Biol Invasions DOI 10.1007/s10530-016-1327-7

ORIGINAL PAPER

Invasion of the acoustic niche: variable responses by native species to invasive American bullfrog calls Camila Ineu Medeiros . Camila Both . Taran Grant . Sandra Maria Hartz

Received: 21 February 2016 / Accepted: 7 November 2016 Ó Springer International Publishing Switzerland 2016

Abstract Biological invasions are a major threat to biodiversity. Invasive species that use acoustic communication can affect native species through interference in the acoustic niche. The American Bullfrog Lithobates catesbeianus is a highly invasive anuran that is widely distributed in the Brazilian Atlantic Rainforest. Adult male bullfrogs emit loud advertisement calls at frequencies that overlap with the calls of several native species of frogs. Given that spectral overlap is a major factor in acoustic masking, the purpose of this study was to test the effects of the acoustic invasion of L. catesbeianus on native frogs that have calls with and without spectral overlap with the invader. In field experiments, we exposed calling males of two overlapping species and two nonoverlapping species to recorded bullfrog vocalizations, white noise, and the vocalization of another

C. I. Medeiros  S. M. Hartz Departamento de Ecologia, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul 91501-970, Brazil C. I. Medeiros (&)  C. Both Departamento de Ecologia e Evoluc¸a˜o, Universidade Federal de Santa Maria, Av. Roraima s/no, Santa Maria, Rio Grande do Sul CEP 97105-900, Brazil e-mail: [email protected] T. Grant Departamento de Zoologia, Instituto de Biocieˆncias, Universidade de Sa˜o Paulo, Sa˜o Paulo, Sa˜o Paulo 05508-090, Brazil

native frog species. To identify effects, we compared calls recorded before, during, and after exposure. Our results showed that native species altered their calls in response to the bullfrog calls. However, we also observed similar responses to white noise and heterospecific native calls. Both the invasive and heterospecific calls were emitted at low frequencies, which suggests that the observed responses might be specific to low-frequency calls. Our results provide evidence that the introduction of new sounds can cause native species to modify their calls, and that responses to exogenous sounds are species- and stimulusspecific. Keywords Atlantic forest  Biological invasions  Amphibian conservation  Bioacoustics  Masking interference  Lithobates catesbeianus

Introduction It is well known that biological invasions are among the most serious threats to biodiversity. Invasions can cause homogenization of ecosystems (Mack et al. 2000) and other ecological impacts through competition, predation, hybridization, or alteration of the dynamics of the native community (Primack and Rodrigues 2001; Sax et al. 2007). Recently, Both and Grant (2012), Farina et al. (2013), Bleach et al. (2015), and Tennessen et al. (2016) drew attention to another

123

C. I. Medeiros et al.

mechanism by which invasive species can affect native species: interference in the acoustic niche. These studies found that native species of amphibians and birds altered their vocalizations, i.e. modified their acoustic niches, in response to vocalizations by invasive species. The acoustic niche minimally comprises the microhabitat used for calling, the time that calling activity takes place, and the acoustic structure of the advertisement call (Sinsch et al. 2012). When sounds, such as anthropogenic noises and sounds produced by the invasive species, are introduced into the environment, they can hinder species communication by attenuating or degrading signals (Brumm et al. 2004). The masking of acoustic signals can directly affect reproduction and/or energy consumption, which can result in higher energy costs, loss of mating or foraging opportunities, unnecessary aggressive interactions, and other critical behaviors (Edge and Marcum 1985; Krausman et al. 1986; Bradbury and Vehrencamp 1998). Acoustic interactions are important in several ecological functions, including mate choice, territorial defense, predator–prey interactions, maintenance of social cohesion, and spatial orientation (Reby et al. 1999; Rheindt 2003; Wells 2007; Hollen and Radford 2009; Nogueira et al. 2012). Species that have evolved in sympatry employ strategies to minimize interference in the acoustic niche and a new sound in the environment can be incompatible with the existing acoustic partitioning by native species (Rabin et al. 2003). The effects of new anthropogenic sound sources have been well studied for diverse animal taxa. For instance, traffic and airplane noise mask acoustic signals by overlapping with the native vocalizations in space, time, and frequency (Halfwerk et al. 2011) and can have significant impacts on amphibian vocalizations and other aspects of their breeding biology (Bee and Swanson 2007; Parris et al. 2009; Cunnington and Fahrig 2010; Kaiser et al. 2011; Tennessen et al. 2014; Vargas-Salinas et al. 2014; Kruger and Preez 2016). In anurans, acoustic interference can have severe consequences for reproductive success, since vocalization is related directly to mate choice (Gerhardt 1991). In this animal group, females of many species choose males based on their vocalization characteristics, such as amplitude and dominant frequency (Ryan 1988; Gerhardt 1991). The calls males emit can identify their sex, identity, reproductive state, and

123

spatial position (Wells 1977) and have been shown to be a reliable indicator of male quality (e.g., Welch et al. 1998; Forsman and Hagman 2006; Jaquie´ry et al. 2010). Acoustic signals spread over a large area and simultaneously attract females and transmit a territorial message to other males (Martof and Thompson 1958; Wells 1977). Anurans naturally experience acoustic competition, given that they often vocalize in aggregations where their calls overlap with those of con- and heterospecifics (Narins 1995). Due to this competition, species commonly use strategies to minimize signal overlap, like increasing signal amplitude and modulating spectral and/or temporal features of calls and temporal segregation of calling activities (Lane and Tranel 1971; Dubois and Martens 1984; Zelick and Narins 1985; Narins 1995; Sueur 2002; Bee and Swanson 2007), and sympatric species whose calls overlap extensively exert selective pressure on each other (Gerhardt and Huber 2002). Similar strategies are used by anurans in response to anthropogenic noise and invasive species calls. Frogs have been shown to change the spectral and temporal patterns of their calls when exposed to such noises. For example, Bee and Swanson (2007), Kaiser and Hammers (2009), Parris et al. (2009), and Cunnington and Fahrig (2010) found that some anuran species alter frequencies in response to high levels of spectral overlap with anthropogenic noise, even when this decreases call propagation, and Both and Grant (2012) reported shifts in call frequencies in response to calls by an invasive species. The results of those studies suggest that spectral overlap is a major factor in signal masking, as observed in birds (Planque and Slabbekoorn 2008). Furthermore, Bleach et al. (2015) and Tennessen et al. (2016) reported changes in temporal parameters caused by exposure to invasive species calls. In these studies, authors reported that some species used strategies like shortening calls or emitting them in gaps between the signals emitted by invasive species. In this study, we assessed the potential effects of invasion of the acoustic niche by the American bullfrog (Lithobates catesbeianus Shaw 1802) on native species that emit calls with and without spectral overlap with the invader. L. catesbeianus is highly invasive, with invasive populations found in more than 40 countries (Lever 2003) and at least 130 Brazilian municipalities (Both et al. 2011). Its vocalization

