Abstract

Sexual signals can be important mechanisms of reproductive isolation among incipient species, but the relative effectiveness and geographic distribution of different signals may vary, particularly when those signals have different modes of inheritance (e.g., cultural vs. genetic). Indeed, novel phenotypes may develop when traits with different modes of inheritance or different selection pressures become spatially decoupled from one another in zones of secondary contact, and such novel phenotypes may facilitate or hinder reproductive isolation. We assessed divergence of a culturally transmitted signal (song) and conspecific responses to that signal across a zone of reduced gene flow where plumage characters show asymmetrical introgression among subspecies of the red-backed fairy-wren (Malurus melanocephalus). We found that, in contrast to plumage, geographic variation in many song characteristics correlated well with geographic genetic structure. Additionally, playback experiments revealed that birds responded most strongly to songs from within the same genetic region irrespective of plumage type. These results indicate that song has been resistant to introgression, yet song traits and preferences have not hindered introgression of plumage traits. We suggest that this is because different modes of inheritance (here, cultural vs. genetic) facilitate decoupling of song and plumage, resulting in a large region where males exhibit a novel combination of sexual traits.

INTRODUCTION

Signals used for communication among conspecifics, particularly signals involved in mate attraction or territory acquisition and maintenance, are important mechanisms of reproductive isolation among incipient species because such signals may facilitate assortative mating (Coyne and Orr 2004; Edwards et al. 2005; Price 2008b). Genetically heritable signals, such as color patches or innate vocalizations, may readily facilitate speciation, particularly when postmating isolation is strong (Price 1998; Slabbekoorn and Smith 2002; Seddon and Tobias 2007; Kirschel et al. 2009; Uy et al. 2009). Signals that are inherited culturally, notably learned avian vocalizations, may also facilitate reproductive isolation and may diverge rapidly across populations (e.g., Baker and Mewaldt 1978; Sorenson et al. 2003; Patten et al. 2004; Grant and Grant, 2009). Variation in learned songs typically mirrors species and subspecies boundaries when reproductive isolation is complete (Price 2008b), but it is less clear when and how learned songs function in reproductive isolation when closely related populations come together and hybridize because learned songs and recognition may either converge (e.g., if interspecific interactions are beneficial) or diverge (e.g., if interspecific interactions are detrimental; Irwin and Price 1999; Haavie et al. 2004; Sattler et al. 2007; Dingle et al. 2010; Secondi et al. 2011). Evaluating how songs and conspecific responses vary across zones of reduced gene flow affords an opportunity to examine such convergence/divergence and also to assess the effectiveness of learned signals as isolating mechanisms.

Songs are just one type of sexual signal that might lead to reproductive isolation. Accordingly, it is important to consider the entire phenotype in which traits (or suites of traits) occur because some traits may facilitate reproductive isolation, whereas others may not, thereby leading to complex and dynamic patterns of isolation and introgression (Lein and Corbin 1990; Patten et al. 2004; Price 2008a; Uy et al. 2009). When traits exhibit differential introgression in zones of secondary contact, one potential outcome will be regions with novel phenotypes composed of unique combinations of parental traits (Parsons et al. 1993; Stein and Uy 2006; Baker 2010; den Hartog et al. 2010). These novel phenotypes may have significant evolutionary consequences (Arnold 1992; Grant et al. 1996; Barton 2001; Price 2008a) because they decouple introgressing (e.g., uniformly advantageous) traits from traits that may otherwise slow introgression (e.g., locally adapted traits).

Considerable research has focused on differential introgression of genetically heritable traits, which require selection to become decoupled from each other (Parsons et al. 1993; Dowling and Secor 1997; Stein and Uy 2006; Price 2008a; den Hartog et al. 2010; Sánchez-Guillén et al. 2011). For example, asymmetry in mating success appears to drive introgression of yellow plumage from golden-collared manakins (Manacus vitellinus) into white-collared manakin males (Manacus candei; Stein and Uy 2006), and asymmetry in dispersal and aggressive behavior appears to drive unidirectional expansion of Townsend’s warbler (Setophaga townsendi) traits into regions with Hermit warbler (Setophaga occidentalis) genetic background (Pearson and Rohwer 2000; Krosby and Rohwer 2009). However, differential introgression of genetic and cultural traits, such as plumage and learned avian song, may be particularly likely to lead to novel combinations of traits because they may be readily decoupled (Lein and Corbin 1990; Feldman and Laland 1996; Baker 2010). For example, birds may learn songs postdispersal (Lein and Corbin 1990; Baptista and Gaunt 1997) or from heterospecifics (Helb et al. 1985;,Grant and Grant 1997; Baker and Boylan 1999; Secondi et al. 2003; Haavie et al. 2004; Qvarnström et al. 2006), or females may disperse across song boundaries and carry with them genes but not learned male songs (Lein and Corbin 1990). It remains unclear how learned songs evolve in systems where unidirectional introgression of other traits has occurred, but songs may hinder introgression of other traits, introgress alongside preferred traits, or evolve independently. We can discriminate among these alternatives by investigating the spatial distribution of songs and conspecific responses to those songs in areas showing unidirectional introgression of genetic traits.

We examined divergence of vocal signals and conspecific responses to those signals across a contact zone where there appears to be a large region of plumage introgression between 2 subspecies of red-backed fairy-wrens, Malurus melanocephalus melanocephalus and Malurus melanocephalus cruentatus. These subspecies are genetically distinct and likely diverged during periods of range restriction and reduced gene flow during the Pleistocene (Lee and Edwards 2008). The subspecies are currently distributed along the northern and eastern coasts of Australia with a broad zone of contact in Northern Queensland (Figure 1). Phenotypically, they are distinguished primarily on the basis of male nuptial plumage color; males in the west have deep red back plumage, whereas males in the east have brighter, orange back plumage (Rowley and Russell 1997). However, there is a significant discrepancy between the geographic pattern of genetic variation and that of plumage traits; nuclear and mtDNA analyses indicate that past restricted gene flow across the Carpentarian barrier allowed for genetic divergence between the 2 subspecies (Lee and Edwards 2008), but currently plumage traits show the greatest divergence at a location much farther east (Rowley and Russell 1997) where there is significant gene flow (Figure 1). A parsimonious explanation for this pattern is that plumage initially diverged across the Carpentarian barrier and has subsequently introgressed from red cruentatus (western) populations to orange melanocephalus (eastern) populations (Baldassarre et al., in review). This introgression of plumage traits may be driven by sexual selection (see Pearson and Rohwer 2000; Stein and Uy 2006), as plumage color is important for mating and dominance interactions in this species (Karubian, 2002; Karubian et al. 2009), and sexual selection via extra-pair mating is common (Webster et al. 2008, 2010).

