Skip to main content
NCBI home page
Ecology and Evolution logoLink to Ecology and Evolution
. 2020 Oct 6;10(21):12115–12128. doi: 10.1002/ece3.6833

Anuran accents: Continental‐scale citizen science data reveal spatial and temporal patterns of call variability

Savannah J Weaver 1,2,3,✉, Corey T Callaghan 3,4, Jodi J L Rowley 1,3
PMCID: PMC7663080  PMID: 33209274

Abstract

Many animals rely on vocal communication for mating advertisement, territorial displays, and warning calls. Advertisement calls are species‐specific, serve as a premating isolation mechanism, and reinforce species boundaries. Nevertheless, there is a great deal of interspecific variability of advertisement calls. Quantifying the variability of calls among individuals within a species and across species is critical to understand call evolution and species boundaries, and may build a foundation for further research in animal communication. However, collecting a large volume of advertisement call recordings across a large geographic area has traditionally posed a logistical barrier. We used data from the continental‐scale citizen science project FrogID to investigate the spatial and temporal patterns of call characteristics in six Australian frog species. We found intraspecific call variability in both call duration and peak frequency across species. Using resampling methods, we show that variability in call duration and peak frequency was related to the number of individuals recorded, the geographic area encompassed by those individuals, and the intra‐annual time difference between those recordings. We conclude that in order to accurately understand frog advertisement call variation, or “anuran accents,” the number of individuals in a sample must be numerous (N ≥ 20), encompass a large geographic area relative to a species' range, and be collected throughout a species' calling season.

Keywords: advertisement call, bioacoustics, frog, geographic variation, isolation by distance, temporal variation

Quantifying the variability in a species' advertisement vocalization is critical to understand call evolution and species boundaries, and may build a foundation for further research on vocalizations. We used data from the continental‐scale citizen science project FrogID to investigate the spatial and temporal patterns of call characteristics in six Australian frog species. We conclude that in order to accurately understand frog advertisement call variation, or “anuran accents,” the number of individuals in a sample must be numerous (N ≥ 20), encompass a large geographic area relative to the species’ range, and be collected throughout the species' calling season. Photo: Litoria chloris.

graphic file with name ECE3-10-12115-g006.jpg

1. INTRODUCTION

Vocal communication is widely used throughout the animal kingdom, including for mating advertisement, territorial displays, and warning calls (Brumm, 2013; Gerhardt & Huber, 2002). Advertisement vocalizations are made by males to attract females (Eriksson & Wallin, 1986; Ryan, 2001; Wells, 1977), and females use call characteristics like pitch, volume, and complexity to choose the best mate (Boul et al., 2007; Pedroso et al., 2013). Many taxa, namely birds (Ballentine, 2004; Clark et al., 2006), frogs (Ryan, 2001; Stebbins & Cohen, 1997; Wells, 2007), and insects (Gerhardt & Huber, 2002; Saarikettu et al., 2005), rely on vocalizations to find and secure mates for reproduction. Because these advertisement calls are species‐specific (Curé et al., 2012; Searcy et al., 1981) and serve as a premating isolation mechanism (Littlejohn, 1969), they reinforce species boundaries (Braune et al., 2008; Pedroso et al., 2013).

Advertisement vocalizations best serve their purpose by being unique to, and consistent within, a species. Yet we consistently find intraspecific variation across animals with species‐specific vocalizations. For example, nonhuman primates (Marler, 1973; Mitani et al., 1999), insects (Zuk et al., 2001), marine mammals (Cerchio et al., 2001; Rendell & Whitehead, 2005; Terhune et al., 2001), and birds (González & Ornelas, 2014; Haftorn & Hailman, 1997; Handford, 1988) all exhibit varying levels of intraspecific variation in vocal communication. Despite a breadth of research on numerous taxa, a clear ecological pattern for intraspecific variability of vocalizations remains enigmatic. Research has investigated spatial and temporal trends of vocalizations, but results from previous studies are inconsistent. For example, local variability of a species‐specific call can change over time (Adret‐Hausberger, 1986; Ince et al., 1980) or stay consistent without measurable temporal variation (Fournet et al., 2018; Whitney, 1992). There is also evidence that vocalization differences can be predicted by geographic distances (Marova et al., 2010; Röhr et al., 2020; Searfoss et al., 2020) or barriers (Jang et al., 2011; Zuk et al., 2001), but we do not know whether geographic isolation is a driver or product of different vocalizations.

Frogs are a useful taxonomic group to test ecological patterns in intraspecific variability because the advertisement call is a species‐specific trait (Oldham & Gerhardt, 1975), serves as a premating isolation mechanism (Boul et al., 2007; Capranica et al., 1973; Littlejohn, 1969), and is used by researchers to distinguish between and describe new species (Davies et al., 1986; Hoskin, 2007; Rowley et al., 2016; Sullivan et al., 1996). Additionally, the developmental biology of most frog species eliminates the possibility of learning as a confounding variable in studies of their vocalizations (Duellman & Trueb, 1994; Wells, 2007), allowing more confidence in a genetic basis for frog call characteristics (Welch et al., 2014), and a more direct connection to phylogeny (Bosch & De la Riva, 2004; Erdtmann & Amézquita, 2009; Ryan & Wilczynski, 1991).

Although species‐specific, frog advertisement calls do vary among individuals and populations, and even within individuals (Bee et al., 2001, 2010; Gambale et al., 2014; Gerhardt, 1991; Gerhardt & Huber, 2002; Hernández‐Herrera & Pérez‐Mendoza, 2020; Pettitt et al., 2013). Factors that correlate with population‐level variation of frog calls include habitat, discrete populations, and geographic isolation by distance (Jang et al., 2011; Littlejohn & Roberts, 1975; Ohmer et al., 2009; Rafiński & Babik, 2000; Rodríguez et al., 2010; Ryan & Wilczynski, 1991), although geographic patterns are not always found (Baraquet et al., 2015; Giacoma et al., 1997). Intra‐annual time difference (measured as difference in days within a species‐specific breeding season) may play a role, but few studies have tested this in frogs (see Gambale et al., 2014; Giacoma et al., 1997; Smith & Hunter, 2005). At the individual level, calls also vary in relation to body size and temperature (Blair, 1964; Kasuya & Shiobara, 1996; Rodríguez et al., 2010; Sullivan & Hinshaw, 1990). However, not all properties covary significantly or consistently, and residual variation remains after body size and temperature effects are accounted for (Castellano et al., 2000; Jang et al., 2011). Quantifying the variability of calls among individuals within a species and across species is critical to understand call evolution and species boundaries, as well as to inform methods for future research on frog vocalizations.

To date, studies of frog advertisement call variability have been limited in geographic, taxonomic, and temporal extent, largely as a result of the logistical challenges in collecting the quantity and quality of data required to investigate macroecological trends. Recently, however, the rise of citizen science projects has enabled data collection on a much greater temporal and spatial scale than ever before (Aceves‐Bueno et al., 2017; Callaghan et al., 2019; Lukyanenko et al., 2016; McKinley et al., 2017; Silvertown, 2009). Such datasets facilitate a breadth of ecological studies that would not be feasible otherwise (Diblíková et al., 2019; Mitchell et al., 2020; Searfoss et al., 2020).

We use continental‐scale citizen science data collected and submitted through the FrogID project to investigate spatial and temporal patterns of call characteristics in six Australian frog species. We hypothesized that within each species, call characteristics would vary across geographic area and breeding season, with both spatial and temporal variability following an isolation by distance/difference model. We also hypothesized that a larger geographic range or larger maximum body size would enable greater vocal variability within a species.

2. METHODS

2.1. FrogID

FrogID (www.frogid.net.au) is a citizen science project led by the Australian Museum where volunteers use their smartphones to record calling frogs. All submissions are validated by experts at the Australian Museum (Rowley & Callaghan, 2020; Rowley et al., 2019). To date, FrogID has received over 150,000 submissions, resulting in approximately 220,000 records of calling frogs, including 198 of Australia's 240 known frog species.