Invasion of the acoustic niche by bullfrogs

covers a broad frequency spectrum, overlapping with the frequencies of calls of many native species, and is loud and conspicuous (Capranica 1968). It also has low environmental degradation and attenuation (Capranica 1968; Llusia et al. 2013). We performed a field experiment introducing L. catesbeianus vocalizations in an L. catesbeianus-free biological research and conservation area in southern Brazil. We tested the effects of L. catesbeianus calls on vocalizations of two native species that exhibit spectral overlap with the invasive species and two that do not. First, we expected that all species would respond to L. catesbeianus acoustic signals by altering their call parameters because any new sound in the environment has the potential to affect the communication between native species. In this case, specifically, we expected changes in temporal parameters like signal length and signal rate because they are highly variable in anurans and are modulated by many species to avoid masking (Kaiser and Hammers 2009; Cunnington and Fahrig 2010; Bleach et al. 2015). Second, previous studies have reported frequency shifts in response to noises (Parris et al. 2009; Cunnington and Fahrig 2010), especially when the frequencies of calls and noises are similar, so we hypothesized that the two species with spectral overlap would also change the dominant frequency of their calls to avoid masking, whereas the dominant frequencies of the species without spectral overlap would remain unchanged.

Materials and methods Study area and species The study was conducted at the Centro de Pesquisas e Conservac¸a˜o da Natureza Pro´-Mata (29°350 S, 050°150 W). To date, L. catesbeianus has never been observed at or near this well-studied locality (Kwet and Di-Bernardo 1999; Kwet 2001; Both et al. 2009). We performed experiments in the breeding season, during austral spring and summer (September 2013– March 2014). We studied the effects of L. catesbeianus vocalizations on the vocalizations of native species with and without spectral overlap with the invasive species (Kwet 2001). The species with spectral overlap were

Bischoff’s tree frog (Hypsiboas bischoffi) and the snouted tree frog (Scinax perereca) and the species without spectral overlap were the fine-lined tree frog (Hypsiboas leptolineatus) and the lesser tree frog (Dendropsophus minutus). We chose these species due to their acoustic properties and local abundance during the field experiments. Descriptive characteristics of calls are provided in Table 1. The invasive species, L. catesbeianus, has a vocal repertoire composed of seven different calls (Capranica 1968), the most common of which is the advertisement call (Fig. 1a). The wave shape of the advertisement call is nearly periodic, with an almost harmonic relationship between spectral components, and it covers a broad frequency band. As a native species control, we also exposed the target species to advertisement calls of the cururu toad (Rhinella icterica), which breeds contemporaneously in sympatry with the target species. Rhinella icterica occurs in central, southeastern, and southern Brazil, northeastern Argentina, and eastern Paraguay (Kwet and DiBernardo 1999) in a variety of habitats, including forested and open areas, and breeds in permanent or temporary water bodies (Silvano et al. 2010). The advertisement call (Fig. 1b) of R. icterica is composed of a single, multipulsed note emitted at low frequencies (Kwet 2001; Pombal 2010). Hypsiboas bischoffi is endemic to southern and southeastern Brazil (Marcelino et al. 2009). The species has two typical kinds of calls: type A notes, which comprise the advertisement call (Fig. 2a; Table 1), and type B notes, which are probably territorial calls (Pombal 2010). Scinax perereca occurs in the Atlantic Rainforest from southern Sa˜o Paulo State in Brazil to Misiones Province in Argentina and produces a single type of advertisement call (Fig. 2b) and two types of aggressive notes (Pombal et al. 1995; Magrini and Giaretta 2001). Hypsiboas leptolineatus occurs in wetlands and river systems in high-elevation grasslands on the Araucaria Plateau in Rio Grande do Sul, Santa Catarina and Parana´ states (Cruz and Caramaschi 1998; Kwet and Di-Bernardo 1999; Hiert and Moura 2007). Two kinds of calls are known for H. leptolineatus, an advertisement call (Fig. 2c) and a territorial call. Finally, the nominal taxon D. minutus is widely distributed in South America where it is typically found in open areas (Frost 2015). However, recent molecular analyses have shown that D. minutus comprises multiple species (Gehara et al. 2014). All of

123

C. I. Medeiros et al. Table 1 Numerical characteristics of calls of the invader Lithobates catesbeianus, native species that were recorded in field experiments, and the native-species control, Rhinella icterica Note duration (s)

Inter-note interval (s)

Frequency range (Hz)

Dominant frequency (Hz)

Lithobates catesbeianus

0.6–1.5

0.8

90–4000

200–400/1000–2000

Hypsiboas bischoffi

0.05–0.1

0.06–0.66

1100–2500

1400–2100

Scinax perereca

0.28–0.37

0.71–1.60

950–7000

1300–2310 and/or 27,400–3900

Hypsiboas leptolineatus

0.04–0.1

0.01–0.12

3500–5500

3500–5200

Dendropsophus minutus

0.05–0.08/0.13–0.17

0.05–0.85

2000–6000

4000–5500

Rhinella icterica

0.03–0.04

0.05–0.058

400–900

500–600

Fig. 1 Spectrograms and oscillograms of the advertisement calls of the anurans that were used as playback stimuli, Lithobates catesbeianus (a) and Rhinella icterica (b). The calls are the same ones used in the experiments and represent a single individual of each species

the samples from the region of our study site pertain to lineage 39 of Gehara et al. (2014). The local frogs have three kinds of notes, referred to as notes A (Fig. 2d), B and C (Cardoso and Haddad 1984), and these notes are combined to form different calls. Acoustic stimuli To test if L. catesbeianus advertisement calls cause the native species to alter their calls, we followed the A– B–A protocol proposed by McGregor et al. (1992), whereby we recorded the calling activity of each individual for 5 min prior to exposure to L. catesbeianus vocalizations (A), 5 min during L. catesbeianus vocalization playback (B), and 5 min postexposure (A). In order to test if alterations were a specific response to L. catesbeianus vocalizations or a general response to environmental noise, we also repeated the same A–B–A procedure using white noise and the advertisement call of R. icterica, a large species that is naturally sympatric with the test species