Figure 1

Sampling sites for recordings and playbacks (numbers correspond to Table 1). Black square indicates pivot point used for calculation of distances between sites (see section “Methods”). Example spectrograms are shown from 2 sites, and time scales for both are equivalent (song length = ~6 s for cruentatus, ~3.5 s for melanocephalus; see Supplementary Figure 1 for additional spectrograms). Dark gray region indicates approximate range of genetic cruentatus subspecies (all with red plumage), intermediate gray indicates genetic melanocephalus subspecies with red plumage, and light gray indicates genetic melanocephalus subspecies with orange plumage. Dotted line indicating the location of greatest genetic divergence (between sites 4 and 5) based on nuclear and mtDNA analysis of Lee and Edwards (2008), and dotted line indicating the location of greatest plumage divergence (between sites 12 and 13) based on Rowley and Russell (1997).

Figure 1

Sampling sites for recordings and playbacks (numbers correspond to Table 1). Black square indicates pivot point used for calculation of distances between sites (see section “Methods”). Example spectrograms are shown from 2 sites, and time scales for both are equivalent (song length = ~6 s for cruentatus, ~3.5 s for melanocephalus; see Supplementary Figure 1 for additional spectrograms). Dark gray region indicates approximate range of genetic cruentatus subspecies (all with red plumage), intermediate gray indicates genetic melanocephalus subspecies with red plumage, and light gray indicates genetic melanocephalus subspecies with orange plumage. Dotted line indicating the location of greatest genetic divergence (between sites 4 and 5) based on nuclear and mtDNA analysis of Lee and Edwards (2008), and dotted line indicating the location of greatest plumage divergence (between sites 12 and 13) based on Rowley and Russell (1997).

Table 1

Sampling locations for all recordings and playbacks in this study

Site Subspecies Plumage Location name GPS coordinates n (males) Song PC1 (mean ± standard deviation) Level 
cruentatus Red Broome, WA 17°49′S, 122°12′E -1.82±0.57 
2* cruentatus Red Mornington Sanctuary, WA 17°31′S, 126°06′E 23 -3.54±0.52 
3* cruentatus Red Camooweal, QLD 19°58′S, 138°06′E 13 -1.99±0.66 
4* cruentatus Red Mount Isa, QLD 20°37′S, 139°30′E 12 -2.02±0.47 
melanocephalus Red Croyden, QLD 18°12′S, 142°15′E -0.69±0.25 
melanocephalus Red Grorgetown, QLD 18°15′S, 143°09′E 0.68±0.27 
7* melanocephalus Red Mount Surprise, QLD 18°08′S, 144°18′E 1.39±0.47 
8* melanocephalus Red Mount Molloy, QLD 16°38′S, 145°20′E 1.66±0.45 
melanocephalus Red Moomin, QLD 17°22′S, 145°25′E 1.82±0.38 
10* melanocephalus Red Herberton, QLD 17°26′S, 145°22′E 1.53±0.16 
11 melanocephalus Red Ravenshoe, QLD 17°38′S, 145°27′E 1.38±0.51 
12 melanocephalus Red Mutarnee, QLD 18°55′S, 146°18′E 3.23±0.49 
13* melanocephalus Orange Mount Ossa, QLD 20°56′S, 148°50′E 14 2.36±0.4 
14 melanocephalus Orange Taunton, QLD 23°33′S, 149°13′E 12 1.79±0.4 
15 melanocephalus Orange Samsonvale, QLD 27°16′S, 152°51′E 13 1.77±0.33 
16 melanocephalus Orange Mullumbimby, NSW 28°34′S, 153°25′E 1.89 f, e 
Site Subspecies Plumage Location name GPS coordinates n (males) Song PC1 (mean ± standard deviation) Level 
cruentatus Red Broome, WA 17°49′S, 122°12′E -1.82±0.57 
2* cruentatus Red Mornington Sanctuary, WA 17°31′S, 126°06′E 23 -3.54±0.52 
3* cruentatus Red Camooweal, QLD 19°58′S, 138°06′E 13 -1.99±0.66 
4* cruentatus Red Mount Isa, QLD 20°37′S, 139°30′E 12 -2.02±0.47 
melanocephalus Red Croyden, QLD 18°12′S, 142°15′E -0.69±0.25 
melanocephalus Red Grorgetown, QLD 18°15′S, 143°09′E 0.68±0.27 
7* melanocephalus Red Mount Surprise, QLD 18°08′S, 144°18′E 1.39±0.47 
8* melanocephalus Red Mount Molloy, QLD 16°38′S, 145°20′E 1.66±0.45 
melanocephalus Red Moomin, QLD 17°22′S, 145°25′E 1.82±0.38 
10* melanocephalus Red Herberton, QLD 17°26′S, 145°22′E 1.53±0.16 
11 melanocephalus Red Ravenshoe, QLD 17°38′S, 145°27′E 1.38±0.51 
12 melanocephalus Red Mutarnee, QLD 18°55′S, 146°18′E 3.23±0.49 
13* melanocephalus Orange Mount Ossa, QLD 20°56′S, 148°50′E 14 2.36±0.4 
14 melanocephalus Orange Taunton, QLD 23°33′S, 149°13′E 12 1.79±0.4 
15 melanocephalus Orange Samsonvale, QLD 27°16′S, 152°51′E 13 1.77±0.33 
16 melanocephalus Orange Mullumbimby, NSW 28°34′S, 153°25′E 1.89 f, e 

Site numbers correspond to Figure 1. Subspecies identity based on Lee and Edwards (2008), plumage color based on Rowley and Russell (1997). Means and standard deviations of song PC1 (details of PC1 in Table 2) are given; levels with a different letter indicate presence of a significant difference in PC1 between sampling sites (2-tailed Student’s t-test). Asterisks indicate sites at which playbacks were conducted.

Table 2

Correlations of song variables with the first 3 principle components of a principle component analysis

 PC 1 PC 2 PC 3 
Eigenvalue 5.37 1.25 0.79 
Percent variation 59.62 13.94 8.77 
Intro note intervala 0.34 -0.07 0.46 
Song length -0.28 0.18 0.43 
Note length -0.42 0.05 0.09 
Peak frequencyb 0.38 0.15 0.16 
High frequency 0.22 0.67 0.16 
Low frequency 0.30 -0.20 0.51 
Note slopec 0.31 0.43 -0.24 
Note bandwidth -0.33 0.52 0.02 
Note rate 0.36 -0.04 -0.47 
 PC 1 PC 2 PC 3 
Eigenvalue 5.37 1.25 0.79 
Percent variation 59.62 13.94 8.77 
Intro note intervala 0.34 -0.07 0.46 
Song length -0.28 0.18 0.43 
Note length -0.42 0.05 0.09 
Peak frequencyb 0.38 0.15 0.16 
High frequency 0.22 0.67 0.16 
Low frequency 0.30 -0.20 0.51 
Note slopec 0.31 0.43 -0.24 
Note bandwidth -0.33 0.52 0.02 
Note rate 0.36 -0.04 -0.47 

aMean time between the first 4 notes of a song.

bMean loudest frequency for each note in song.

cNote bandwidth divided by note length.