2.2. Study species

We selected Crinia insignifera, Crinia parinsignifera, Limnodynastes dorsalis, Limnodynastes peronii, Litoria chloris, and Litoria xanthomera to be the focus of our study (Figure 1). Each has a high number of FrogID submissions, distributed throughout their geographic range. These three congeneric pairs of species are phylogenetically closely related and have similar male advertisement calls, but are allopatric, and have different geographic range sizes.

FIGURE 1.

FIGURE 1

Open in a new tab

Study species, geographic range, and spectrograms of representative advertisement calls. Relative amplitudes over time are as follows: Crinia insignifera, 3 calls: ±500, 7 s; Crinia parinsignifera, 2 calls: ±500, 6 s; Limnodynastes dorsalis, 1 call: ±20, 2 s; Limnodynastes peronii, 2 calls: ±6, 8 s; Litoria chloris, 1 call, 11 notes: ±1, 8 s; Litoria xanthomera, 1 call, 14 notes: ±25, 15 s. Photographs by Jodi Rowley (C. parinsignifera, Lim. peronii, Lit. chloris, Lit. xanthomera) and Stephen Mahoney (C. insignifera, Lim. dorsalis)

2.3. Call selection and analysis

Increasingly, descriptive studies of frog calls are using smartphone recordings (Modak et al., 2016; Roh et al., 2014). The audio clarity and quality of FrogID recordings vary, resulting from differences in smartphones used, the distance between the observer and the calling frog, and the amount of background noise, including other calling frogs, in the recording (Rowley & Callaghan, 2020; Rowley et al., 2019). Although different phone models vary in their detection range and frequency response (Zilli, 2015), almost all smartphone models have a flat frequency response up to a threshold (Kardous & Shaw, 2014). Recordings made with the FrogID app are saved as MPEG AAC audio files, a form of audio compression that affects all frequencies uniformly (Brandenburg, 1999). While frequencies above 17 kHz are not represented in FrogID recordings (Rowley et al., 2019), all known Australian frogs have advertisement calls below 10kHz (Loftus‐Hills, 1973; Rowley et al., unpublished data). The dominant frequency of the advertisement calls of frog species in this study was all below 6 kHz.

We used FrogID recordings verified up to October 2019 and filtered for those in which only the target species was calling, increasing the likelihood that a recording would be of sufficient quality for analysis. We then filtered by locality to maintain spatial representation in those species with large numbers of FrogID recordings. For C. insignifera, C. parinsignifera, and Lim. dorsalis, we only used one recording from a given latitude and longitude combination to avoid unintentionally using several recordings of the same individual. For Lim. peronii, a species with a large geographic range and numerous FrogID recordings, we only used recordings collected at least 1 km apart from each other. We did not eliminate any recordings based on location for Lit. chloris and Lit. xanthomera because fewer FrogID recordings were available. This process of geographic filtering nullifies questions of temporal change at the same location or of the same individual, but we were more interested in among individual, rather than within individual, macroecological trends. After filtering by locality, we excluded recordings that were of insufficient quality (too many overlapping calls to pick out a single individual or too faint) or duration (not capturing enough calls of the individual). Table 1 lists number of recordings of each species before and after filtering.

TABLE 1.

Range size, maximum male snout–vent length (SVL), recording details, and call values for each species

Species Range size (km2) Maximum Male SVL (mm) Earliest recorded individual Latest recorded individual Recordings evaluated Recordings analyzed (%) Mean call duration ± SD (s) Mean peak frequency ± SD (Hz)
Crinia insignifera 16,853 25 4‐May‐18 30‐Aug‐19 226 123 (54) 0.22 ± 0.06 3,353 ± 437
Crinia parinsignifera 1,116,599 22 10‐Nov‐17 8‐Sep‐19 383 221 (58) 0.27 ± 0.13 3,746 ± 400
Limnodynastes dorsalis 311,818 64 24‐Nov‐17 14‐Sep‐19 181 82 (45) 0.16 ± 0.07 609 ± 130
Limnodynastes peronii 1,071,422 69 4‐Nov‐17 20‐Sep‐19 487 232 (48) 0.12 ± 0.07 1,033 ± 327
Litoria chloris 276,209 62 13‐Nov‐17 8‐Mar‐19 108 48 (44) 0.97 ± 0.19 1873 ± 190
Litoria xanthomera 14,196 56 21‐Nov‐17 15‐May‐19 102 56 (55) 0.75 ± 0.14 2,160 ± 198
Open in a new tab

We define a call as the entire assemblage of acoustic signals for a given vocalization and a note as an individual unit of sound, following definitions presented by Duellman and Trueb (1986). For C. insignifera, C. parinsignifera, Lim. dorsalis, and Lim. peronii, each call was a single note, and for Lit. chloris and Lit. xanthomera, each call consisted of several notes. For consistency across species, we analyzed the notes of the species in the genus Litoria as calls. FrogID recordings were converted into WAV format at a 48 kHz sampling rate and 16 bits per sample. We used Raven Sound Software (Pro Version 1.5, Cornell Lab of Ornithology) to visualize waveform and spectrogram for each recording. We set the spectrogram window size to 512 with 50% overlap and band‐filtered all recordings to reduce background noise. For one individual frog in each recording, we measured call duration and peak frequency, which are well‐established call characteristics to analyze (Gergus et al., 1997; Giacoma et al., 1997; Köhler et al., 2017; Littlejohn & Roberts, 1975; Mitchell et al., 2020; Penna & Veloso, 1990). Although we cannot ensure every recording is of a unique individual, it is likely the case, especially for the species which were filtered by locality.

For each recording, we selected 3–10 consecutive advertisement calls of a single individual, a sample size consistent with previous acoustic research (Bionda et al., 2008; Penna & Veloso, 1990). For Lit. chloris and Lit. xanthomera, we used all consecutive notes in a single call. For most individuals, we used the first 10 calls recorded to minimize bias. If quality was an issue, we selected the largest possible group of good quality consecutive notes. Quality was based on interference (i.e., wind, human activity, and insects), with preference given to the loudest individual with the most calls recorded. We only included advertisement calls in our analyses because it is the most commonly heard and most taxonomically informative (Köhler et al., 2017). Non‐advertisement calls are less frequent and vary with social context (Perrill & Bee, 1996; Ryan & Wilczynski, 1991; Sullivan & Wagner, 1988).

2.4. Statistical analysis

Each individual represents a unique recording location and date. Individual call duration and peak frequency were measured as means of all calls analyzed for each individual, as we were interested in variation among individuals of a species, rather than within individuals. We used resampling approaches to investigate how the intraspecific variability in call duration and peak frequency was influenced by the number of individuals, geographic area covered, and intra‐annual time difference of individuals in a sample of a population.

To test the effect of sample size (i.e., the number of individuals measured as a representative sample of the population), we drew random samples of N = 2, 3, 4, … 48 individuals. We used 48 as the maximum number of individuals in a sample because that was the lowest number of individuals we measured for a single species (Lit. chloris; Table 1). For each sample size, N, we resampled individuals 1,000 times with replacement and calculated the standard deviations of call duration and peak frequency for each sample. We drew qualitative conclusions about the resulting pattern using visual representation (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

As the number of individuals measured increases, deviation of measured variance from actual variance decreases. Shapes and line types denote genera: Crinia as squares with solid lines, Limnodynastes as circles with dashed lines, and Litoria as triangles with dotted lines. Colors distinguish species within each genus: green versus blue.

To test the effect of geographic area and intra‐annual time difference on call variability, we randomly sampled 20 calls with replacement 1,000 times for each species. We calculated the following parameters for each random sample: (a) standard deviation of call duration; (b) standard deviation of peak frequency; (c) geographic area, measured as the convex hull area based on each individual's recording location; and (d) intra‐annual time difference between calls, measured as number of days between the first and last individuals recorded within a breeding season. For species in the genus Litoria that have breeding seasons spanning across the calendar year, we relativized the time of year based on the first observations of the breeding season recorded. We set N = 20 because it roughly corresponded to the number of individuals necessary for a sample to achieve reliable estimates of variation for a species (Figure 3), and it was approximately half the number of individuals measured for the species with the least individuals (Lit. chloris; Table 1), allowing for sufficient probability of resampling. We tested other sample sizes (N = 10, 30, 40), but found no qualitative differences in our key results, thus only present results for N = 20.