123

and also has a low dominant frequency like L. catesbeianus. The acoustic stimuli of L. catesbeianus and R. icterica were obtained by recording vocalizations of solitary males. Recordings were obtained with a unidirectional microphone (Sennheiser ME 66-K6P) and a digital recorder (Sony PCM-M10, sampling rate of 96 kHz, sample size 24 bits). For all the stimuli, we standardized the sound pressure level (SPL) at 75 dB (C-weighting; Instrutemp ITDEC-4000) 1 m from the source. Lithobates catesbeianus calls were recorded in the municipality of Faxinal do Soturno, southern Brazil (29883´S, 053882´W). The stimulus-playback of L. catesbeianus (snout–vent length = 128 mm; 20.7 °C) has advertisement call notes of 6.6 s and a dominant frequency of 187.5 Hz, separated by regular intervals of 30 s (for details see Both and Grant 2012). The R. icterica calls were recorded at our study site (male snout–vent length: 136 mm; 19.1 °C). The stimulus of R. icterica has advertisement call notes of 4.44 s and a dominant frequency of 689.1 Hz,

Invasion of the acoustic niche by bullfrogs Fig. 2 Spectrograms and oscillograms of the advertisement calls of the native species exposed to playback stimuli. Species with spectral overlap with L. catesbeianus, Hypsiboas bischoffi (a) and Scinax perereca (b); and species without spectral overlap: H. leptolineatus (c) and Dendropsophus minutus (d)

separated by regular intervals of 30 s, similar to the L. catesbeianus call. The white noise was the only stimulus with a different temporal structure, being reproduced continuously for 5 min during the stimulus-playback period. The white noise stimulus has frequency band limits between 0 and 8268.8 Hz and was obtained from the SimplyNoise website (available from: https://simplynoise.com/). Like the other playbacks, exposure to white noise was standardized as SPL 75 dB at 1 m distance from the source (sampling rate of 24 bits/96 kHz). Experiment We recorded the responses of 18 actively calling males each of D. minutus, H. bischoffi, and H. leptolineatus and 12 of S. perereca. Each individual was exposed to a single stimulus type, with one-third of the individuals of each species assigned randomly to each stimulus. Recordings were obtained from the same equipment that was used to record the stimulus, using

the same parameters. The distances of the microphone and the speaker from the focal male were standardized at 1 m and all focal males were at least 5 m apart. Before each experiment, we measured the temperature, relative humidity, and environmental sound (SPL). The SPL was measured 1 m from the water body with a sound level meter (C-weighting; Instrutemp ITDEC-4000). Prior to each experiment, we actively searched for conspecific males within 5 m of the focal male and removed all that we detected. After each experiment, focal males were captured, photographed, measured in the laboratory with digital calipers (Starret—798), and held captive for up to 7 days to avoid pseudo-replication, after which they were released at the point of capture. Captive individuals were kept in containers with vegetation and wet cotton at ambient temperatures and provided with food (ants, small crickets, and termites) ad libitum. All experimental procedures were approved by the Centro Nacional de Pesquisa e Conservac¸a˜o de Re´pteis e Anfı´bios—Instituto Chico Mendes de

123

C. I. Medeiros et al.

Conversac¸a˜o da Biodiversidade (RAN-ICMBIO— Permit No. 42411-1). Call analyses We divided the 15 min recordings for each individual into three 5 min files corresponding to pre-stimulus, stimulus, and post-stimulus time periods and used Raven Pro v.1.5 (Bioacoustics Research Program 2014) to construct oscillograms. To calculate the signal rate, given as (notes-1)/min, we counted the advertisement and aggressive notes emitted in each 5 min period. To measure other advertisement call parameters, we randomly selected 20 notes in each time period for S. perereca and D. minutus (60 per individual) and 10 notes for H. leptolineatus and H. bischoffi (30 per individual), because some individuals of the latter species emitted fewer than 20 notes per period. Hypsiboas leptolineatus and H. bischoffi emitted aggressive calls as well, which we counted and analyzed separately by randomly selecting 20 notes per period (60 per individual). Notes were randomized in Excel (rand function; Microsoft Excel 2010. available from: https://products.office.com/pt-BR/ excel). Spectrograms were constructed using 16 bit resolution, 44.1 kHz sampling rate and 256 point fast Fourier transform. We analyzed the following temporal and spectral parameters of the randomly selected notes: Dominant frequency (frequency at peak amplitude, Hz), Center frequency (the frequency that divides the signal into two frequency intervals of equal energy, Hz), Note duration (time that a single note lasted, s) and Inter-note interval (time between sequential two sequential notes, s). All parameters were calculated using Raven Pro v. 1.5. We measured the temporal overlap between the emitted calls and the biotic stimuli by counting the number of emitted notes that overlapped with any note of the stimulus-species (L. catesbeianus and R. icterica) and the number of emitted notes that did not overlap with any notes of the stimulus-species.

Statistical analyses Differences in call parameters among the three time periods were tested by Analyses of Variance with randomizing tests. Stimulus type and period (A-B-A)

123

were treated as fixed factors and the individuals were treated as blocks. We used the null hypothesis that any note and/or any signal rate could be emitted by a given individual in any period and during any stimulus. We utilized Pseudo-F statistics as the test criteria and obtained null distributions from 1000 permutations (Anderson 2001). In order to determine which periods and stimuli differed, we employed pairwise contrast analyses (Pillar and Orlo´ci 1996). All analyses were performed in MULTIV v. 2.42 (Pillar 2006). In order to assess if males adjusted the timing of their calls during the stimulus period to avoid masking by the anuran stimuli, we compared the observed proportion of overlapping notes with the proportion of overlapping notes expected by chance. The expected proportions were calculated through randomization of the stimulus-playbacks, where we considered the total time (s) in the 300 s stimulus period occupied by L. catesbeianus or R. icterica notes (36.45 s occupied by L. catesbeianus notes and 91.56 s occupied by R. icterica notes). We used Fisher’s Exact Test (Fisher 1935) to test if the degree of temporal overlap between the recorded males and the acoustic signals differed from that expected by chance. Analyses were performed and figures constructed in the R environment (R 2.2.1, Development Core Team 2012). The oscillograms and spectrograms were constructed in the Seewave package (Sueur et al. 2008).

Results Hypsiboas bischoffi Of the 18 recorded individuals, 14 emitted both advertisement and aggressive calls, in all three time periods. The advertisement signal rate differed among periods (F = 0.228; p = 0.039) but not among stimulus types (F = 0.022; p = 0.226). The interaction term was non-significant (F = 0.073, p = 0.721). Signal rate increased during the stimuli and was even higher after the exposure (21 notes/min; Fig. 3a), but post hoc comparisons showed that only pre (14 calls/ min) and post-stimuli signal rates (21 calls/min) differed. No modifications were observed in the aggressive signal rate. Among the call parameters evaluated, inter-note interval and center frequency did not differ among periods or stimulus type for either advertisement or aggressive calls (p [ 0.1 for both). In