In contrast to plumage color, which likely has a large genetic component (see section “Discussion”), songs in red-backed fairy-wrens are likely in large part learned. The Malurus fairy-wrens are oscine passerines (Gardner et al. 2010), a group that is known to have significant cultural contributions to song phenotypes (Kroodsma and Miller 1996). Indeed, males of the closely related splendid fairy-wren (Malurus splendens) have been shown to learn their songs through social association with minimal genetic contributions to similarity of songs across individuals (Greig et al. 2012). In red-backed fairy-wrens, such cultural contributions to song phenotypes, paired with limited male dispersal (Varian-Ramos and Webster 2012) and high extra-pair paternity rates (Webster et al. 2008), potentially create an opportunity for songs to become disassociated from genetic traits; extra-pair offspring may learn songs of socially associated but unrelated individuals, and dispersing females may carry with them genes but not learned male songs (see Lein and Corbin 1990). Given the apparent plumage introgression, this system provides an ideal opportunity to compare introgression patterns of sexual signals that are transmitted genetically (plumage) versus culturally (song).

Several different patterns are possible for geographic variation in song, plumage, and genes, and these depend on the patterns of coupling among phenotypic traits and on the patterns of selection (e.g., through responses of conspecifics to song). First, we predict that song will show a pattern different from that of plumage traits, and likely more similar to that of neutral genetic markers, if song and plumage traits are readily decoupled and introgression patterns reflect independent selective pressures on each trait. In such a scenario, songs may either be resistant to introgression (and mirror genes) or show introgression patterns that reflect asymmetrical social or environmental selective pressures on songs (which may or may not mirror the geographic patterns of plumage). Second, song and plumage traits may introgress in concert if these traits are tightly coupled by common patterns of inheritance or reinforcing selective pressures (e.g., favoring introgression of both sexual traits together). For example, if plumage has introgressed because cruentatus males disperse in a unidirectional manner into melanocephalus populations, then they may bring with them learned songs, as well as genes affecting male plumage. Alternatively, males with a favored plumage type could be hindered from dispersal if they sing an unfavored song, and thus, learned songs could slow the introgression of plumage traits. Here, we describe geographic patterns of song and compare these to patterns of genetic and plumage variation. Additionally, we assess how geographic variation in song is related to the strength of conspecific response during territorial interactions. This allows us to discriminate among the different introgression scenarios described above and evaluate hypotheses regarding the possible mechanisms of trait dispersal.

MATERIALS AND METHODS

Study species

Red-backed fairy-wrens are cooperatively breeding insectivorous passerines that inhabit open forest in northern and eastern Australia (Rowley and Russell 1997). These birds are nonmigratory (Rowley and Russell 1997), and male dispersal is limited (Varian-Ramos and Webster 2012), which allows us to assess song divergence in relation to genetic and plumage divergence in a system uncomplicated by postdispersal song learning. Males exhibit delayed plumage maturation, so they may retain female-like brown plumage either as auxiliary helpers or as primary breeders for multiple seasons before adopting red–black nuptial plumage (Rowley and Russell 1997; Webster et al. 2008, 2010). Extra-pair paternity is common, occurring in approximately 63% of broods (Webster et al. 2008). Both males and females sing a typical Malurus reel, notably during dawn chorus displays and during territorial interactions (Rowley and Russell 1997).

Song recording and acoustic analysis

We recorded bouts of male dawn chorus songs from 15 populations (Figure 1, Table 1: n = 132 males, 3–23 males per site, see Supplementary Figure 1 for examples of spectrograms from 12 sites). We recorded each male for 5–10min using a Marantz PMD 661 solid-state digital recorder at 96kHz sampling rate, 24-bit depth, or using a Marantz PMD 670 at 48kHz sampling rate, 16-bit depth (D&M Professional, Itasca, IL), combined with ME66 shotgun microphone capsules and K6 power modules (Sennheiser Electronic Corporation, Old Lyme, CT; frequency response 0.04–20.0kHz). Recordings were collected between October 2010–January 2011 and November–December 2011, when birds were in breeding condition. We included 1 individual from an additional site (Mullumbimby, QLD) using an archived recording of a dawn chorus from the Macaulay Library of Natural Sounds (Cornell Lab of Ornithology, Ithaca, NY, ML#149220). For analyses, we chose 1 example song from each male that was of high recording quality based on spectrograms digitized in Raven 1.4 (Bioacoustics Research Program, 2004; 16-bit sample format; discrete Fourier transform, DFT = 512 samples; frequency resolution = 124 Hz; time resolution = 11.6ms; frame overlap = 50%). We measured individual note characteristics in Luscinia (Lachlan 2007) and summarized 9 acoustic parameters for each song (Table 2). We chose these particular acoustic measurements because they captured straightforward aspects of variation in songs.

Playback protocol

At a subset of the song recording sites (Table 1), we conducted playbacks to focal individuals using songs from 3 primary playback categories: 1) a positive control, which was a homotypic recording from the same subspecies and plumage color as the focal population; 2) a negative control, which was a heterospecific white-winged fairy-wren (M. leucopterus); and 3) a treatment of interest, which for focal cruentatus populations was a song from either red or orange melanocephalus populations and for focal melanocephalus populations was either a cruentatus song or a song from a differently colored melanocephalus population. Taken together, these playbacks allowed us to tease apart the effect on focal bird response of hearing a playback from a different genetic subspecies or a different plumage region; if birds discriminate against songs from the opposite subspecies irrespective of plumage type (which we expect if song variation tracks subspecies identity), then we should observe strong responses to songs from the same genetic subspecies but low responses to the opposite subspecies of any plumage type (i.e., focal cruentatus individuals should respond to cruentatus songs but not to melanocephalus songs from either plumage region, and focal melanocephalus individuals should respond to melanocephalus songs from either plumage region but not to cruentatus songs). Alternatively, if birds discriminate against songs from the opposite plumage region irrespective of subspecies identity (which we expect if song variation tracks plumage variation), then we should observe strong responses to songs from the same plumage type, but low responses to the opposite plumage type regardless of subspecies identity (i.e., focal cruentatus individuals should respond strongly to cruentatus songs and melanocephalus songs from the red plumage region but not to melanocephalus songs from the orange plumage region, and focal melanocephalus individuals should respond to melanocephalus or cruentatus songs from the same plumage region but not to melanocephalus songs from the opposite plumage region).

