pattern of acoustic stimuli inﬂuenced the bats’
The rapid responses of bats to the echolocation pulses of
approaching conspeciﬁcs exemplify the use of echolocation signals
for social interactions. By eavesdropping on echolocation signals,
males can sing when they are certain a conspeciﬁc is in audible
range. Interestingly, echolocation calls of approaching conspeciﬁcs
also trigger song in the greater sac-winged bat, Saccopteryx bili-
neata (Knörnschild et al. 2012). In this species, different songs are
used based on the gender of the approaching conspeciﬁc, indicating
that males can identify gender based on echolocation call acoustics
(Knörnschild et al. 2012). Although we only played male echolo-
cation calls, it is unlikely that free-tailed bats recognize the gender
of the caller since research has shown that there is a great deal of
within-individual variation in echolocation calls in this species and
no differences between males and females in echolocation call
structure (Gillam & McCracken 2007).
Echolocation-evoked songs were more likely to include all three
of the main song phrases but were shorter than spontaneous songs.
The shorter duration of these songs was not attributable to fewer
numbers of phrases but instead likely reﬂect shorter durations of
the phrases themselves. Echolocation-evoked singing likely notiﬁes
passing bats of the presence and location of an occupied and suit-
able day roost, a critical need for migratory bats that may not be
familiar with the local territory. These shorter songs are tailored to
the context in which it they are used because they must rapidly
convey sufﬁcient information to capture the attention of a ﬂying
conspeciﬁc that would pass beyond hearing range within 1e2s.
Although these data do not show whether echo-induced songs are
directed exclusively at females or males, these songs may serve
both functions; attracting females and warning males, much like
what is seen in songbirds (Kroodsma & Miller 1996). The function of
spontaneous singing is less clear. Song variation may reﬂect a va-
riety of social contexts in these densely crowded roosts, such
as attracting females to the male’s location within the roost,
deﬁning or maintaining dominance hierarchies, or promoting
Our ﬁnding that free-tailed bats vary the syntax of their songs in
different social contexts is unique among mammals. Chick-a-dee
calls that have similar combinatorial syntax also show correla-
tions between sequence order and social context (Ficken et al.
1994). However, even songbirds that do not use combinatorial
syntax, per se, and can vary song structure and syntax across
different contexts (Sossinka & Böhner 1980;Dunn & Zann 1996;
Woolley & Doupe 2008;Byers & Kroodsma 2009). This variability
has been closely tied to neural activity in an avian analogue of the
mammalian striatothalamocortical network, the songbirds’ante-
rior forebrain pathway (Jarvis et al. 1998,2005;Kao et al. 2005). A
similar network is believed to be important for speech (Jarvis
2004), and, in a pathological state, may underlie a plethora of hu-
man speech disorders (Lieberman 2007). Striatal dopamine sub-
serves vertebrate vocal plasticity and appears to be a key substance
for inducing song variability (Jarvis et al. 1998;Kao et al. 2005). In
mammals, incorporation of the striatothalamic network into the
vocal control circuitry may have been a critical step towards the
evolution of human speech (Doupe & Kuhl 1999;Jarvis 2004;
Jürgens 2009), yet vocal plasticity is rare in mammals with corre-
spondingly little evidence that the basal ganglia participates in
mammalian vocal communication (Jürgens 2009). Unlike most
mammals, however, echolocating bats routinely modulate the
timing and acoustic structure of their vocalizations. Free-tailed bats
show vocalization-related neuronal activity in the dorsolateral
striatum (Schwartz & Smotherman 2011) and pharmacological
manipulations of striatal dopamine profoundly inﬂuence
sensorimotor control of the free-tailed bat’s voice (Tressler et al.
2011). We have yet to show whether bat song composition is
directly inﬂuenced by basal ganglia activity, but similarities in the
way free-tailed bats and songbirds vary song composition indicate
the two disparate groups may share common design elements
within their song control circuitry. Identifying the neural basis for
these similarities may provide a useful new avenue for exploring
how, when and why vocal plasticity evolved in mammals.
We thank our undergraduate research assistants N. Tedford, S.
Trent, K. Rogers and M. Gutierrez for help conducting experiments
and collecting data at the ﬁeld sites. We thank B. Earnest and J.