Invasion of the acoustic niche by bullfrogs

Fig. 3 Main effects (means and SE) of playback stimuli on advertisement and aggressive call parameters of Hypsiboas bischoffi. Circles with solid lines indicate the Lithobates

catesbeianus call effects, squares with dashed lines indicate the Rhinella icterica call effects, and triangles with dotted lines indicate the white-noise effects

contrast, males altered the dominant frequency of the advertisement calls, with the changes depending on the stimulus type (F = 0.117; p = 0.023; see also Table 2) and on its interaction with time periods (F = 0.031; p = 0.014; Fig. 3b). The dominant frequency of the advertisement call significantly decreased during and after both the L. catesbeianus (from 1808.8 to 1762.8 and 1755 Hz, respectively) and R. icterica (from 1850.3 to 1802.6 and 1800.8 Hz, respectively) stimuli (Fig. 3b), but increased during and after the white-noise stimulus (from 1775.2 to 1810.5 and 1802.9 Hz, respectively). Advertisement call duration was affected by time period (F = 0.055; p = 0.001; Fig. 3c) and stimulus type (F = 0.404; p = 0.001), with no interaction effect (F = 0.001; p = 0.958). Calls were significantly longer during the stimulus (0.085 s) and post-stimulus periods (0.084 s)

than during the pre-stimulus period (0.079 s). Calls recorded during the R. icterica stimulus were slightly shorter in all periods (Fig. 3c). Aggressive call duration was affected by time period (F = 0.060; p = 0.001), stimulus type (F = 0.153; p = 0.01, and the interaction between them (F = 0.059; p = 0.001; Fig. 3d). Post-hoc comparisons showed that aggressive calls were shorter during the stimuli, and this was especially strong during the L. castebeianus (stimulus 0.0255 vs. pre- 0.0275 and post- 0.0285 s) and white noise (stimulus 0.0199 vs. pre- 0.0234 and post0.0265 s; Fig. 3d) stimuli. Other parameters of aggressive calls did not change (p [ 0.1 for all cases). Fisher’s Exact Test showed that the proportion of calls of H. bischoffi that overlapped the acoustic signals from both stimuli did not differ from the null expectation (p = 0.29 for L. catesbeianus; p = 0.11

123

C. I. Medeiros et al. Table 2 Summary of predictions and results observed for all four native species exposed to invasive Lithobates catesbeianus calls and experimental controls Species

Treatment

Prediction

Outcome

Invasive species

Shift in spectral and temporal parameters

Dominant frequency decreased, signal rate and note duration increased

No shift in spectral parameters. Possibly shift in temporal parameters

Dominant frequency decreased, signal rate and note duration increased

No or weak effects on spectral parameters

Dominant frequency increased, signal rate and note duration increased

Overlap with invasive species Hypsiboas bischoffi

L. catesbeianus Native species control R. icterica

Synthetic noise control White-noise

Possibly shift in temporal parameters Scinax perereca

Invasive species

Shift in spectral and temporal parameters

Dominant frequency increased, note duration decreased

No shift in spectral parameters. Possibly shift temporal parameters

Dominant frequency increased, inter-note interval and note duration decreased

No or weak shift in spectral parameters, possibly shift in temporal parameters

Dominant frequency decreased, notes increased and inter-note interval increased

No shift in spectral parameters, possibly shift in temporal parameters

Dominant and center frequencies decreased, signal rate increased and note duration decreased

No shift in spectral parameters. Possibly shift temporal parameters

Dominant and center frequencies decreased and note duration increased, signal rate and internote interval decreased

Synthetic noise control White-noise

No or weak shift in spectral parameters, possibly shift in temporal parameters

Dominant and center frequencies increased, signal rate increased and note duration decreased

Invasive species

No shift in spectral parameters, possibly shift in temporal parameters

Dominant and center frequencies decreased, note duration increased

No shift in spectral parameters, possibly shift in temporal parameters

Note duration decreased

No or weak shift in spectral parameters, possibly shift in temporal parameters

Dominant and center frequencies increased

L. catesbeianus Native species control R. icterica

Synthetic noise control White-noise

Non Overlap with invasive species Hypsiboas leptolineatus

Invasive species L. catesbeianus

Native species control R. icterica

Dendropsophus minutus

L. catesbeianus

Native species control R. icterica

Synthetic noise control White-noise

123

Invasion of the acoustic niche by bullfrogs

for R. icterica). A summary of results for H. bischoffi and all other species is provided in Table 2.

Scinax perereca The signal rate of this species was not affected by any stimulus tested in this study (p [ 0.05 for all cases). The dominant frequencies changed depending on stimulus type (F = 0.0247; p = 0.003), and the interaction with time period (F = 0.0389; p = 0.001). Overall the dominant frequency increased during and after L. catesbeianus and R. icterica stimuli and decreased in response to the white noise. Post-hoc comparisons showed that the dominant frequency was significantly higher during the L. catesbeianus poststimulus period (3129.7 Hz) than in the pre-stimulus time (2959.1 Hz). For the other two stimuli dominant frequencies differed among the three time periods (Fig. 4a). Stimuli (F = 0.822; p = 0.001), time periods (F = 0.041; p = 0.002), and the interaction term

(F = 0.044; p = 0.001) also had significant effects on call duration (Fig. 4b). Calls were significantly shorter in the R. icterica post-stimulus period (0.2968 s) than pre- or stimulus periods (0.335 and 0.326 s, respectively; Fig. 4c). They also tended to decrease in response to L. catesbeianus and increase in response to R. icterica, although post hoc comparisons did not identify significant differences. Stimulus type (F = 0.024; p = 0.001) and the interaction term (F = 0.041; p = 0.01) had significant effects on inter-note interval. The inter-note interval did not differ significantly during L. catesbeianus calls but increased during and after the white noise (from 1.9138 to 3.2799 and 3.8865 s, respectively) and decreased in response to R. icterica calls (from 2.8784 to 1.5192 and 0.5274 s; Fig. 4c). Fisher’s Exact Test for temporal overlap showed that the proportion of calls of S. perereca that overlapped the stimulus signals did not differ from that expected by chance (p = 0.58 for L. catesbeianus; p = 0.67, for R. icterica).