We presented playbacks from all 3 categories (i.e., 2 controls and 1 treatment) to every focal individual so that we could assess within-individual variation in response, and we balanced the order of playbacks so that every category was equally represented in every order position. In our analysis, playback order had no effect on either male or female response strength (P > 0.2), so we subsequently removed that factor from our analysis. We separated playbacks to the same individual by a minimum of 5min and kept track of the focal birds (either by sight or by sound) until all playbacks were completed to ensure that playbacks within a series were in fact to the same bird. Trials were separated by a longer interval of time if focal birds were not in clear sight after 5min (e.g., if they were foraging in tall grass or flying to another location); we waited until the birds were in view and stationary before beginning a trial. Focal birds typically resumed preplayback behavior (e.g., foraging) within 1–2min after the playback, so responses tended to be immediate and brief.

To the extent possible, we used a different playback exemplar for every trial to focal birds of each subspecies. We used a total of 20 stimuli from heterospecifics, 28 stimuli from the genetic cruentatus (red) region, and 25 stimuli from the genetic melanocephalus region, of which 13 were from the region where birds have red plumage and 12 were from the region where birds have orange plumage. Playback stimuli consisted of 1 song repeated twice with an interval of approximately 10 s between songs. We used an amplified speaker (Pignose Legendary 7–100, Pignose-Gorilla, Las Vegas, NV; frequency response, 0.1–12.0kHz) positioned approximately 1 m off the ground and an iPod nano (Apple Inc., Cupertino, CA) with uncompressed .WAV files for playback. We standardized the amplitude of playbacks in Raven 1.4 and tested them with the field playback equipment using a sound–pressure level meter (model number 33–2050, Radio Shack Corporation, Fort Worth, TX) set at C-weighting, fast response (approximately 89.0 dB at 1 meter for all playbacks). This amplitude was similar to observed natural dawn song levels. The initial distance between the speaker and the focal birds ranged from approximately 10–20 m, and our analysis revealed no effect of distance on either male or female response strength (P > 0.2), so we subsequently removed this factor from the analysis.

All playbacks were conducted during the breeding season between November 2010 and January 2011. Focal males were in red–black breeding plumage (rather than brown plumage) for all but 5 trials. We observed nesting behavior at every playback site but were unsure of the precise nesting stage of most focal individuals. Although nesting stage may have an effect on focal bird response, our playback design allowed us to assess within-individual variation in response while holding nesting stage constant. The primary focal individuals were males, but additional brown birds were present in all but 3 playbacks, so we also recorded their vocal responses and dictated their behavioral responses. Brown birds were likely female mates of red–black plumaged males, and we refer to these birds as “putative females” below, although some may have been auxiliary helpers.

One person (E.G.) conducted each trial and observed focal birds for a minimum of 5min before beginning a playback series to confirm that birds were not engaged in a territorial interaction or disturbed by the observer’s presence. Behavioral responses were dictated for 5-min postplayback into the same recorder that was used to record vocal responses of focal individuals. We noted the occurrence of and latency to the behaviors of approaching (flying) toward the speaker, flying over the speaker, and singing. We chose these responses because they were unambiguous and thus minimally subject to observer bias and because they typically reflect territorial aggression (Kroodsma and Byers 1991). Birds tended to respond either quickly or not at all (for birds that responded, median latency to focal male approach = 15.0 s, sing = 31.5 s; putative female approach = 7.0 s, sing = 34.0 s), so we did not use latency to respond in our final analysis because it was no more informative than presence/absence of response. We classified behaviors as a positive response if they occurred within 60 s of the beginning of the playback; approach or song behaviors after this time were relatively infrequent (6/87 total approach responses and 38/116 total song responses occurred after 60 s). Although this method of classification may underestimate positive responses, it is also likely to exclude behaviors that were not in response to playback and that occurred for other reasons during the 5-min postplayback observation. We noted the distance of nearest approach to speaker, but we did not use this in our analysis because it appeared to be dependent on the available tall vegetation. We also attempted to record the duration of response, but it was often unclear what constituted the end of a response (birds would sometimes begin foraging, begin preening, or would simply stop singing and perch calmly), so we felt this was too subjective to quantitatively include in our analysis. We therefore created overall response scores for focal males and putative females using the sum of the nominal responses approach, sing, and fly over; thus, an individual’s response could range from 0 (no response) to 3 (individual approached, flew over speaker, and sang).

Statistical analyses

First, we assessed if song variation corresponded spatially to either the location of subspecies (i.e., genetic) or plumage divergence. To do this, we summarized the 9 acoustic parameters for each song using a principle components analysis (Table 2). Using the first principle component (which explained 59.62% of the variation) as our metric of song variation, we used a linear mixed model approach with sampling site (n = 16) as a random effect to assess the relative importance of the plumage and subspecies regions on song variation. Additionally, we created song pairs (assigning songs randomly to a pair and using each song only once, so pairs were independent) for songs 1) within each subspecies and plumage region, 2) between subspecies regions, and 3) between plumage regions; we then calculated the Euclidean distances between the mean song scores (PC1, PC2, PC3) for each song pair and compared overall song similarity within and between subspecies and plumage regions using Student’s t-tests. Finally, we used agglomerative hierarchical clustering (Ward 1963), with all song variables in Table 2, to classify songs into 2 categories; we then assessed the distribution of those categories along the sampling transect to estimate which location explained the greatest proportion of variation in song.

To assess the influence of receiving a playback from a different subspecies or plumage region on individual response to playback, we used 2 statistical methods. First, we used a generalized linear mixed model (GLMM) approach with Laplace approximation (implemented in R 2.10.1 using the “glmer” function in the package lmer4; R Core Development Team 2008) with individual as a random effect and subspecies region and plumage region as fixed effects to test for an overall effect of subspecies identity or plumage region on response to playback. We specified a Poisson distribution for the response variable to account for the zero skew in focal bird response scores (Bolker et al. 2009). We included playback order and initial distance in the initial model but removed these factors from the final analysis because they were not significant predictors of response strength. We used a similar GLMM approach to assess if response strength could be explained by acoustic similarity between focal population and playback type irrespective of subspecies identity and plumage region. For this analysis, we calculated acoustic similarity as the Euclidean distance between the mean song scores of the focal population and the scores of the actual playback stimulus (PC1, PC2, PC3). Songs from more distant locations were acoustically more different (r2 = 0.65, F1,127 = 235.3, P < 0.0001), so we accounted for geographic distance between playback source and focal population in this analysis. For most sites, we used the shortest linear distance between the 2 locations as our measure of geographic distance. Because of the arch-shaped distribution along the coast, some linear distances crossed areas that were outside the species range, which are not available for movement of individuals. We corrected for this underestimate of distance by choosing a pivot point on the inner edge of the range (Figure 1) and calculating the distance between sites as the sum of the distances between each site and the pivot point. We used GPS coordinates (GPS 60CSx, Garmin International, Inc., Olathe, KS) collected at the time of playback to make all distance calculations.