We used intraspecific z‐scores of each variable to fit individual models for each species (N = 6) and to fit a model encompassing the variability across our study species, for four unique comparisons: (a) call duration as a function of intra‐annual time difference; (b) call duration as a function of geographic area; (c) peak frequency as a function of intra‐annual time difference; and (d) peak frequency as a function of geographic area. First, for each comparison, we fit a linear mixed model with species as a random effect to test for the overall influence of the response variable on the predictor variable across species. Second, we fit linear models for each of the six species (i.e., 24 models). Lastly, we fit another linear model to investigate whether the model estimates (i.e., correlation coefficients) extracted from the models for each relationship for each species were associated with range size or body size. Range sizes were extracted from FrogID distribution maps (www.frogid.net.au), which were originally calculated based on species records in the Atlas of Living Australia (www.ala.org.au) and modified by expert opinion. Maximum male body sizes, measured in mm as snout–vent length (SVL), were taken from Anstis (2017).

Temporal aspects of calls often vary with temperature (Bionda et al., 2008; Gambale et al., 2014; Rodríguez et al., 2010; Sullivan & Hinshaw, 1990). Therefore, we tested a generalized linear model for individual call variables across species as a function of temperature, using temperature estimates based on methods from Mitchell et al. (2020). We found that call duration was nonsignificantly positively correlated with temperature (GLM, estimate = .0031, SE = .001752, t = 1.773, p‐value =.077). Typically, temperature shows a direct correlation with temporal call properties such as call duration because as ectotherms, frogs' vocal speed ability is determined by ambient air temperature (Ryan, 2001). However, we found no such correlation. Conversely, peak frequency was significantly negatively correlated (GLM, estimate = −61, SE = 8.52, t = −7.16, p‐value < .0001), even though spectral properties are not usually correlated with temperature, and usually rely more on body size. We acknowledge that temperature effects exist at the individual level, but these models suggest that including temperature in our analyses of call variability would be uninformative. Based on the nonsignificant and counterintuitive trends we found, we chose not to include temperature in our statistical analyses. Further, our resampling approach samples many individuals (N = 1,000) at many potential temperatures, so temperature effects are unlikely to influence our macroecological results. Our goal was to encompass the potential range of environmental conditions, rather than to eliminate them. Humidity may also affect body condition generally, but would only affect vocalizations indirectly (Köhler et al., 2017), so we do not consider humidity.

3. RESULTS

Of the 1,487 individuals with recordings that met our filtering criteria, 762 were deemed of sufficient quality for inclusion in our analyses (Table 1). Call duration and peak frequency measurements of individuals clustered by species, but there was a great deal of variation within each intraspecific cluster (Figure 2; Table 1). We found that as the number of individuals in a sample increased, the sample more accurately represented the variability in call characteristics present in the population. Deviation from true population variability decreased as the number of individuals measured increased (Figure 3). Deviation appears to be greatest and decrease the most between 0 and 20 individuals sampled, with a less drastic decrease after including approximately 20 individuals, and this pattern was similar among all six species (Figure 3).

FIGURE 2.

FIGURE 2

Open in a new tab

Call duration and peak frequency, measured as the mean value for each individual analyzed, clustered by genus and species. Shapes denote genera: Crinia as squares, Limnodynastes as circles, and Litoria as triangles. Colors distinguish species within each genus: green versus blue

We found strong correlations of the variability of call duration and peak frequency with geographic area encompassed by and the number of days between individuals in a sample (Figure 4). Across species (i.e., using species as a random effect), the variability of call duration was positively correlated with geographic area and intra‐annual time difference. This relationship was statistically significant for geographic area (GLMM, estimate = .0418, t = 3.247, df = 5,998, p‐value = .001), but nonsignificant for intra‐annual time difference (GLMM, estimate = .0018, t = .142, df = 5,998, p‐value = .887). The variability of peak frequency was positively correlated with and statistically significant for both geographic area (GLMM, estimate = .0822, t = 6.39, df = 5,998, p‐value < .0001) and difference in days (GLMM, estimate = .0501, t = 3.88, df = 5,998, p‐value = .0001)

FIGURE 4.

FIGURE 4

Open in a new tab

Vocal variability correlates with (a) intra‐annual time difference and (b) geographic area. Shaded area around each linear model represents standard error, which varies based on number of individuals within a given section of the model. Colors distinguish models for call duration (green) versus peak frequency (blue)

When species were considered separately, the trends in variability of call duration and peak frequency were not the same across species (Table 2, Figures 4 and 5). Variability in call duration increased as geographic area increased for all species except C. parinsignifera. Variability in call duration increased as intra‐annual time difference increased for C. insignifera, Lit. chloris, and Lit. xanthomera, but decreased for C. parinsignifera, Lim. dorsalis, and Lim. peronii. Variability in peak frequency increased as the geographic area increased for C. parinsignifera, Lim. peronii, Lit. chloris, and Lit. xanthomera, but decreased for Lim. dorsalis, and showed no clear trend for C. insignifera. Variability in peak frequency in relation to intra‐annual time difference appeared to differ among genera, but not within species pairs: It increased as intra‐annual time difference increased for species in the Crinia and Litoria genera, but decreased for species in the Limnodynastes genus.

TABLE 2.

Summary statistics for species‐specific generalized linear models

Species Predictor variable Dependent variable Estimate t‐statistic df p‐value
Crinia insignifera Area Call duration −0.01050 −0.332 2 .7403
Peak frequency 0.01813 0.573 2 .5668
Days Call duration 0.08177 2.592 2 .0097
Peak frequency 0.05816 1.840 2 .0660
Crinia parinsignifera Area Call duration −0.04339 −1.372 2 .1704
Peak frequency 0.05662 1.792 2 .0735
Days Call duration −0.05686 −1.799 2 .0723
Peak frequency 0.13137 4.186 2 <.0001
Limnodynastes dorsalis Area Call duration 0.07446 2.359 2 .0185
Peak frequency −0.04959 −1.568 2 .1171
Days Call duration −0.03186 −1.007 2 .3142
Peak frequency −0.05238 −1.657 2 .0978
Limnodynastes peronii Area Call duration 0.10425 3.311 2 .0010
Peak frequency 0.04311 1.363 2 .1731
Days Call duration −0.06749 −2.137 2 .0328
Peak frequency −0.07535 −2.387 2 .0172
Litoria chloris Area Call duration 0.03968 1.254 2 .2100
Peak frequency 0.34275 11.526 2 <.0001
Days Call duration 0.03333 1.053 2 .2924
Peak frequency 0.13560 4.324 2 <.0001
Litoria xanthomera Area Call duration 0.08680 2.753 2 .0060
Peak frequency 0.08237 2.611 2 .0092
Days Call duration 0.05209 1.648 2 .0997
Peak frequency 0.10302 3.272 2 .0011
Open in a new tab

Area refers to geographic area (km2) of the convex hull encompassed by individuals in a sample. Days refers to the intra‐annual time difference (days) between individuals in a sample.

FIGURE 5.

FIGURE 5

Open in a new tab

Correlation coefficients for call duration and peak frequency as a function of intra‐annual time difference and geographic area. Error bars represent 95% confidence intervals. Colors distinguish model estimates for call duration (green) versus peak frequency (blue)

A species' range size did not correlate with the relationships of call duration as a function of geographic area (GLM, estimate = −1.5e‐8, t = −.264, df = 5, p‐value = .805), peak frequency as a function of geographic area (GLM, estimate = −3.2e‐8, t = −.242, df = 5, p‐value = .821), or peak frequency as a function of intra‐annual time difference (GLM, estimate = −4.1e‐8, t = −.45, df = 5, p‐value = .676), but it did significantly correlate with the relationship of call duration as a function of intra‐annual time difference (GLM, estimate = −1.1e‐7, t = −4.05, df = 5, p‐value = .015). The maximum male SVL for each species correlated with the relationship of call duration as a function of geographic area (GLM, estimate = .0026, t = 4.85, df = 5, p‐value = .0083). A species' maximum SVL did not correlate with the relationships of peak frequency as a function of geographic area (GLM, estimate = .0013, t = .417, df = 5, p‐value = .698), call duration as a function of intra‐annual time difference (GLM, estimate = −6.8e‐4, t = −.468, df = 5, p‐value = .664), or peak frequency as a function of intra‐annual time difference (GLM, estimate = −.0023, t = −1.203, df = 5, p‐value = .295).