Fig. 4 Main effects (means and SE) of playback stimuli on advertisement call parameters of Scinax perereca. Circles with solid lines indicate the Lithobates catesbeianus call effects, squares with dashed lines indicate the Rhinella icterica call effects, and triangles with dotted lines indicate the white-noise effects

123

C. I. Medeiros et al.

Hypsiboas leptolineatus Among the 18 recorded individuals, 14 emitted aggressive calls during the three time periods. The advertisement signal rate differed among the three time periods (F = 1.115; p = 0.003) and also depended on the stimulus type (F = 4.475; p = 0.043), but the interaction effect was non-significant (p [ 0.06). Post-hoc comparisons showed that males called at higher rates after being exposed to the L. castesbeianus and white noise stimuli compared to pre-stimuli periods. After exposure to L. catesbeianus calls males increased call rate by approximately 11 notes/min (from 18.73 notes/min to 29.61), and they increased it even more during and after the white noise stimulus, emitting on average 17 notes/min more than before exposure (before: 22.98 notes/min; after: 40.13 notes/min; Fig. 6a). No changes signal rate were observed in response to the R. icterica stimulus (see Fig. 5a). Dominant frequency responses depended on the stimulus type (F = 0.018; p = 0.044), and the interaction term was significant (F = 0.020; p = 0.042). Exposure to L. catesbeianus calls caused a significant decrease in the dominant frequency of calls, and frequencies were even lower after the stimulus ceased (before: 4125.8 Hz, during: 4017.1 Hz; after: 3900.4 Hz; Fig. 5b). The center frequencies showed the same trend, being affected only by L. catesbeianus stimuli and significantly decreasing across the three time periods (before: 4124.4 Hz; during: 4000.2 Hz; after: 3920.8 Hz; F = 0.003; p = 0.003; Fig. 5c). No other stimuli affected spectral parameters. Stimulus type, time period, and the interaction had significant effects on note duration, being shorter during and after white noise (0.06048, 0.05674 and 0.05225 s) and after the L. catesbeianus calls (0.070821, 0.07569 and 0.063672 s), but longer during and shorter after the R. icterica calls (0.059599, 0.061922 and 0.055894 s; F = 0.021; p = 0.010; Fig. 6). Inter-note intervals were significantly different among time periods (F = 0.0127; p = 0.031; Fig. 5e) and did not respond to stimulus type (p [ 0.06 for both), being shorter during and after exposure to stimuli than in the prestimulus period (before: 3.59 s; during: 3.27 s; after: 3.22 s). All parameters of the aggressive calls of H. leptolineatus were altered except the inter-note interval. The dominant frequency was affected by the

123

stimulus type (F = 0.095; p = 0.001) and period (F = 0.065; p = 0.001) and depended on the interaction between them (F = 0.041; p = 0.001). The dominant frequency was distinctly lower during the L. catesbeianus stimulus (3904.1 Hz) and post-stimulus (3885.1 Hz) periods than pre exposure (4138.7 Hz; Fig. 6a). Dominant frequencies were higher during and after the white noise and R. icterica stimulus periods than in the post-stimulus periods (4152.2, 4203.1 and 4059 Hz for white noise; 3993.2, 4155.3, and 4125 Hz for R. icterica), although the responses were weaker (Fig. 6a). The center frequency was also affected by stimuli (F = 0.0969; p = 0.001), periods (F = 0.0649; p = 0.001), and the interaction between them (F = 0.045; p = 0.001; Fig. 6b). Overall, the center frequency was distinctly lower during (3905.6 Hz) and after (3885.3 Hz) the L. catesbeianus stimulus than in the pre-stimulus period (4138.8 Hz) and significantly decreased after R. icterica stimulus (4062.4) compared to pre(4152.3 Hz) and stimulus periods (4203 Hz). Stimuli (0.375; p = 0.001), periods (F = 0.031; p = 0.001), and the interaction between them (F = 0.101; p = 0.001) also affected aggressive note duration (Fig. 6c). Note duration increased after the L. catesbeianus stimulus (before: 0.0227957 s; during: 0.0214015 s; after: 0.028106 s). Notes emitted during white noise stimulus (0.0213 s) were also longer than notes emitted before (0.0192 s). Fisher’s Exact Test indicated that the proportion of calls of H. leptolineatus that overlapped the stimulus signals did not differ from that expected by chance in either of the anuran playbacks tested (p = 0.73 for L. catesbeianus; p = 0.08 for R. icterica). Dendropsophus minutus The calls of D. minutus are complex, comprising three types of notes. The results below are based on type ‘‘A’’ notes (see Cardoso and Haddad 1984), which are emitted most frequently. The signal rate and inter-note interval were not altered by any of the stimuli (p [ 0.05, for all cases). The dominant frequency was affected by periods (F = 0.008; p = 0.031) and stimuli (F = 0.0747; p = 0.035), and the interaction term was significant (F = 0.012; p = 0.043). This frequency was distinctly lower during after L. catesbeianus stimulus (4205.5 Hz) than in pre- (4537.8 Hz) and stimulus (4502.9 Hz) periods; Fig. 7a). It was also

Invasion of the acoustic niche by bullfrogs

Fig. 5 Main effects (means and SE) of playback stimuli on advertisement call parameters of Hypsiboas leptolineatus. Circles with solid lines indicate the Lithobates catesbeianus

call effects, squares with dashed lines indicate the Rhinella icterica call effects, and triangles with dotted lines indicate the white-noise effects

distinctly higher during the white noise stimulus (4582.7 Hz) than in the other two periods (4347.6 and 4408.5 Hz; Fig. 7a). The center frequency was also affected by both factors (stimuli: F = 0.0713; p = 0.02; periods: F = 0.008; p = 0.015) and the interaction between them (F = 0.014; p = 0.021), showing the same patterns observed for the dominant frequency (Fig. 7b). Period (F = 0.027; p = 0.001), stimulus type (F = 0.051; p = 0.001), and the interaction term (F = 0.0153; p = 0.005) also had

significant effects on note duration. In general, notes were approximately 30 ms shorter during the R. icterica stimuli (0.164649 s) than in the other two periods (0.193992 and). Note duration significantly increased after L. catesbeianus stimuli (pre0.185953 s vs post-stimulus 0.223992 s; Fig. 7c). The proportion of D. minutus calls that overlapped with stimulus signals of both anurans did not differ from that expected by chance (p = 0.82 for L. catesbeianus; p = 0.97 for R. icterica).

123

C. I. Medeiros et al. Fig. 6 Main effects (means and SE) of playback stimuli on aggressive call parameters of Hypsiboas leptolineatus. Circles with solid lines indicate the Lithobates catesbeianus call effects, squares with dashed lines indicate the Rhinella icterica call effects, and triangles with dotted lines indicate the white-noise effects

Discussion As expected, all species responded to L. catesbeianus vocalizations by modifying temporal call parameters, although the specific responses varied across species. Moreover, all species responded to both the sympatric native species R. icterica calls and white noise by changing temporal parameters, which suggest that this may be a generalized response to environmental noise. Similar modifications of temporal parameters have been observed in response to both anthropogenic noise (e.g. Rheindt 2003; Bee and Swanson 2007; Kaiser and Hammers 2009; Parris et al. 2009; Cunnington and Fahrig 2010) and invasive species (Bleach et al. 2015; Tennessen et al. 2016). Although we expected changes in temporal call parameters in all species, we only expected changes in spectral parameters in species whose frequencies overlapped with L. catesbeianus calls. However, contrary to our expectations, all of the species we studied altered the frequencies of their vocalizations in response to L. catesbeianus calls. Scinax perereca was the only species that responded

123

according to our expectations by increasing the dominant frequency and decreasing spectral overlap with L. catesbeianus. Previous studies have reported frequency shifts that reduce spectral overlap with the invasive species L. catesbeianus (Both and Grant 2012) and anthropogenic noise (Roca et al. 2016). Scinax perereca also increased its dominant frequency in response to calls of the native species R. icterica, suggesting that, at least in this species, this is a strategy to avoid acoustic masking by low frequency sounds by anuran vocalizations and not a specific response to an invasive species. Indeed, the ability to employ alternative spectral bands is known for Scinax perereca, which can alternate between low and high frequency calls or present the two bands of high energy (Magrini and Giaretta. 2001). Contrary to our expectations, all other species also responded to L. catesbeianus calls by changing the spectral parameters of their vocalizations. Even more unexpectedly, they did so by lowering their dominant frequencies during or after the exposure, thereby increasing spectral overlap with the exogenous sound.