Next, we assessed within-individual variation in response strength for each focal subspecies to the homotypic control, the heterospecific control, and the respective treatments using non–parametric matched comparisons (Wilcoxon signed rank tests). This allowed us to verify patterns found in the previous analysis and to compare the response tendencies of cruentatus and melanocephalus populations and determine whether responses were symmetrical between the 2 subspecies. We used between-subject nonparametric Wilcoxon rank sum tests to compare the response strengths of each subspecies for the various playback types (i.e., controls and treatments) when those could be meaningfully compared (for instance, there is only 1 plumage type in cruentatus populations, so there is no analogue to the within-subspecies, different plumage type playbacks that can be conducted in melanocephalus populations). All statistical analyses were 2-tailed and were conducted in JMP 5.1.2 (SAS Institute Inc.) or R 2.10.1 (R Core Development Team 2008).

RESULTS

Song variation across the species range

A linear mixed model comparing the effects of subspecies identity (i.e., genetic region) and plumage region on song variation (PC1) revealed that genetic subspecies identity was a significant predictor of song structure (F1,130 = 46.5, P < 0.0001), but plumage region was not (F1,130 = 1.05, P = 0.306). Thus, song variation was best explained by genetic subspecies boundaries, irrespective of plumage variation. Site-level variation was present to some degree, but many sites within subspecies boundaries had statistically indistinguishable songs, whereas all sites that crossed subspecies boundaries were significantly different (Table 1). This was supported by our assessment of song similarity based on Euclidean distances within and between subspecies and plumage regions; between-subspecies song distances were the largest irrespective of plumage type, but between-plumage song differences were equal to or less than song differences within subspecies and plumage types (Figure 3, see Supplementary Table 1 for details of Euclidean distance analysis). Finally, agglomerative hierarchical clustering using all 9 song variables revealed 2 distinct clusters, with each cluster corresponding to a particular geographic range and the break between clusters falling near the geographic location of subspecies (i.e., genetic) divergence, rather than the location of plumage divergence (Figure 2). In summary, song variation mirrored spatial patterns of genetic variation, rather than plumage variation, resulting in a region composed of songs characteristic of melanocephalus individuals and plumage characteristic of cruentatus individuals (intermediate gray area in Figure 1).

Figure 3

The first 3 principle components for the sampled song from each male (n = 133 males). Gray points indicate songs from red cruentatus populations, black points indicate songs from red melanocephalus populations, and white points indicate songs from orange melanocephalus populations.

Figure 3

The first 3 principle components for the sampled song from each male (n = 133 males). Gray points indicate songs from red cruentatus populations, black points indicate songs from red melanocephalus populations, and white points indicate songs from orange melanocephalus populations.

Figure 2

PC1 for the sampled song from each male (n = 133 males) plotted against the distance from the furthest western sampling site (Broome, WA). Dotted lines indicate the approximate geographic locations of the genetic and plumage divides relative to our sampling locations. Filled and unfilled points distinguish 2 main song types based on cluster analysis.

Figure 2

PC1 for the sampled song from each male (n = 133 males) plotted against the distance from the furthest western sampling site (Broome, WA). Dotted lines indicate the approximate geographic locations of the genetic and plumage divides relative to our sampling locations. Filled and unfilled points distinguish 2 main song types based on cluster analysis.

Response to playbacks

Mixed models revealed that both males and putative females responded more strongly to playbacks from the same genetic subspecies than to playbacks from a different subspecies (male, n = 66, z = 3.30, P < 0.001; female, n = 64, z = 3.10, P = 0.002). In contrast, receiving a playback from a different plumage region did not affect response to playback (male, n = 66, z = 0.26, P = 0.378; female, n = 64, z = -0.08, P = 0.937). Thus, birds tended to discriminate against songs from the other subspecies irrespective of plumage type. When controlling for geographic distance, acoustic divergence between playback and focal population was a significant predictor of putative female response strength (n = 64, acoustic: z = -1.97, P = 0.048, geographic: z = -.028, P = 0.781) and a marginally nonsignificant predictor of male response strength (n = 66, acoustic: z = -1.84, P = 0.065, geographic: z = -0.97, P = 0.329). Overall, therefore, greater response strength was associated with greater genetic similarity (i.e., subspecies identity) irrespective of plumage similarity and was likely driven by the high degree of acoustic similarity within but not between subspecies.

Within-individual matched comparisons of responses for each focal subspecies supported the previous results and revealed that the pattern was primarily driven by discrimination against cruentatus songs in melanocephalus populations. Focal males of both cruentatus and melanocephalus populations showed the strongest responses to playbacks from within the same subspecies and plumage region (i.e., homotypic positive controls), melanocephalus males showed moderately high responses to melanocephalus playbacks from a different plumage region, and males from both populations showed low responses to heterospecific playbacks (Table 3, Figure 4). However, cruentatus males showed moderate responses to playback of melanocephalus songs from either plumage region, whereas melanocephalus males showed very low responses to cruentatus playbacks, similar to their responses to heterospecifics (Table 3, Figure 4). Putative females showed similar patterns of discrimination against the opposite subspecies, although female responses were lower overall, and subspecies discrimination was more obvious in cruentatus populations (Table 3, Figure 4). Females from melanocephalus populations did not show any significant differences in responses to playbacks of homotypic controls, treatments, and heterospecific controls (Table 3). However, this appears to reflect low female responses to all playbacks rather than a lack of discrimination against treatments and heterospecifics. Mixed models revealed complementary results; in melanocephalus populations, male response was affected by genetic subspecies identity but not plumage type of playback (subspecies: z = 3.46, P < 0.001, plumage: z = 1.19, P = 0.236, whereas female response was not affected by either factor (P > 0.1, see Supplementary Table 2 for full GLMM report). In cruentatus populations, female response was affected by genetic subspecies identity but not plumage type of playback (subspecies: z = 2.07, P < 0.038, plumage: z = 0.47, P = 0.640), whereas male response was not affected by either factor (P > 0.2, see Supplementary Table 2).