Jarvis for help training bats and maintaining the captive colony, and
P. Narins and W. Metzner for comments on an earlier version of the
manuscript. We are particularly grateful to the Texas A&M
Department of Athletics for providing access to the bats and
allowing us to conduct these experiments within and around their
facilities. This research was supported by Texas A&M University.
Supplementary material for this article is available, in the online
version, at http://dx.doi.org/10.1016/j.anbehav.2013.04.002.
Arnold, K. & Zuberbühler, K. 2006. Semantic combinations in primate calls. Nature,
Balaban, E. 1988. Bird song syntax: learned intraspeciﬁc variation is meaningful.
Proceedings of the National Academy of Sciences, U.S.A.,85, 3657e3660.
Behr, O. & von Helversen, O. 2004. Bat serenades: complex courtship songs of the
sac-winged bat (Saccopteryx bilineata). Behavioral Ecology and Sociobiology,56,
Bohn, K. M., Schmidt-French, B., Ma, S. T. & Pollak, G. D. 2008. Syllable acoustics,
temporal patterns and call composition vary with behavioral context in
Mexican free-tailed bats. Journal of the Acoustical Society of America,124, 1838e
Bohn, K. M., Schmidt-French, B., Schwartz, C., Smotherman, M. & Pollak, G.
2009. Versatility and stereotypy of free-tailed bat songs. PLoS One,4, e6746.
Boughman, J. W. 1998. Vocal learning by greater spear-nosed bats. Proceedings of
the Royal Society B,265,227e233.
Byers, B. E. & Kroodsma, D. E. 2009. Female mate choice and songbird song rep-
ertoires. Animal Behaviour,77,13e22.
Catchpole, C. K. & Slater, P. J. B. 2008. Bird Song: Biological Themes and Variations.
Cambridge: Cambridge University Press.
Cerchio, S., Jacobsen, J. K. & Norris, T. F. 2001. Temporal and geographical variation
in songs of humpback whales, Megaptera novaeangliae: synchronous change in
Hawaiian and Mexican breeding assemblages. Animal Behaviour,62,313e329.
Clarke, E., Reichard, U. H. & Zuberbühler, K. 2006. The syntax and meaning of wild
gibbon songs. PLoS One,1, e73.
Davidson, S. M. & Wilkinson, G. S. 2002. Geographic and individual variation in
vocalizations by male Saccopteryx bilineata (Chiroptera: Emballonuridae).
Journal of Mammalogy,83, 526e535.
Deecke, V. B., Ford, J. K. B. & Spong, P. 2000. Dialect change in resident killer
whales: implications for vocal learning and cultural transmission. Animal
Doupe, A. J. & Kuhl, P. K.1999. Birdsong and human speech: common themes and
mechanisms. Annual Review of Neuroscience,22, 567e631.
Dunn, A. M. & Zann, D. A. 1996. Undirected song in wild zebra ﬁnch ﬂocks: con-
texts and effects of mate removal. Ethology,102, 529e539.
Ficken, M. S., Hailman, E. D. & Hailman, J. P. 1994. The chick-a-dee call system of
the Mexican chickadee. Condor,96,70e82.
Foote, A. D., Grifﬁn, R. M., Howitt, D., Larsson, L., Miller, P. J. O. & Hoelzel, A. R.
2006. Killer whales are capable of vocal learning. Biology Letters,2, 509e512 .
Garland, E. C., Goldizen, A. W., Rekdahl, M. L., Constantine, R., Garrigue, C.,
Hauser, N. D., Poole, M. M., Robbins, J. & Noad, M. J. 2011. Dynamic horizontal
cultural transmission of humpback whale song at the ocean basin scale. Current
Gillam, E. H. & McCracken, G. F. 2007. Variability in the echolocation of Tadarida
brasiliensis: effects of geography and local acoustic environment. Animal
Hailman, J. P. & Ficken, M. S. 1986. Combinational animal communication with
computable syntax: chick-a-dee calling qualiﬁes as ‘language’by structural
linguistics. Animal Behaviour,34, 1899e1901.
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491149 0
Hailman, J. P., Ficken, M. S. & Ficken, R. W. 1987. Constraints on the structure of
combinatorial ‘chick-a-dee’calls. Ethology,75,62e80.
Hammerschmidt, K. & Fischer, J. 2008. Constraints in primate vocal production.