4. DISCUSSION

We leveraged a unique continental‐scale citizen science dataset to quantify the variability in call duration and peak frequency across broad spatial and temporal scales. Although only half of all FrogID recordings we examined were used in our study, the dataset gave us a remarkable degree of spatial and temporal representation. Among only the calls we analyzed, recordings were distributed throughout most of each species' geographic range (Figure S1) and spanned the duration of the species' breeding seasons over the 2 year study period (Table 1). FrogID recordings which did not meet our quality standards still remain important biodiversity records and may be suitable for other bioacoustic studies.

For six frog species across two families and three genera, we found high intraspecific variability of call duration and peak frequency (Figure 2). Our results highlight the importance of considering the number of individuals analyzed when examining vocalizations: a sample size of less than 20 individuals is unlikely to capture the true variability in a species' advertisement call (Figure 3). Male advertisement calls are often used to delineate and describe species as new to science, but the number of individuals analyzed can be relatively small, sometimes only 1–5 individuals (Matsui, 1997; Pombal et al., 1995; Roberto et al., 2017). We recommend that researchers analyze advertisement calls from at least 20 individuals per clade or taxon (Martins & Jim, 2004). Although our recommendation is based on only six species, the geographic and taxonomic scope of our study was relatively broad. We encourage future studies to test whether this trend holds true in other regions and taxonomic groups. We acknowledge that 20 or more recordings may be logistically challenging, especially for rarely detected species in remote locations, but future call descriptions should strive to sample 20 individuals, captured over time and space, to truly describe a species' vocalization.

Call duration and peak frequency both varied across time and space for all six species. We found that variability of both call duration and peak frequency was positively correlated with both geographic area encompassed by locations of individuals and intra‐annual time difference between individuals in a sample. When we considered each species separately, this trend was relatively consistent.

As the intra‐annual time difference between individuals in a sample increased, the vocalizations of those individuals became more variable (Figures 4 and 5). This correlation could be related to changes in individuals throughout the breeding season, such as age, body size, hormone fluctuations, or vocal maturation as the breeding season progresses (Ryan, 2001; Wells, 2007). While some studies have investigated temporal patterns, it is unclear whether call variability does (Smith & Hunter, 2005) or does not (Gambale et al., 2014; Giacoma et al., 1997) have temporal trends. Regardless of the reason for intra‐annual variability in call characteristics, our findings highlight the importance of considering time of year when measuring frog vocalizations.

As the geographic area among individuals increased, their vocalizations also became more variable (Figures 4 and 5). These results suggest that vocalizations follow an isolation by distance model, which has also been suggested by previous studies of advertisement calls (Marova et al., 2010; Rafiński & Babik, 2000; Ryan et al., 1996). Generally, there is strong evidence for spatial, population‐based differences in call characteristics (Baraquet et al., 2015; Capranica et al., 1973; Hernández‐Herrera & Pérez‐Mendoza, 2020; Jang et al., 2011; Littlejohn & Roberts, 1975; Ohmer et al., 2009; Rodríguez et al., 2010). In addition to the phenotypic gradient represented across an isolation by distance model, there are geographic barriers to gene flow like mountain ranges, ocean divides, and cities, which separate groups and may lead to vocal divergence (Jang et al., 2011; Klymus et al., 2012). Our findings emphasize the need to measure vocalizations across a species' entire geographic range to encompass its call variability.

While variability in both call duration and peak frequency most often had positive correlations with time and space, this relationship was not true for every species we analyzed. In fact, several relationships had strong negative correlations (e.g., C. parinsignifera, Lim. dorsalis, and Lim. peronii (Figure 4a); C. parinsignifera and Lim. dorsalis (Figure 4b)). We are unsure what may drive the negative correlations we observed. One potential driver is character displacement (Brown & Wilson, 1956; Jang & Gerhardt, 2006), so that within a species, individuals from the same population alter their calls when overlapping in time or space to distinguish themselves from their neighbors when competing to attract a mate. This pattern of sympatric character displacement has been observed between bird species (Kirschel et al., 2009; Wallin, 1986), but has yet to be studied in frog species. One study suggests that C. parinsignifera and Lim. dorsalis have “tuned hearing” that only picks up the specific frequency of conspecific calls because these species are often found in multi‐species choruses (Loftus‐Hills & Johnstone, 1970). Tuned hearing may explain the negative trends we observed because although small vocal variation may occur in these species, over more space or time, their calls would be constrained by their hearing, resulting in less variation in broader spatial or temporal scales than at smaller scales. Alternatively, breeding season is unlikely to play a role, as C. insignifera and Lim. dorsalis breed in the winter, Lit. chloris, Lit. xanthomera, and Lim. peronii breed in the summer, and C. parinsignifera breeds year‐round (Barker et al., 1995). Similarly, call structure was unlikely to be responsible, as it was relatively consistent within each genus studied. Further, we are unsure of the role of plasticity, which could potentially contribute to either positive or negative correlations observed (Price et al., 2003). Roles of these and other factors should be discerned by future research to determine the mechanisms resulting in the observed patterns.

Each species has a unique life history, so each species is likely affected differently and to varying degrees by the ecological and evolutionary drivers of vocal variation. We investigated species' range size as a potential reason for the interspecific differences across correlations, but found little evidence that range size influenced the relationships of either call characteristics as a function of space or time. While we did find evidence suggesting that a larger range size could lead to more variable call duration over time, overall, our findings suggest that a species' range size does not determine the vocal variation possible across its range or throughout time. We also considered whether body size had a relationship with advertisement call variability. Studies show that vocalizations vary among individuals and among species based on body size (Blair, 1964; McClelland et al., 1996; Rodríguez et al., 2010). It was previously unknown whether a larger maximum body size enables more vocal variability within a species, and we found a correlation between maximum male SVL of a species and the relationship of call duration as a function of geographic area. However, most body size correlations were weak and nonsignificant. Other factors likely to influence the species‐specific relationships observed include ecological considerations such as weather, habitat type, elevation, or anthropogenic effects, and evolutionary constraints such as call complexity, vocal repertoire, and morphology (Ryan, 2001; Wells, 2007). Every species is likely to respond to these factors differently, which leads to great potential vocal diversity within some species, as well as potential species divergence. The distinction between intraspecific diversity and speciation is yet to be determined, although we are confident in the species delineations for those analyzed in this study.

Ultimately, our findings reveal strong temporal and spatial patterns of frog vocalizations. However, several limitations of this study should be improved upon in future studies. Due to our macroecological approach and reliance on citizen science recordings, we cannot ensure every recording is a unique individual. To increase this likelihood, we filtered recordings by location. These methods produced meaningful results for our investigation of variability among individuals, but we suggest additional studies also investigate call variation within individuals of a specific locale over time. We were also unable to rule out whether temperature or body size plays a role in the patterns we present due to their influence at the individual level (Cocroft & Ryan, 1995; McClelland et al., 1996; Wilczynski et al., 1993). Frogs also modify their calls depending on social context (Hernández‐Herrera & Pérez‐Mendoza, 2020), but we were unable to incorporate information about whether frogs were calling solitarily or as part of a chorus. Rather, we filtered the recordings we used to those with only the target species calling and minimum background noise, which eliminated most chorus calling from our analyses. Future studies could incorporate this variable, but should be aware of the trade‐offs between citizen science data volume and scientific specificity. For example, to address social context using our dataset would have required post hoc determination of singularity versus chorus and estimations of chorus size based on audio interpretation. Finally, we measure intra‐annual time difference as the maximum range of days among individuals in a sample. This measure currently does not account for the distribution of individuals in the random sample in regard to the breeding season (i.e., more individuals could be sampled closer to one end of the calling season than the other end), potentially limiting our analyses related to intra‐annual time difference. Future work should focus on testing our observed patterns in other species, both within related genera and across various lineages. Testing the patterns we present at several phylogenetic levels would help to determine whether phylogeny or environment has the greatest influence on call variability (Bosch & De la Riva, 2004; Erdtmann & Amézquita, 2009; Welch et al., 2014). Lastly, we used two key vocal characteristics (i.e., call duration and peak frequency), but other bioacoustic variables may be valuable. Our approach could be generalized to include these additional call characteristics, as well as more species, in different habitats, and from other continents.