Table 3

For each focal sex and subspecies, mean response strengths and Wilcoxon signed rank comparisons for the homotypic control playbacks (control) and the different treatment types (including heterospecific controls)

Focal Control Treatment type Treatment n z P 
Males 
cruentatus 1.19 melanocephalus (red) 0.76 21 24 0.18 
0.93 melanocephalus (orange) 0.57 14 0.344 
1.09 heterospecific 0.26 34 123 <0.001 
melanocephalus 1.29 melanocephalus (different plumage) 1.00 0.812 
1.35 cruentatus (red) 0.12 17 42 0.002 
1.32 heterospecific 0.18 22 45 <0.001 
Females 
cruentatus 0.68 melanocephalus (red) 0.16 19 22.5 0.020 
0.38 melanocephalus (orange) 0.08 13 0.312 
0.58 heterospecific 0.06 31 39 <0.001 
melanocephalus 0.57 melanocephalus (different plumage) 0.57 1.000 
0.47 cruentatus (red) 0.12 17 7.5 0.062 
0.55 heterospecific 0.23 22 11 0.148 
Focal Control Treatment type Treatment n z P 
Males 
cruentatus 1.19 melanocephalus (red) 0.76 21 24 0.18 
0.93 melanocephalus (orange) 0.57 14 0.344 
1.09 heterospecific 0.26 34 123 <0.001 
melanocephalus 1.29 melanocephalus (different plumage) 1.00 0.812 
1.35 cruentatus (red) 0.12 17 42 0.002 
1.32 heterospecific 0.18 22 45 <0.001 
Females 
cruentatus 0.68 melanocephalus (red) 0.16 19 22.5 0.020 
0.38 melanocephalus (orange) 0.08 13 0.312 
0.58 heterospecific 0.06 31 39 <0.001 
melanocephalus 0.57 melanocephalus (different plumage) 0.57 1.000 
0.47 cruentatus (red) 0.12 17 7.5 0.062 
0.55 heterospecific 0.23 22 11 0.148 

Response means differ slightly for the homotypic control playbacks (which were playbacks from the same subspecies and plumage region) because each matched comparison was composed of a different subsets of trials.

Figure 4

Mean response strength for males (gray) and females (white) of each subspecies for the different treatment types. Playbacks to cruentatus populations are shown above the dashed line and playbacks to melanocephalus populations below the dashed line. Response means here differ slightly from those listed in Table 3 because these represent all playbacks irrespective of the matched comparisons. For playbacks to melanocephalus populations, “same plumage” includes playbacks to both red and orange populations, and “different plumage” includes playbacks of songs from orange populations to red populations and vice versa. Error bars indicate 1 standard error. See text for comparisons of treatment types between subspecies.

Figure 4

Mean response strength for males (gray) and females (white) of each subspecies for the different treatment types. Playbacks to cruentatus populations are shown above the dashed line and playbacks to melanocephalus populations below the dashed line. Response means here differ slightly from those listed in Table 3 because these represent all playbacks irrespective of the matched comparisons. For playbacks to melanocephalus populations, “same plumage” includes playbacks to both red and orange populations, and “different plumage” includes playbacks of songs from orange populations to red populations and vice versa. Error bars indicate 1 standard error. See text for comparisons of treatment types between subspecies.

Between-subspecies comparisons for each playback type revealed similar patterns. For both males and females, responses to playbacks from the same subspecies and plumage region (i.e., homotypic positive controls) were similar in cruentatus and melanocephalus populations, as were responses to heterospecific playbacks (Figure 4; P > 0.05 for between-subspecies comparisons of homotypic and heterospecific playback responses). There was, however, a significant difference among subspecies in the strength of male response to playbacks from a different genetic region; melanocephalus males responded very weakly to cruentatus songs (similar to their response to heterospecifics), but cruentatus males responded at an intermediate level to melanocephalus songs of either plumage type (Figure 4; comparison of cruentatus response to melanocephalus playbacks of both plumage types vs. melanocephalus response to cruentatus playbacks: n = 59, z = -2.68, P = 0.007).

Complementing the asymmetric discrimination tendencies among subspecies evident from the above analysis, we found that acoustic similarity between playback source and focal population had different relative effects in melanocephalus and cruentatus populations. In cruentatus populations, response strength of neither males (n = 38) nor putative females (n = 36) was significantly correlated with acoustic distance when controlling for geographic distance (P > 0.1 for all correlations; see Supplementary Table 2 for full GLMM report). In contrast, in melanocephalus populations, male (n = 28) response was significantly correlated with acoustic distance when controlling for geographic distance (acoustic: z = -2.56, P = 0.010, geographic: z = -0.89, P = 0.375), although female (n = 28) response was not (P > 0.2; see Supplementary Table 2).

In summary, cruentatus males (but not females) were moderately responsive to a broad range of song stimuli, including acoustically distant members of the other subspecies irrespective of plumage type. Melanocephalus males only responded strongly to songs that were acoustically quite similar to their own (irrespective of plumage type) and discriminated strongly against acoustically distant songs of the other subspecies.

DISCUSSION

We found strong concordance between geographic patterns of song and genetic variation in red-backed fairy-wrens: the “break” between divergent songs occurred near the same geographic location as the break between genetically defined subspecies (Figure 2). This is in striking contrast to the geographic pattern of plumage, with the “break” in plumage types falling much farther east of the genetic divide (Figure 2;Lee and Edwards 2008, Baldassarre et al., in review). Mirroring this geographic pattern of song variation, we found that focal birds tended to discriminate against songs from the opposite genetic subspecies but not to songs from the opposite plumage region when subspecies identity was taken into account. This was particularly evident in melanocephalus populations, where birds essentially did not respond to cruentatus songs. These patterns of territorial response suggest strong differences in the abilities of males with foreign song types to establish breeding territories. We suspect that responses of putative females to our playbacks reflect territorial aggression rather than sexual attraction (Cooney and Cockburn 1995), but nonetheless females may show similar or greater levels of discrimination during mating interactions (Searcy et al. 2002; Danner et al. 2011). Thus, our results illustrate that songs have been resistant to introgression, which may be due in part to discrimination against foreign songs in territorial, and potentially sexual, interactions.

Although song and response divergence across populations, as found in this study, might be expected to reduce gene flow, song divergence does not appear to have hindered the introgression of plumage traits. We suggest that this is because song in this system, as a learned trait, can be readily decoupled from introgressing genetic traits, and thus, song may be a relatively ineffective barrier to introgression. Although song inheritance has not been studied specifically in red-backed fairy-wrens, songs are learned from a male’s social father in the closely related splendid fairy-wren (M. splendens; Greig et al. 2012), and social fathers are often unrelated to the sons that they tutor due to high rates of extra-pair paternity in both splendid (Webster et al. 2004) and red-backed fairy-wrens (Webster et al. 2008). In contrast, male plumage color differences across red-backed fairy-wren populations are likely to be, at least in part, due to genetic differences because carotenoid-based plumage appears to have significant genetic underpinnings in other species (Brush 1990; Griffith et al. 2006), and geographic variation in red-backed fairy-wren plumage suggests a genetic basis (Baldassarre et al., in review). Note that a significant affect of environment on male plumage color (e.g., through diet) would still represent a significant deviation in mode of transmission from learned song and therefore could permit ready decoupling from song.