Invasion of the acoustic niche by bullfrogs Fig. 7 Main effects (means and SE) of playback stimuli on advertisement call parameters of Dendropsophus minutus. Circles with solid lines indicate the Lithobates catesbeianus call effects, squares with dashed lines indicate the Rhinella icterica call effects, and triangles with dotted lines indicate the white-noise effects

Hypsiboas bischoffi, whose dominant frequency overlaps with the L. catesbeianus call frequency spectrum, lowered its dominant frequency in response to both L. catesbeianus and R. icterica calls. Hypsiboas leptolineatus exhibited the strongest response to the invasive calls among all the species we tested—even though its dominant frequency does not overlap with the L. catesbeianus call spectrum—decreasing the dominant frequency of both advertisement and aggressive calls in response to L. catesbeianus stimuli. This species also altered its dominant frequency in response to the R. icterica calls, but it did so by raising the frequency instead of lowering it. The dominant frequency of D. minutus is also higher than the L. catesbeianus call spectrum, yet this species also decreased its dominant frequency after the L. catesbeianus stimuli as well. However, in contrast to H. leptolineatus, D. minutus males also decreased the dominant frequency in response to R. icterica calls, as did H. bischoffi. These unexpected results show that responses to a new sound source are species- and stimulus-specific and can result in either increased or decreased spectral overlap.

A possible explanation for the unexpected decrease of dominant frequencies in response to heterospecific vocalizations regardless of the spectral overlap or anuran source (invasive or native) is that lower frequencies propagate better in the environment than higher frequencies (Marten et al. 1977; Marten and Marler 1977; Boonman and Kurniati 2011), regardless of the structure of habitat (Boonman and Kurniati 2011; Goutte et al. 2013; Llusia et al. 2013). Given that these species evolved to call synchronously with a temporally and spatially variable collection of heteroand conspecifics that exhibit differing degrees of spectral overlap (Gerhardt and Huber 2002), it is expected that strategies would evolve to improve call propagation. By slightly decreasing frequencies, they increase the probability of reaching a higher number of receivers. Moreover, communication in many anuran is multimodal (Ho¨dl and Ame´zquita 2001), so this strategy might serve to draw distant individuals closer where other kinds of short-range communication, like visual communication, can be employed. White noise also elicited significant frequency shifts in all species. Dendropsophus minutus, H. bischoffi,

123

C. I. Medeiros et al.

and H. leptolineatus all increased the dominant frequency of their calls in response to white noise, whereas S. perereca decreased its dominant frequency. The occurrence of frequency shifts for all species in our study suggests that this might be a common response to sounds introduced in the environment and not only to other anurans. Nevertheless, in all species responses were stronger to anuran calls than to white noise, which suggests that these species might have evolved a more acute awareness of anuran sounds and could explain the different directions of the spectral shift. Our results suggest that the primary impact of the invasion of a ‘‘noisy species’’ such as L. catesbeianus might not be masking due to spectral overlap with the vocalizations of native species, which would engender specific responses by affected species, but rather a generalized and highly variable response across the acoustic community. Changes in the acoustic niche of species can entail higher energy costs to the animals (Wells 1977) and compromise the fidelity and integrity of the signal for the receiver affecting mate selection and, consequently, fitness (Gerhardt 1991; Welch et al. 1998). Further, although no study has examined the physiological effects of invasive species calls to date, chronic exposure to anthropogenic noise can increase physiological stress and impede reproduction (Tennessen et al. 2014; Kaiser et al. 2015). Finally, in our experiments, we introduced only one calling male, and L. catesbeianus usually calls in a chorus of 4–8 males in some invaded Neotropical localities (Both and Grant 2012; Medeiros et al. 2016). In a real invasion scenario, calling L. catesbeianus males might exert stronger impacts on the native acoustic community. While many native species can modify their calls in response to new sounds, including human-caused noises, other species cannot (e.g. Lengagne 2008). Therefore, we can expect changes in the native acoustic structure with an introduction of any new source of sound. These changes may occur not only in a L. catesbeianus invasion event, but in any other animal groups that use acoustic communication for ecological interactions. Here, we demonstrated that the animals respond to the introduction of new sound sources, and that these responses are speciesand stimulus-specific. Thus, in agreement with studies of anthropogenic and stream noises (Dubois and Martens 1984; Feng et al. 2006), our study shows that invasive species that vocalize at low frequencies have the potential to impact the native acoustic community.

123

Acknowledgements We are grateful to Abner Pontelli Perez, Rafael Henrique, Dilson Peixoto, Patrı´cia Barcarolo, Friedrich Keppeler, Ro´gger Antunes, Andre´ Luza, Mariane Bosholn and Pedro Aure´lio Lima for their help in field and/or laboratory activities. We also thank the INSTITUTO CHICO MENDES ˜ O DA BIODIVERSIDADE (ICMBio) for DE CONSERVAC¸A authorization to conduct this research (No. 42411-1), the ˜O COORDENAC¸A DE APERFEIC¸OAMENTO DE PESSOAL DE NI´VEL SUPERIOR (CAPES) for the award of a scholarship to CIM, the CONSELHO NACIONAL DE ´ GICO DESENVOLVIMENTO CIENTI´FICO E TECNOLO (CNPq) for the research fellowships of SMH (process 304820/2014-8), TG (305234/2014-5) and CB (401076/20148), and finally the UNIVERSIDADE FEDERAL DO RIO GRANDE DO SUL (UFRGS) Postgraduate program in Ecology ` PESQUISA DO ˜ O DE AMPARO A and the FUNDAC¸A ˜ O PAULO (FAPESP) (2012/10000-5) for ESTADO DE SA financial support of the equipment and field activities.