While many studies have tested spatial patterns of vocal variability (Baraquet et al., 2015; Capranica et al., 1973; Jang et al., 2011; Littlejohn & Roberts, 1975; Rodríguez et al., 2010; Ryan et al., 1996), few have tested temporal patterns (Gambale et al., 2014; Giacoma et al., 1997; Smith & Hunter, 2005). Our results highlight the value of using citizen science data to assess the patterns of acoustic or morphological variability at scales previously not possible. We clearly highlight the inherent variability in advertisement calls, which should be accounted for in future bioacoustic studies. Comparisons of frog calls and descriptions of new species that only measure a few individuals from a single locale at a single point in time likely fail to properly capture the variability that exists within a species' vocalization. In order to accurately understand anuran accents, the number of individuals in a sample must be numerous (N ≥ 20), encompass a large geographic area relative to the species range, and be collected throughout its calling season. Citizen science will continue to play a role enabling such studies, and coupled with targeted fieldwork, could supply ecologists with increasingly robust datasets to find and explore nuances in similar macroecological patterns.

CONFLICT OF INTERESTS

The authors report no conflicts of interest.

AUTHOR CONTRIBUTION

Savannah J. Weaver: Data curation (equal); Formal analysis (equal); Visualization (equal); Writing‐original draft (equal); Writing‐review & editing (equal). Corey T. Callaghan: Conceptualization (equal); Formal analysis (equal); Methodology (equal); Visualization (equal); Writing‐original draft (equal); Writing‐review & editing (equal). Jodi J. L. Rowley: Conceptualization (equal); Visualization (equal); Writing‐original draft (equal); Writing‐review & editing (equal).

OPEN RESEARCH BADGES

This article has been awarded Open Data and Open Materials Badges. All materials and data are publicly accessible via the Open Science Framework at https://github.com/science-with-sav/anuran_accents_data_code.

Supporting information

Fig S1: Geographic range for each species (shaded area) and location of each FrogID recording (black circles) used in our analyses. Horizontal axis is longitude. Vertical axis is latitude

Click here for additional data file. (5.3MB, png)

ACKNOWLEDGMENTS

We thank the Bucknell University Presidential Fellowship Office for the student funding to endeavor this project. We thank the Citizen Science Grants of the Australian Government for providing funding for the FrogID project, the Impact Grants program of IBM Australia for providing the resources to build the FrogID App, and Bunnings and Fyna Foods for supporting FrogID. We thank the thousands of citizen scientists who contribute to FrogID records, and the FrogID team, who validates submissions, thus creating the robust dataset which made this study possible.

Weaver SJ, Callaghan CT, Rowley JJL. Anuran accents: Continental‐scale citizen science data reveal spatial and temporal patterns of call variability. Ecol. Evol. 2020;10:12115–12128. 10.1002/ece3.6833

[Correction Statement: Correction added on 16 October 2020 after first online publication: The link to the Open Science Framework has been updated in this version]

DATA AVAILABILITY STATEMENT

Figures and analyses were completed in R version 3.6.3 (R Core Team, 2020) and relied heavily on the tidyverse workflow (Wickham, 2017). Data and annotated code are archived in Zenodo (https://doi.org/10.5281/zenodo.4016911).