The result of inheritance differences between song and plumage may be differential movement of such traits across zones of secondary contact. This could occur via multiple mechanisms; consider, for example, a dispersive red-backed fairy-wren male with cruentatus plumage and song that secures extra-pair paternity in the nest of a male with melanocephalus plumage and song. In this scenario, offspring would inherit (genetically) more cruentatus-like plumage but would learn the songs of their cuckolded melanocephalus father (note that this scenario requires melanocephalus females to accept as extra-pair mates males with cruentatus songs). Alternatively, because red-backed fairy-wren females are the more dispersive sex (Varian-Ramos and Webster 2012), genetic traits may be carried by females more readily than learned male songs (Lein and Corbin 1990); a dispersive cruentatus female could carry with her genes affecting the plumage of her sons, but if the female paired with a melanocephalus male, her sons would learn melanocephalus songs. In either scenario, the learning process constrains spread of cultural traits but not genetic traits. We cannot distinguish between these scenarios with the data set presented here, but our results do suggest that plumage introgression has not occurred via the asymmetrical dispersal of cruentatus males. If this were the case, we would expect dispersive cruentatus males to carry with them their western song phenotypes and for song variation to mirror variation in plumage.

In this system, the consequence of plumage-song decoupling is a large region with a novel phenotype, that is, males with melanocephalus songs (and genes) but cruentatus plumage. This novel phenotype suggests an important consequence of trait decoupling in zones of secondary contact: novel combinations of traits arising from differential introgression may facilitate further introgression because “novel” individuals may enjoy advantages of having one type of sexual signal (e.g., red plumage) without disadvantages of having another (e.g., a foreign song). Novel trait combinations might also arise if one or both traits respond to different environmental factors or if different traits diverged stochastically in response to different divergence events. However, these latter explanations seem unlikely for red-backed fairy-wrens because environmental factors appear to have had little influence on plumage traits (Baldassarre et al., in review), and song divergence appears to have occurred across the same geographic barrier that has affected genetic divergence (Lee and Edwards 2008). Although it is possible that plumage traits diverged stochastically across the more eastern Burdekin Barrier (Lee and Edwards 2008), rather than having introgressed from the Carpentarian Barrier, we expect stochastic trait divergence to be most dramatic, following the greatest periods of isolation, as is suggested by song divergence in this system.

Both subspecies showed the strongest response to local songs, indicating that there were not strong asymmetries between subspecies in their song preferences, but the strength of discrimination against songs from the opposite subspecies was not equal. Rather, cruentatus males were less discriminating than melanocephalus males and showed moderate responses to songs from melanocephalus populations irrespective of plumage type. In contrast, melanocephalus individuals showed extremely low responses to cruentatus songs, suggesting that dispersal of cruentatus males into melanocephalus populations (the direction that would complement plumage introgression) may be more difficult than the reverse. These results emphasize that any introgression of plumage has occurred despite opposing pressures on songs (i.e., discrimination against foreign songs), which in turn emphasizes that trait decoupling may significantly facilitate additional introgression.

In summary, we have shown that traits with different modes of inheritance may exhibit vastly different introgression patterns when divergent populations hybridize, leading to the generation of large regions with novel combinations of parental population traits. Although decoupling of traits has been widely documented in hybrid zones, particularly where unidirectional introgression has occurred (Anderson and Stebbins 1954; Parsons et al. 1993; Chiba 1997; Johnston et al. 2004; Jørgensen and Mauricio 2005; Stein and Uy 2006), this study provides an example where traits have different modes of inheritance and therefore emphasizes the potential complexity of phenotypes when genetic and cultural factors interact (Lein and Corbin 1990; Baker and Boylan 1999; Sattler et al. 2007; Baker 2010). These different modes of inheritance likely facilitate decoupling of traits and thus introgression of some traits (e.g., plumage) despite potential reproductive barriers imposed by others (e.g., song). More generally, our results illustrate the ability of cultural traits and conspecific responses to those traits to remain divergent even when reproductive isolation is incomplete. Documenting the mechanism of plumage introgression and investigating the simultaneous effects of divergent plumage and song in territorial and sexual interactions will be necessary to fully understand the complexities of this system. Nonetheless, this work highlights the degree to which cultural traits may be decoupled from genetic traits and suggests mechanistic hypotheses for how such decoupling could occur.

SUPPLEMENTARY MATERIAL

Supplementary material can be found at http://www.beheco.oxfordjournals.org/.

FUNDING

Support provided by National Science Foundation (M.S.W). All work was conducted with approval from appropriate animal ethics and permitting agencies: Cornell University Animal Care and Use Committee Approval (2009-0105); James Cook University Ethics Approval (A1340); Scientific Purposes Permit (WISP07773610); Regulation 17 license (SF007698).

We thank Dan Baldassarre, Emily Cramer, Jenélle Dowling, Kristen Hook, Sara Kaiser, Trevor Price, Aaron Rice, Benjamin Risk, Nathan Senner, Benjamin Taft and 4 anonymous reviewers for discussion and comments. Alex White assisted in the field. Matt Medler provided recordings from the Macaulay Library. Duncan Yandell assisted with acoustic measurements. Staff at the Mornington Wildlife Sanctuary, Michelle Hall, Karen and Angus Emmott, Joanne Heathcote, Jan Lewis, Doug Barron and Tim Daniels provided logistical advice and assistance during fieldwork.