References Anderson MJ (2001) Permutation tests for univariate or multivariate analysis of variance and regression. Can J Fish Aquat Sci 58:626–639 Bee MA, Swanson EM (2007) Auditory masking of anuran advertisement calls by road traffic noise. Anim Behav 74:1765–1776 Bioacoustics Research Program (2014) Raven Pro: interactive sound analysis software (Version 1.5) [Computer software]. Ithaca, NY: The Cornell Lab of Ornithology. Available from http://www.birds.cornell.edu/raven 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 Sociobiol 69:675–683 Boonman A, Kurniati H (2011) Evolution of high-frequency communication in frogs. Evolut Ecol Res 13:197–207 Both C, Grant T (2012) Biological invasions and the acoustic niche: the effect of bullfrog calls on the acoustic signals of white-banded tree frogs. Biol Lett 8:714–716 Both C, Sole´ M, Dos Santos TG, Cechin SZ (2009) The role of spatial and temporal descriptors for neotropical tadpole communities in southern Brazil. Hydrobiologia 624:125–138 Both C, Lingnau R, Santos-Jr AP, Lima LP, Madalozzo B, Grant T (2011) Widespread occurrence of the American Bullfrog, Lithobates catesbeianus (Shaw, 1802) (Anura: Ranidae), in Brazil. South Am J Herpetol 6:127–134 Bradbury JW, Vehrencamp SL (1998) Principles of animal communication. Sinauer Associates, Sunderland Brumm H, Voss K, Ko¨llmer I, Todt D (2004) Acoustic communication in noise: regulation of call characteristics in a New World monkey. J Exp Biol 207:443–448 Capranica RR (1968) The vocal repertoire of the bullfrog (Rana catesbeiana). Behaviour 31:302–324 Cardoso AJ, Haddad CFB (1984) Variabilidade acu´stica em diferentes populac¸o˜es e interac¸o˜es agressivas de Hyla minuta (Amphibia: Anura). Cien Cult 36:1393–1399

Invasion of the acoustic niche by bullfrogs Cruz CAG, Caramaschi U (1998) Definic¸a˜o, composic¸a˜o e distribuic¸a˜o geogra´fica do grupo de Hyla polytaenia Cope, 1870 (Amphibia, Anura, Hylidae). Bol do Museu Nacional Zool (NS) 392:1–19 Cunnington GM, Fahrig L (2010) Plasticity in the vocalizations of anurans in response to traffic noise. Acta Oecol 36:463–470 Dubois A, Martens J (1984) A case of possible vocal convergence between frogs and a bird in Himalayan torrents. J Ornithol 125:455–463 Edge WD, Marcum CL (1985) Movements of elk in relation to logging disturbances. J Wildl Manag 49:741–744 Farina A, Pieretti N, Morganti N (2013) Acoustic patterns of an invasive species: the Red-billed Leiothrix (Leiothrix lutea Scopoli 1786) in a Mediterranean shrubland. Bioacoustics. doi:10.1080/09524622.2012.761571 Feng AS, Narins PM, Xu C-H, Lin W-Y, Yu Z-L, Qiu Q, Xu ZM, Shen JX (2006) Ultrasonic communication in frogs. Nature 440:333–336 Fisher RA (1935) The logic of inductive inference. J R Stat Soc Ser A 98:39–54 Forsman A, Hagman M (2006) Calling is an honest indicator of paternal genetic quality in poison frogs. Evolution 60:2148–2157 Frost DR (2015) Amphibian species of the World: an Online Reference. Version 6.0. Electronic database accessible at http://research.amnh.org/herpetology/amphibia/index. html. American Museum of Natural History, New York. Accessed 26 Apr 2015 Gehara M, Crawford AJ, Orrico VGD, Rodrı´guez A, Lo¨tters S et al (2014) High levels of diversity uncovered in a widespread nominal taxon: continental phylogeography of the neotropical tree frog Dendropsophus minutus. PLoS ONE 9(9):e103958. doi:10.1371/journal.pone.0103958 Gerhardt H (1991) Female mate choice in treefrogs: static and dynamic acoustic criteria. Anim Behav 42:615–635 Gerhardt HC, Huber F (2002) Acoustic communication in insects and anurans: common problems and diverse solutions. The University of Chicago Press, Chicago Goutte S, Dubois A, Legendre F (2013) The importance of ambient sound level to characterise anuran habitat. PLoS ONE 8(10):e78020. doi:10.1371/journal.pone.0078020 Halfwerk W, Holleman LJM, Lessells CM, Slabbekoorn H (2011) Negative impact of traffic noise on avian reproductive success. J Appl Ecol 48:210–219 Hiert C, Moura MO (2007) Anfı´bios do Parque Municipal das Arauca´rias, Guarapuava- Parana´. Editora Unicentro, Guarapuava Ho¨dl W, Ame´zquita A (2001) Visual signaling in anuran amphibians. In: Ryan MJ (ed) Anuran communication. Smithsonian Institution Press, Washington, pp 121–141 Hollen LI, Radford AN (2009) The development of alarm call behavior in mammals and birds. Anim Behav 78:791–800 Jaquie´ry J, Broquet T, Aguilar C, Evanno G, Perrin N (2010) Good genes drive female choice for mating partners in the lek-breeding european treefrog. Evolution 64:108–115 Kaiser K, Hammers JL (2009) The effect of anthropogenic noise on male advertisement call rate in the neotropical treefrog, Dendropsophus triangulum. Behaviour 146:1053–1069 Kaiser K, Scofield DG, Alloush M, Jones RM, Marczak S, Martineau K, Oliva MA (2011) When sounds collide: the