In: Evolution of Communicative Flexibility: Complexity, Creativity, and Adaptability
in Human and Animal Communication (Ed. by D. K. Oller & U. Griebel), pp. 93e
119. Cambridge, Massachusetts: MIT Press.
Holy, T. E. & Guo, Z. 2005. Ultrasonic songs of male mice. PLoS Biology,3, e386.
Janik, V. M. 2000. Whistle matching in wild bottlenose dolphins (Tursiops trunca-
tus). Science,289, 1355e1357.
Jarvis, E. D. 2004. Learned birdsong and the neurobiology of human language.
Behavioral Neurobiology of Birdsong,1016,749e777.
Jarvis, E. D., Scharff, C., Grossman, M. R., Ramos, J. A. & Nottebohm, F. 1998. For
whom the bird sings: context-dependent gene expression. Neuron,21, 775e788.
Jarvis, E. D., Gunturkun, O., Bruce, L., Csillag, A., Karten, H., Kuenzel, W.,
Medina, L., Paxinos, G., Perkel, D. J., Shimizu, T., et al. 2005. Avian brains and a
new understanding of vertebrate brain evolution. Nature Reviews Neuroscience,
Jürgens, U. 2009. The neural control of vocalizations in mammals: a review. Journal
Kao, M. H., Doupe, A. J. & Brainard, M. S. 2005. Contributions of an avian basal
gangliaeforebrain circuit to real-time modulation of song. Nature,433, 638e
Knörnschild, M., Behr, O. & Von Helversen, O. 2006. Babbling behavior in the sac-
winged bat (Saccopteryx bilineata). Naturwissenschaften,93,451e454.
Knörnschild, M., Jung, K., Nagy, M., Metz, M. & Kalko, E. 2012. Bat echolocation
calls facilitate social communication. Proceedings of the Royal Society B,279,
Knornschild, M., Nagy, M., Metz, M., Mayer, F. & von Helversen, O. 2010. Complex
vocal imitation during ontogeny in a bat. Biology Letters,6,156e159.
Kroodsma, D. E. & Miller, E. H. 1996. Ecology and Evolution of Acoustic Communi-
cation in Birds. Ithaca, New York: Comstock.
Lieberman, P. 2007. The evolution of human speech. Current Anthropology,48,39e
Maynard Smith, J. & Szathmáry, E. 1995. The Major Transitions in Evolution. Oxford:
Oxford University Press.
Marler, P. & Slabbekoorn, H. 2004. Nature’s Music, the Science of Birdsong.
Amsterdam: Elsevier Academic Press.
Mitani, J. C. & Marler, P. 1989. A phonological analysis of male gibbon singing
Ouattara, K., Lemasson, A. & Zuberbühler, K. 2009. Campbell’s monkeys concat-
enate vocalizations into context-speciﬁc call sequences. Proceedings of the Na-
tional Academy of Sciences, U.S.A.,106, 22026e22031.
Payne, R. S. & McVay, S. 1971. Songs of humpback whales. Science,173 , 585e597.
Rusch, K. M., Pytte, C. L. & Ficken, M. S. 1996. Organization of agonistic vocali-
zations in black-chinned hummingbirds. Condor,98,557e566.
Schwartz, C. & Smotherman, M. 2011. Mapping vocalization-related immediate
early gene expression in echolocating bats. Behavioral Brain Research,224,358e
Snowdon, C. T. 2009. Plasticity of communication in nonhuman primates. Advances
in the Study of Behavior,40 (40), 239e276.
Soard, C. M. & Ritchison, G. 2009. ‘Chick-a-dee’calls of Carolina chickadees convey
information about degree of threat posed by avian predators. Animal Behaviour,
Sossinka, R. & Böhner, J. 1980. Song types in the zebra ﬁnch Poephila guttata
castanotis.Zeitschrift für Tierpsychologie,53, 123132.
Tressler, J., Schwartz, C., Wellman, P., Hughes, S. & Smotherman, M. 2011.
Regulation of bat echolocation pulse acoustics by striatal dopamine. Journal of
Experimental Biology,214, 3238e3247.
Woolley, S. C. & Doupe, A. J. 2008. Social context-induced song variation affects
female behavior and gene expression. PLoS Biology,6, e62.
Wright, T. F. & Dahlin, C. R. 2007. Pair duets in the yellow-naped amazon (Amazona
auropalliata): phonology and syntax. Behaviour,144, 207e228.
K. M. Bohn et al. / Animal Behaviour 85 (2013) 1485e1491 1491