REFERENCES

  1. Aceves‐Bueno, E. , Adeleye, A. S. , Feraud, M. , Huang, Y. , Tao, M. , Yang, Y. , & Anderson, S. E. (2017). The accuracy of citizen science data: A quantitative review. Bulletin of the Ecological Society of America, 98(4), 278–290. 10.1002/bes2.1336 [DOI] [Google Scholar]
  2. Adret‐Hausberger, M. (1986). Temporal dynamics of dialects in the whistled songs of Starlings. Ethology, 71(2), 140–152. 10.1111/j.1439-0310.1986.tb00579.x [DOI] [Google Scholar]
  3. Anstis, M. (2017). Tadpoles and frogs of Australia. New Holland Publishers Pty Ltd.. [Google Scholar]
  4. Ballentine, B. (2004). Vocal performance influences female response to male bird song: An experimental test. Behavioral Ecology, 15(1), 163–168. 10.1093/beheco/arg090 [DOI] [Google Scholar]
  5. Baraquet, M. , Grenat, P. R. , Salas, N. E. , & Martino, A. L. (2015). Geographic variation in the advertisement call of Hypsiboas cordobae (Anura, Hylidae). Acta Ethologica, 18(1), 79–86. 10.1007/s10211-014-0188-2 [DOI] [Google Scholar]
  6. Barker, J. , Grigg, G. C. , & Tyler, M. J. (1995). A Field Guide to Australian Frogs. Surrey Beatty and Sons. [Google Scholar]
  7. Bee, M. A. , Cook, J. M. , Love, E. K. , O'Bryan, L. R. , Pettitt, B. A. , Schrode, K. , & Vélez, A. (2010). Assessing acoustic signal variability and the potential for sexual selection and social recognition in Boreal Chorus Frogs (Pseudacris maculata). Ethology, 116(6), 564–576. [Google Scholar]
  8. Bee, M. A. , Kozich, C. E. , Blackwell, K. J. , & Gerhardt, H. C. (2001). Individual variation in advertisement calls of territorial male green frogs, Rana clamitans: Implications for individual discrimination. Ethology, 107(1), 65–84. 10.1046/j.1439-0310.2001.00640.x [DOI] [Google Scholar]
  9. Bionda, C. D. L. , Salas, N. , & Di Tada, I. E. (2008). Effect of temperature on the advertisement call of Physalaemus bilingonigerus (Anura: Leptodactylidae). Biological Association of Herpetology, 2008(19), 19–22. [Google Scholar]
  10. Blair, W. F. (1964). Isolating mechanisms and interspecies interactions in anuran amphibians. The Quarterly Review of Biology, 39(4), 334–344. 10.1086/404324 [DOI] [PubMed] [Google Scholar]
  11. Bosch, J. , & De la Riva, I. (2004). Are frog calls modulated by the environment? An analysis with anuran species from Bolivia. Canadian Journal of Zoology, 82(6), 880–888. 10.1139/z04-060 [DOI] [Google Scholar]
  12. Boul, K. E. , Chris Funk, W. , Darst, C. R. , Cannatella, D. C. , & Ryan, M. J. (2007). Sexual selection drives speciation in an Amazonian frog. Proceedings of the Royal Society B: Biological Sciences, 274(1608), 399–406. 10.1098/rspb.2006.3736 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Brandenburg, K. (1999). MP3 and AAC Explained. Presented at the Audio Engineering Society Conference: 17th International Conference: High‐Quality Audio Coding. [Google Scholar]
  14. Braune, P. , Schmidt, S. , & Zimmermann, E. (2008). Acoustic divergence in the communication of cryptic species of nocturnal primates (Microcebus ssp.). BMC Biology, 6(19), 1–10. 10.1186/1741-7007-6-19 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Brown, W. L. , & Wilson, E. O. (1956). Character displacement. Systematic Zoology, 5(2), 49–64. 10.2307/2411924 [DOI] [Google Scholar]
  16. Brumm, H. (Ed.). (2013). Animal communication and noise (Vol. 2). Springer. [Google Scholar]
  17. Callaghan, C. T. , Rowley, J. J. L. , Cornwell, W. K. , Poore, A. G. B. , & Major, R. E. (2019). Improving big citizen science data: Moving beyond haphazard sampling. PLOS Biology, 17(6), e3000357 10.1371/journal.pbio.3000357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Capranica, R. R. , Frishkopf, L. S. , & Nevo, E. (1973). Encoding of geographic dialects in the auditory system of the cricket frog. Science, 182(4118), 1272–1275. [DOI] [PubMed] [Google Scholar]
  19. Castellano, S. , Giacoma, C. , & Dujsebayeva, T. (2000). Morphometric and advertisement call geographic variation in polyploid green toads. Biological Journal of the Linnean Society, 70(2), 341–360. 10.1111/j.1095-8312.2000.tb00214.x [DOI] [Google Scholar]
  20. 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(2), 313–329. 10.1006/anbe.2001.1747 [DOI] [Google Scholar]
  21. Clark, J. A. , Boersma, P. D. , & Olmsted, D. M. (2006). Name that tune: Call discrimination and individual recognition in Magellanic penguins . Animal Behaviour, 72(5), 1141–1148. 10.1016/j.anbehav.2006.04.002 [DOI] [Google Scholar]
  22. Cocroft, R. B. , & Ryan, M. J. (1995). Patterns of advertisement call evolution in toads and chorus frogs. Animal Behaviour, 49(2), 283–303. 10.1006/anbe.1995.0043 [DOI] [Google Scholar]
  23. Curé, C. , Mathevon, N. , Mundry, R. , & Aubin, T. (2012). Acoustic cues used for species recognition can differ between sexes and sibling species: Evidence in shearwaters. Animal Behaviour, 84(1), 239–250. 10.1016/j.anbehav.2012.04.039 [DOI] [Google Scholar]
  24. Davies, M. , McDonald, K. R. , & Adams, M. (1986). A new species of green tree frog (Anura: Hylidae) from Queensland, Australia. Proceedings of the Royal Society of Victoria, 98(2), 63–71. [Google Scholar]
  25. Diblíková, L. , Pipek, P. , Petrusek, A. , Svoboda, J. , Bílková, J. , Vermouzek, Z. , Procházka, P. , & Petrusková, T. (2019). Detailed large‐scale mapping of geographical variation of Yellowhammer (Emberiza citrinella) song dialects in a citizen science project. Ibis, 161(2), 401–414. 10.1111/ibi.12621 [DOI] [Google Scholar]
  26. Duellman, W. E. , & Trueb, L. (1986). Biology of amphibians. McGraw‐Hill Book Company. [Google Scholar]
  27. Duellman, W. E. , & Trueb, L. (1994). Biology of amphibians. Johns Hopkins University Press. [Google Scholar]
  28. Erdtmann, L. , & Amézquita, A. (2009). Differential evolution of advertisement call traits in dart‐poison frogs (Anura: Dendrobatidae). Ethology, 115(9), 801–811. 10.1111/j.1439-0310.2009.01673.x [DOI] [Google Scholar]
  29. Eriksson, D. , & Wallin, L. (1986). Male bird song attracts females: A field experiment. Behavioral Ecology and Sociobiology, 19(4), 297–299. 10.1007/BF00300645 [DOI] [Google Scholar]
  30. Fournet, M. E. H. , Gabriele, C. M. , Culp, D. C. , Sharpe, F. , Mellinger, D. K. , & Klinck, H. (2018). Some things never change: Multi‐decadal stability in humpback whale calling repertoire on Southeast Alaskan foraging grounds. Scientific Reports, 8(1), 1–13. 10.1038/s41598-018-31527-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Gambale, P. , Signorelli, L. , & Bastos, R. (2014). Individual variation in the advertisement calls of a Neotropical treefrog (Scinax constrictus). Amphibia‐Reptilia, 35, 271–281. 10.1163/15685381-00002949 [DOI] [Google Scholar]
  32. Gergus, E. W. A. , Sullivan, B. K. , & Malmos, K. B. (1997). Call variation in the Bufo microscaphus complex: Implications for species boundaries and the evolution of mate recognition. Ethology, 103(12), 979–989. 10.1111/j.1439-0310.1997.tb00140.x [DOI] [Google Scholar]
  33. Gerhardt, H. C. (1991). Female mate choice in treefrogs: Static and dynamic acoustic criteria. Animal Behaviour, 42(4), 615–635. 10.1016/S0003-3472(05)80245-3 [DOI] [Google Scholar]
  34. Gerhardt, H. C. , & Huber, F. (2002). Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. University of Chicago Press. [Google Scholar]
  35. Giacoma, C. , Zugolaro, C. , & Beani, L. (1997). The advertisement calls of the green toad (Bufo viridis): Variability and role in mate choice. Herpetologica, 53(4), 454–464. [Google Scholar]
  36. González, C. , & Ornelas, J. F. (2014). Acoustic divergence with gene flow in a lekking hummingbird with complex songs. PLoS One, 9(10), e109241 10.1371/journal.pone.0109241 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Haftorn, S. , & Hailman, J. P. (1997). Do the Siberian tits Parus cinctus in Scandinavia and Siberia speak the same language? Bioacoustics, 8(3–4), 223–247. 10.1080/09524622.1997.9753364 [DOI] [Google Scholar]
  38. Handford, P. (1988). Trill rate dialects in the Rufous‐collared Sparrow, Zonotrichia capensis, in northwestern Argentina. Canadian Journal of Zoology, 66(12), 2658–2670. 10.1139/z88-391 [DOI] [Google Scholar]