effect of anthropogenic noise on a breeding assemblage of frogs in Belize, Central America. Behaviour 148:215–232 Kaiser K et al (2015) Effects of anthropogenic noise on endocrine and reproductive function in White’s treefrog, Litoria caerulea. Conserv Physiol 3(1):cou061 Krausman PR, Leopold BD, Scarbrough DL (1986) Desert mule deer response to aircraft. Wildl Soc Bull 14:68–70 Kruger DJD, Preez LHD (2016) The effect of airplane noise on frogs: a case study on the critically endangered Pickersgill’s reed frog (Hyperolius pickersgilli). Ecol Res 31:393–405 Kwet A (2001) Fro¨sche im brasilianischen Araukarienwald. Anurengemeinschaft des Araukarienwaldes von Rio Grande do Sul: Diversita¨t, Reproduktion und Ressourcenaufteilung. Mu¨nster. Natur und Tier-Verlag, 192 Kwet A, Di-Bernardo M (1999) Anfibios—Amphibien—Amphibians. EDIPUCRS, Porto Alegre Lane H, Tranel B (1971) The Lombard sign and the role of hearing in speech. J Speech Lang Hear Res 14:677–709 Lengagne T (2008) Traffic noise affects communication behaviour in a breeding anuran, Hyla arborea. Biol Conserv 141(8):2023–2031 Lever C (2003) Naturalized amphibians and reptiles of the world. Oxford University Press, New York Llusia D, Go´mez M, Penna M, Ma´rquez R (2013) Call Transmission efficiency in native and invasive anurans: competing hypotheses of divergence in acoustic signals. PLoS ONE 8(10):e77312. doi:10.1371/journal.pone.0077312 Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout MN, Bazzazz F (2000) Biotic invasions: causes, epidemiology, global consequences and control. Issues Ecol 5:1–20 Magrini L, Giaretta AA (2001) Calls of two Brazilian species of Scinax of the S. ruber clade (Anura: Hylidae). Herpetol Notes 3:121–126 Marcelino VR, Haddad CFB, Alexandrino J (2009) Geographic Distribution and Morphological Variation of Striped and Nonstriped Populations of the Brazilian Atlantic forest Treefrog Hypsiboas bischoffi (Anura: Hylidae). J Herpetol 43:351–361 Marten K, Marler P (1977) Sound transmission and its significance for animal vocalization. I. Temperate habitats. Behav Ecol Sociobiol 2:271–290 Marten K, Quine D, Marler P (1977) Sound transmission and its significance for animal vocalization. II. Tropical forest habitats. Behav Ecol Sociobiol 2:291–302 Martof BS, Thompson EF (1958) Reproductive behavior of the chorus frog (Pseudacris nigrita). Behaviour 13:243–258 McGregor PK, Dabelsteen T, Shepherd M, Pedersen SB (1992) The signal value of matched singing in Great Tits: evidence from in- teractive playback experiments. Anim Behav 43:987–998 Medeiros C, Both C, Kaefer I, Cechin SZ (2016) Reproductive phenology of the American Bullfrog in subtropical Brazil: photoperiod as a main determinant of seasonal activity. An Acad Bras Cieˆnc [Internet]. doi:10.1590/00013765201620150694 Narins PM (1995) Frog communication. Sci Am 273:78–83 Nogueira SSC, Pedroza JP, Nogueira-Filho SLG, Tokumaru RS (2012) The Function of Click Call Emission in Capybaras (Hydrochoerus Hydrochaeris). Ethology 118:1001– 1009

123

C. I. Medeiros et al. Parris KM, Velik-Lord M, North JMA (2009) Frogs call at a higher pitch in traffic noise. Ecol Soc 14(1):25 (online) Pillar VD (2006) MULTIV: multivariate exploratory analysis, randomization testing and bootstrap resampling. Universidade Federal do Rio Grande do Sul. See http://ecoqua. ecologia.ufrgs.br/ecoqua/MULTIV.html Pillar VD, Orlo´ci L (1996) On randomization testing in vegetation science: multifactor comparisons of releve´ groups. J Veg Sci 7:585–592 Planque R, Slabbekoorn H (2008) Spectral overlap in songs and temporal avoidance in a peruvian bird assemblage. Ethology 114:262–271 Pombal JP Jr (2010) O espac¸o acu´stico em uma taxocenose de anuros (Amphibia) do sudeste do brasil. Arquivos do Museu Nacional 68(1–2):135–144 Pombal JP Jr, Bastos RP, Haddad CFB (1995) Vocalizac¸o˜es de algumas espe´cies do geˆnero Scinax (Anura, idae) do Sudeste do Brasil e comenta´rios taxonoˆmicos. Naturalia 20:213–225 Primack BP, Rodrigues E (2001) Biologia da Conservac¸a˜o. Editora Planta, Londrina R Development Core Team (2012) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna Rabin LA, Mccowan B, Hooper SL, Owings DH (2003) Anthropogenic noise and its effect on animal communication: an interface between comparative psychology and conservation biology. Int J Comp Psychol 16:172–192 Reby D, Cargnelutti B, Joachim E, Aulagnier S (1999) Spectral acoustic structure of barking in roe deer (Capreolus capreolus). Sex-, age- and individual-related variations. Comptes Rendus Acad Sci Ser III Sci Vie 322:271–279 Rheindt FE (2003) The impact of roads on birds: does song frequency play a role in determining susceptibility to noise pollution? J Ornithol 144:295–306 Roca IT, Desrochers L, Giacomazzo M, Bertolo A, Bolduc P, Deschesnes R et al (2016) Shifting song frequencies in response to anthropogenic noise: a meta-analysis on birds and anurans. Behav Ecol 27:1269–1274 Ryan MJ (1988) Constraints and patterns in the evolution of anuran acoustic communication. In: Fritzsch B, Ryan MJ, Wilczynski W, Hetherington TE, Walkowiak W (eds) The

123

evolution of the amphibian auditory system. Wiley, New York, pp 637–677 Sax DF, Stachowicz JJ, Brown JH, Bruno JF, Dawson MN et al (2007) Ecological and evolutionary insights from species invasions. Trends Ecol Evol 22:465–471 Silvano D, Scott N, Aquino L, Kwet A, Baldo D (2010) Rhinella icterica. In: IUCN. Red list of threatened species. Version 2010.4. http://www.iucnredlist.org. Accessed 24 June 2015 Sinsch U, Lu¨mkemann K, Rosar K, Schwarz C, Dehling JM (2012) Acoustic niche partitioning in an anuran community inhabiting an Afromontane wetland (Butare, Rwanda). Afr Zool 47(1):60–73 Sueur J (2002) Cicada acoustic communication: potential sound partitioning in a multispecies community from Mexico (Hemiptera: Cicadomorpha: Cicadidae). Biol J Linn Soc 75:379–394 Sueur J, Aubin T, Simonis C (2008) Seewave: a free modular tool for sound analysis and synthesis. Bioacoustics 18:213–226 Tennessen JB, Parks SE, Langkilde T (2014) Traffic noise causes physiological stress and impairs breeding migration behaviour in frogs. Conserv Physiol. doi:10.1093/conphys/ cou032 Tennessen JB, Parks SE, Tennessen TP, Langkilde T (2016) Raising a racket: invasive species compete acoustically with native treefrogs. Anim Behav 114:53–61 Vargas-Salinas F, Cunnington GM, Ame´zquita A, Fahrig L (2014) Does traffic noise alter calling time in frogs and toads? A case study of anurans in Eastern Ontario, Canada. Urban Ecosyst 17:945–953 Welch AM, Semlitsch RD, Gerhardt HC (1998) Call duration as an indicator of genetic quality in male gray tree frogs. Science 280(5371):1928–1930 Wells KD (1977) The courtship of frogs. In: Taylor DH, Guttman SI (eds) The reproductive biology of amphibians. Plenum Press, New York, pp 233–262 Wells KD (2007) The ecology and behavior of amphibians. The University of Chicago Press, Chicago Zelick R, Narins PM (1985) Characterization of the advertisement call oscillator in the frog Elutherodactylus coqui. J Comp Physiol A 156:223–229