  39. Hernández‐Herrera, C. I. , & Pérez‐Mendoza, H. A. (2020). Acoustic and morphological variation on two populations of Dryophytes arenicolor in central México. Bioacoustics, 1–12, 10.1080/09524622.2020.1760937 [DOI] [Google Scholar]
  40. Hoskin, C. J. (2007). Description, biology and conservation of a new species of Australian tree frog (Amphibia: Anura: Hylidae: Litoria) and an assessment of the remaining populations of Litoria genimaculata Horst, 1883: Systematic and conservation implications of an unusual speciation event. Biological Journal of the Linnean Society, 91(4), 549–563. 10.1111/j.1095-8312.2007.00805.x [DOI] [Google Scholar]
  41. Ince, S. A. , Slater, P. J. B. , & Weismann, C. (1980). Changes with time in the songs of a population of Chaffinches. The Condor, 82(3), 285–290. 10.2307/1367393 [DOI] [Google Scholar]
  42. Jang, Y. , & Gerhardt, H. C. (2006). Divergence in the calling songs between sympatric and allopatric populations of the southern wood cricket Gryllus fultoni (Orthoptera: Gryllidae). Journal of Evolutionary Biology, 19(2), 459–472. 10.1111/j.1420-9101.2005.01014.x [DOI] [PubMed] [Google Scholar]
  43. Jang, Y. , Hahm, E. H. , Lee, H.‐J. , Park, S. , Won, Y.‐J. , & Choe, J. C. (2011). Geographic variation in advertisement calls in a tree frog species: Gene flow and selection hypotheses. PLoS One, 6(8), e23297 10.1371/journal.pone.0023297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Kardous, C. A. , & Shaw, P. B. (2014). Evaluation of smartphone sound measurement applications. The Journal of the Acoustical Society of America, 135(4), EL186–EL192. 10.1121/1.4865269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Kasuya, E. , & Shiobara, S. (1996). Variation in the advertisement call in the foam‐nesting Treefrog Rhacophorus arboreus . Bioacoustics, 7(1), 1–11. 10.1080/09524622.1996.9753311 [DOI] [Google Scholar]
  46. Kirschel, A. N. G. , Blumstein, D. T. , & Smith, T. B. (2009). Character displacement of song and morphology in African tinkerbirds. Proceedings of the National Academy of Sciences of the United States of America, 106(20), 8256–8261. 10.1073/pnas.0810124106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Klymus, K. E. , Humfeld, S. C. , & Gerhardt, H. C. (2012). Geographical variation in male advertisement calls and female preference of the wide‐ranging canyon treefrog, Hyla arenicolor . Biological Journal of the Linnean Society, 107(1), 219–232. 10.1111/j.1095-8312.2012.01931.x [DOI] [Google Scholar]
  48. Köhler, J. , Jansen, M. , Rodríguez, A. , Kok, P. J. R. , Toledo, L. F. , Emmrich, M. , Glaw, F. , Haddad, C. F. B. , Rödel, M.‐O. , & Vences, M. (2017). The use of bioacoustics in anuran taxonomy: Theory, terminology, methods and recommendations for best practice. Zootaxa, 4251(1), 1–124. 10.11646/zootaxa.4251.1.1 [DOI] [PubMed] [Google Scholar]
  49. Littlejohn, M. J. (1969). The systematic significance of isolating mechanisms In Systematic Biology: Proceedings of an International Conference (pp. 459–482). National Academies Press. [Google Scholar]
  50. Littlejohn, M. J. , & Roberts, J. D. (1975). Acoustic analysis of an intergrade zone between two call races of the Limnodynastes tasmaniensis complex (Anura: Leptodactylidae) in South‐eastern Australia. Australian Journal of Zoology, 23, 113–122. 10.1071/ZO9750113 [DOI] [Google Scholar]
  51. Loftus‐Hills, J. J. (1973). Comparative aspects of auditory function in Australian anurans. Australian Journal of Zoology, 21(3), 353–367. 10.1071/zo9730353 [DOI] [Google Scholar]
  52. Loftus‐Hills, J. J. , & Johnstone, B. M. (1970). Auditory function, communication, and the brain‐evoked response in anuran amphibians. Journal of the Acoustical Society of America, 47(4B), 1131–1138. 10.1121/1.1912015 [DOI] [PubMed] [Google Scholar]
  53. Lukyanenko, R. , Parsons, J. , & Wiersma, Y. F. (2016). Emerging problems of data quality in citizen science. Conservation Biology, 30(3), 447–449. 10.1111/cobi.12706 [DOI] [PubMed] [Google Scholar]
  54. Marler, P. (1973). A comparison of vocalizations of red‐tailed monkeys and Blue Monkeys, Cercopithecus ascanius and C. mitis, in Uganda. Zeitschrift Für Tierpsychologie, 33(3–4), 223–247. 10.1111/j.1439-0310.1973.tb02093.x [DOI] [PubMed] [Google Scholar]
  55. Marova, I. M. , Ivanitskii, V. V. , & Veprintseva, O. D. (2010). Individual, population, and geographic differentiation in advertising song of the Blyth's reed warbler, Acrocephalus dumetorum (Sylvidae). Biology Bulletin, 37(8), 846–860. 10.1134/S106235901008008X [DOI] [Google Scholar]
  56. Martins, I. A. , & Jim, J. (2004). Advertisement call of Hyla jimi and Hyla elianeae (Anura, Hylidae) in the Botucatu Region, São Paulo, Brazil. Brazilian Journal of Biology, 64(3B), 645–654. 10.1590/S1519-69842004000400012 [DOI] [PubMed] [Google Scholar]
  57. Matsui, M. (1997). Call characteristics of Malaysian Leptolalax with a description of two new species (Anura: Pelobatidae). Copeia, 1997(1), 158–165. 10.2307/1447851 [DOI] [Google Scholar]
  58. McClelland, B. E. , Wilczynski, W. , & Ryan, M. J. (1996). Correlations between call characteristics and morphology in male cricket frogs (Acris crepitans). Journal of Experimental Biology, 199(9), 1907–1919. [DOI] [PubMed] [Google Scholar]
  59. McKinley, D. C. , Miller‐Rushing, A. J. , Ballard, H. L. , Bonney, R. , Brown, H. , Cook‐Patton, S. C. , Evans, D. M. , French, R. A. , Parrish, J. K. , Phillips, T. B. , Ryan, S. F. , Shanley, L. A. , Shirk, J. L. , Stepenuck, K. F. , Weltzin, J. F. , Wiggins, A. , Boyle, O. D. , Briggs, R. D. , Chapin, S. F. , & Soukup, M. A. (2017). Citizen science can improve conservation science, natural resource management, and environmental protection. Biological Conservation, 208, 15–28. 10.1016/j.biocon.2016.05.015 [DOI] [Google Scholar]
  60. Mitani, J. C. , Hunley, K. L. , & Murdoch, M. E. (1999). Geographic variation in the calls of wild chimpanzees: A reassessment. American Journal of Primatology, 47(2), 133–151. [DOI] [PubMed] [Google Scholar]
  61. Mitchell, B. A. , Callaghan, C. T. , & Rowley, J. J. L. (2020). Continental‐scale citizen science data reveal no changes in acoustic responses of a widespread tree frog to an urbanisation gradient. Journal of Urban Ecology, 6(1), 1–12. 10.1093/jue/juaa002 [DOI] [Google Scholar]
  62. Modak, N. , Dahanukar, N. , Ogale, H. , & Padhye, A. (2016). Advertisement calls of Amboli leaping frog Indirana chiravasi (Anura: Ranixalidae) from northern Western Ghats, India. Current Science, 110(12), 2220–2223. [Google Scholar]
  63. Ohmer, M. E. , Robertson, J. M. , & Zamudio, K. R. (2009). Discordance in body size, colour pattern, and advertisement call across genetically distinct populations in a Neotropical anuran (Dendropsophus ebraccatus). Biological Journal of the Linnean Society, 97(2), 298–313. 10.1111/j.1095-8312.2009.01210.x [DOI] [Google Scholar]
  64. Oldham, R. S. , & Gerhardt, H. C. (1975). Behavioral isolating mechanisms of the treefrogs Hyla cinerea and H. gratiosa . Copeia, 1975(2), 223–231. 10.2307/1442880 [DOI] [Google Scholar]
  65. Pedroso, S. S. , Barber, I. , Svensson, O. , Fonseca, P. J. , & Amorim, M. C. P. (2013). Courtship sounds advertise species identity and male quality in sympatric Pomatoschistus spp. Gobies. PLoS One, 8(6), e64620 10.1371/journal.pone.0064620 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Penna, M. , & Veloso, A. (1990). Vocal diversity in frogs of the South American temperate forest. Journal of Herpetology, 24(1), 23–33. 10.2307/1564285 [DOI] [Google Scholar]
  67. Perrill, S. A. , & Bee, M. A. (1996). Responses to conspecific advertisement calls in the green frog (Rana Clamitans) and their role in male‐male communication. Behaviour, 133(3–4), 283–301. 10.1163/156853996X00152 [DOI] [Google Scholar]
  68. Pettitt, B. A. , Bourne, G. R. , & Bee, M. A. (2013). Advertisement call variation in the golden rocket frog (Anomaloglossus beebei): Evidence for individual distinctiveness. Ethology, 119(3), 244–256. [Google Scholar]
  69. Pombal, J. P. , Haddad, C. F. B. , & Kasahara, S. (1995). A new species of Scinax (Anura: Hylidae) from southeastern Brazil, with comments on the genus. Journal of Herpetology, 29(1), 1–6. 10.2307/1565078 [DOI] [Google Scholar]
  70. Price, T. D. , Qvarnström, A. , & Irwin, D. E. (2003). The role of phenotypic plasticity in driving genetic evolution. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270(1523), 1433–1440. 10.1098/rspb.2003.2372 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Rafiński, J. , & Babik, W. (2000). Genetic differentiation among northern and southern populations of the moor frog Rana arvalis Nilsson in central Europe. Heredity, 84(5), 610–618. 10.1046/j.1365-2540.2000.00707.x [DOI] [PubMed] [Google Scholar]
  72. R Core Team (2020). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; https://www.R‐project.org/ [Google Scholar]
  73. Rendell, L. , & Whitehead, H. (2005). Spatial and temporal variation in sperm whale coda vocalizations: Stable usage and local dialects. Animal Behaviour, 70(1), 191–198. 10.1016/j.anbehav.2005.03.001 [DOI] [Google Scholar]
  74. Roberto, I. J. , Araujo‐Vieira, K. , de Carvalho‐e‐Silva, S. P. , & Ávila, R. W. (2017). A new species of Sphaenorhynchus (Anura: Hylidae) from Northeastern Brazil. Herpetologica, 73(2), 148–161. 10.1655/HERPETOLOGICA-D-16-00021 [DOI] [Google Scholar]
  75. Rodríguez, A. , de la Nuez, D. , & Alonso, R. (2010). Intraspecific variation in the advertisement call of the cloud‐forest frog Eleutherodactylus glamyrus (Anura: Eleutherodactylidae). Journal of Herpetology, 44(3), 457–466. 10.1670/09-038.1 [DOI] [Google Scholar]
  76. Roh, G. , Borzée, A. , & Jang, Y. (2014). Spatiotemporal distributions and habitat characteristics of the endangered treefrog, Hyla suweonensis, in relation to sympatric H. japonica . Ecological Informatics, 24, 78–84. 10.1016/j.ecoinf.2014.07.009 [DOI] [Google Scholar]
  77. Röhr, D. L. , Camurugi, F. , Paterno, G. B. , Gehara, M. , Juncá, F. A. , Álvares, G. F. R. , Brandão, R. A. , & Garda, A. A. (2020). Variability in anuran advertisement call: A multi‐level study with 15 species of monkey tree frogs (Anura, Phyllomedusidae). Canadian Journal of Zoology, 98(8), 495–504. 10.1139/cjz-2020-0018 [DOI] [Google Scholar]
  78. Rowley, J. , & Callaghan, C. T. (2020). The FrogID dataset: Expert‐validated occurrence records of Australia's frogs collected by citizen scientists. ZooKeys, 912, 139–151. 10.3897/zookeys.912.38253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Rowley, J. J. L. , Callaghan, C. , & Cutajar, T. (2019). FrogID: Citizen scientists provide validated biodiversity data on frogs of Australia. Herpetological Conservation and Biology, 14(1), 155–170. [Google Scholar]
  80. Rowley, J. J. L. , Tran, D. T. A. , Le, D. T. T. , Dau, V. Q. , Peloso, P. L. V. , Nguyen, T. Q. , Hoang, H. D. , Nguyen, T. T. , & Ziegler, T. (2016). Five new, microendemic Asian leaf‐litter frogs (Leptolalax) from the southern Annamite mountains, Vietnam. Zootaxa, 4085(1), 63–102. [DOI] [PubMed] [Google Scholar]
  81. Ryan, M. J. (2001). Anuran communication. Smithsonian Institution Press. [Google Scholar]
  82. Ryan, M. J. , Rand, A. S. , & Weigt, L. A. (1996). Allozyme and advertisement call variation in the Túngara frog, Physalaemus pustulosus . Evolution, 50(6), 2435–2453. 10.1111/j.1558-5646.1996.tb03630.x [DOI] [PubMed] [Google Scholar]
  83. Ryan, M. J. , & Wilczynski, W. (1991). Evolution of intraspecific variation in the advertisement call of a cricket frog (Acris crepitans, Hylidae). Biological Journal of the Linnean Society, 44(3), 249–271. 10.1111/j.1095-8312.1991.tb00619.x [DOI] [Google Scholar]
  84. Saarikettu, M. , Liimatainen, J. O. , & Hoikkala, A. (2005). The role of male courtship song in species recognition in Drosophila montana . Behavior Genetics, 35(3), 257–263. 10.1007/s10519-005-3218-z [DOI] [PubMed] [Google Scholar]
  85. Searcy, W. A. , Marler, P. , & Peters, S. S. (1981). Species song discrimination in adult female song and swamp sparrows. Animal Behaviour, 29(4), 997–1003. 10.1016/S0003-3472(81)80053-X [DOI] [Google Scholar]
  86. Searfoss, A. M. , Liu, W. , & Creanza, N. (2020). Geographically well‐distributed citizen science data reveals range‐wide variation in the chipping sparrow's simple song. Animal Behaviour, 161, 63–76. 10.1016/j.anbehav.2019.12.012 [DOI] [Google Scholar]
  87. Silvertown, J. (2009). A new dawn for citizen science. Trends in Ecology & Evolution, 24(9), 467–471. 10.1016/j.tree.2009.03.017 [DOI] [PubMed] [Google Scholar]
  88. Smith, M. J. , & Hunter, D. (2005). Temporal and geographic variation in the advertisement call of the Booroolong Frog (Litoria booroolongensis: Anura: Hylidae). Ethology, 111(12), 1103–1115. 10.1111/j.1439-0310.2005.01101.x [DOI] [Google Scholar]
  89. Stebbins, R. , & Cohen, N. (1997). Voice In A natural history of amphibians (pp. 76‐88). Princeton University Press. [Google Scholar]
  90. Sullivan, B. K. , & Hinshaw, S. H. (1990). Variation in advertisement calls and male calling behavior in the spring peeper (Pseudacris crucifer). Copeia, 1990(4), 1146–1150. 10.2307/1446500 [DOI] [Google Scholar]
  91. Sullivan, B. K. , Malmos, K. B. , & Given, M. F. (1996). Systematics of the Bufo woodhousii complex (Anura: Bufonidae): Advertisement call variation. Copeia, 1996(2), 274–280. 10.2307/1446843 [DOI] [Google Scholar]
  92. Sullivan, B. K. , & Wagner, W. E. (1988). Variation in advertisement and release calls, and social influences on calling behavior in the Gulf Coast toad (Bufo valliceps). Copeia, 1988(4), 1014–1020. 10.2307/1445726 [DOI] [Google Scholar]
  93. Terhune, J. M. , Healey, S. R. , & Burton, H. R. (2001). Easily measured call attributes can detect vocal differences between Weddell Seals from two areas. Bioacoustics, 11(3), 211–222. 10.1080/09524622.2001.9753463 [DOI] [Google Scholar]
  94. Wallin, L. (1986). Divergent character displacement in the song of two allospecies: The Pied Flycatcher Ficedula hypoleuca, and the Collared Flycatcher Ficedula albicollis . Ibis, 128(2), 251–259. 10.1111/j.1474-919X.1986.tb02672.x [DOI] [Google Scholar]
  95. Welch, A. M. , Smith, M. J. , & Gerhardt, H. C. (2014). A multivariate analysis of genetic variation in the advertisement call of the gray treefrog, Hyla versicolor . Evolution, 68(6), 1629–1639. [DOI] [PubMed] [Google Scholar]
  96. Wells, K. D. (1977). The courtship of frogs In Taylor D. H., & Guttman S. I. (Eds.), The reproductive biology of amphibians (pp. 233–262). Springer. [Google Scholar]
  97. Wells, K. D. (2007). The ecology & behavior of amphibians. University of Chicago Press. [Google Scholar]
  98. Whitney, C. L. (1992). Temporal stability of song in a local population of Wood Thrushes. Wilson Bulletin, 104(3), 516–520. [Google Scholar]
  99. Wickham, H. (2017). tidyverse: Easily Install and Load the ‘Tidyverse’: Vol. R package version 1.2.1 Retrieved from https://style.tidyverse.org/
  100. Wilczynski, W. , McClelland, B. E. , & Rand, A. S. (1993). Acoustic, auditory, and morphological divergence in three species of neotropical frog. Journal of Comparative Physiology A, 172(4), 425–438. 10.1007/BF00213524 [DOI] [PubMed] [Google Scholar]
  101. Zilli, D. (2015). Smartphone‐powered citizen science for bioacoustic monitoring (Phd, University of Southampton). Retrieved from https://eprints.soton.ac.uk/382943/ [Google Scholar]
  102. Zuk, M. , Rotenberry, J. T. , & Simmons, L. W. (2001). Geographical variation in calling song of the field cricket Teleogryllus oceanicus: The importance of spatial scale. Journal of Evolutionary Biology, 14(5), 731–741. 10.1046/j.1420-9101.2001.00329.x [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

  1. Wickham, H. (2017). tidyverse: Easily Install and Load the ‘Tidyverse’: Vol. R package version 1.2.1 Retrieved from https://style.tidyverse.org/

Supplementary Materials

Fig S1: Geographic range for each species (shaded area) and location of each FrogID recording (black circles) used in our analyses. Horizontal axis is longitude. Vertical axis is latitude

Click here for additional data file. (5.3MB, png)

Data Availability Statement

Figures and analyses were completed in R version 3.6.3 (R Core Team, 2020) and relied heavily on the tidyverse workflow (Wickham, 2017). Data and annotated code are archived in Zenodo (https://doi.org/10.5281/zenodo.4016911).

Articles from Ecology and Evolution are provided here courtesy of Wiley

RESOURCES