REFERENCES

Anderson
E
Stebbins
GL
Jr
1954
.
Hybridization as an evolutionary stimulus
.
Evolution
 .
8
:
378
388
.
Arnold
ML
.
1992
.
Natural hybridization as an evolutionary process
.
Annu Rev Ecol Syst
 .
23
:
237
261
.
Baker
MC
.
2010
.
Hybrid zones and cultural traits of the Australian ringneck Platycercus zonarius
.
J Avian Biol
 .
41
:
50
63
.
Baker
MC
Boylan
JT
.
1999
.
Singing behavior, mating associations and reproductive success in a population of hybridizing Lazuli and Indigo Buntings
.
Condor
 .
101
:
493
504
.
Baker
MC
Mewaldt
LR
.
1978
.
Song dialects as barriers to dispersal in white-crowned sparrows, Zonotrichia leucophrys nuttalli
.
Evolution
 .
32
:
712
722
.
Baptista
LF
Gaunt
SLL
.
1997
.
Social interaction and vocal development in birds
. In:
Snowdon
CT
Hausberger
M
, editors.
Social influences on vocal development
 .
Cambridge
:
Cambridge University Press
.
Barton
NH
.
2001
.
The role of hybridization in evolution
.
Mol Ecol
 .
10
:
551
568
.
Bioacoustics Research Program
2004
.
Raven Pro: Interactive Sound Analysis Software Version 1.2
 .
Ithaca (NY)
:
The Cornell Lab of Ornithology
[Computer program]
Bolker
BM
Brooks
ME
Clark
CJ
Geange
SW
Poulsen
JR
Stevens
MH
White
JS
.
2009
.
Generalized linear mixed models: a practical guide for ecology and evolution
.
Trends Ecol Evol (Amst)
 .
24
:
127
135
.
Brush
AH
1990
.
Metabolism of carotenoid pigments in birds
.
FASEB J
 .
4
:
2969
2977
.
Chiba
S
.
1997
.
Novel colour polymorphisms in a hybrid zone of Mandarina (Gastropoda: Pulmonata)
.
Biol J Linn Soc
 .
61
:
369
384
.
Cooney
R
Cockburn
A
.
1995
.
Territorial defense is the major function of female song in the Superb Fairy-Wren, Malurus cyaneus
.
Anim Behav
 .
49
:
1635
1647
.
Coyne
JA
Orr
HA
editors.
2004
.
Behavioral and nonecological isolation
. In:
Speciation
 .
Sunderland (MA)
:
Sinauer Associates, Inc
.
Danner
JE
Danner
RM
Bonier
F
Martin
PR
Small
TW
Moore
IT
.
2011
.
Female, but not male, tropical sparrows respond more strongly to the local song dialect: implications for population divergence
.
Am Nat
 .
178
:
53
63
.
den Hartog
PM
den Boer-Visser
AM
ten Cate
C
.
2010
.
Unidirectional hybridization and introgression in an avian contact zone: evidence from genetic markers, morphology and comparisons with laboratory-raised F1 hybrids
.
Auk
 .
127
:
605
616
.
Dingle
C
Poelstra
JW
Halfwerk
W
Brinkhuizen
DM
Slabbekoorn
H
.
2010
.
Asymmetric response patterns to subspecies-specific song differences in allopatry and parapatry in the gray-breasted wood-wren
.
Evolution
 .
64
:
3537
3548
.
Dowling
TE
Secor
CL
.
1997
.
The role of hybridization and introgression in the diversification of animals
.
Annu Rev Ecol Syst
 .
28
:
593
619
.
Edwards
SV
Kingan
SB
Calkins
JD
Balakrishnan
CN
Jennings
WB
Swanson
WJ
Sorenson
MD
.
2005
.
Speciation in birds: genes, geography, and sexual selection
.
Proc Natl Acad Sci USA
 .
102
Suppl 1
6550
6557
.
Feldman
MW
Laland
KN
.
1996
.
Gene-culture coevolutionary theory
.
Trends Ecol Evol
 .
11
:
453
457
.
Gardner
JL
Trueman
JW
Ebert
D
Joseph
L
Magrath
RD
.
2010
.
Phylogeny and evolution of the Meliphagoidea, the largest radiation of Australasian songbirds
.
Mol Phylogenet Evol
 .
55
:
1087
1102
.
Grant
PR
Grant
BR
.
1997
.
Mating patterns of Darwin’s finch hybrids determined by song and morphology
.
Biol J Linn Soc
 .
60
:
317
343
.
Grant
PR
Grant
BR
.
2009
.
The secondary contact phase of allopatric speciation in Darwin’s finches
.
Proc Natl Acad Sci USA
 .
106
:
20141
20148
.
Grant
PR
Grant
BR
Deutsch
JC
.
1996
.
Speciation and hybridization in island birds
.
Proc R Soc Lond B Biol Sci
 .
351
:
765
772
.
Greig
EI
Taft
B
Pruett-Jones
S
.
2012
.
Sons learn songs from their social fathers in a cooperatively breeding bird
.
Proc R Soc Lond B Biol Sci
 .
279
:
3154
3160
.
Griffith
SC
Parker
TH
Olson
VA
.
2006
.
Melanin- versus carotenoid-based sexual signals: is the difference really so black and red?
Anim Behav
 .
71
:
749
763
.
Haavie
J
Borge
T
Bures
S
Garamszegi
LZ
Lampe
HM
Moreno
J
Qvarnström
A
Török
J
Saetre
GP
.
2004
.
Flycatcher song in allopatry and sympatry–convergence, divergence and reinforcement
.
J Evol Biol
 .
17
:
227
237
.
Helb
HW
Dowsettlemaire
F
Bergmann
HH
Conrads
K
.
1985
.
Mixed singing in European songbirds: a review
.
Z Tierpsychol
 .
69
:
27
41
.
Irwin
DE
Price
T
.
1999
.
Sexual imprinting, learning and speciation
.
Heredity
 .
82
:
347
354
.
Johnston
JA
Donovan
LA
Arnold
ML
.
2004
.
Novel phenotypes among early generation hybrids of two Louisiana iris species: flooding experiments
.
J Ecol
 .
92
:
967
976
.
Jørgensen
S
Mauricio
R
.
2005
.
Hybridization as a source of evolutionary novelty: leaf shape in a Hawaiian composite
.
Genetica
 .
123
:
171
179
.
Karubian
J
.
2002
.
Costs and benefits of variable breeding plumage in the red-backed fairy-wren
.
Evolution
 .
56
:
1673
1682
.
Karubian
J
Swaddle
JP
Varian-Ramos
CW
Webster
MS
.
2009
.
The relative importance of male tail length and nuptial plumage on social dominance and mate choice in the red-backed fairy-wren Malurus melanocephalus: evidence for the multiple receiver hypothesis
.
J Avian Biol
 .
40
:
559
568
.
Kirschel
AN
Blumstein
DT
Smith
TB
.
2009
.
Character displacement of song and morphology in African tinkerbirds
.
Proc Natl Acad Sci USA
 .
106
:
8256
8261
.
Kroodsma
DE
Byers
BE
.
1991
.
The function(s) of bird song
.
Am Zool
 .
31
:
318
328
.
Kroodsma
DE
Miller
EH
.
1996
.
Ecology and evolution of acoustic communication in birds
.
Ithaca (NY)
:
Comstock Pub
.
Krosby
M
Rohwer
S
.
2009
.
A 2000 km genetic wake yields evidence for northern glacial refugia and hybrid zone movement in a pair of songbirds
.
Proc R Soc Lond B Biol Sci
 .
276
:
615
621
.
Lachlan
RF
.
2007
.
Luscinia: a bioacoustics analysis computer program
. Version 1.0 [Computer program]
Lee
JY
Edwards
SV
.
2008
.
Divergence across Australia’s Carpentarian barrier: statistical phylogeography of the red-backed fairy wren (Malurus melanocephalus)
.
Evolution
 .
62
:
3117
3134
.
Lein
MR
Corbin
KW
.
1990
.
Song and plumage phenotypes in a contact zone between subspecies of the White-crowned Sparrow (Zonotrichia leucophrys)
.
Can J Zool
 .
68
:
2625
2629
.
Parsons
TJ
Olson
SL
Braun
MJ
.
1993
.
Unidirectional spread of secondary sexual plumage traits across an avian hybrid zone
.
Science
 .
260
:
1643
1646
.
Patten
MA
Rotenberry
JT
Zuk
M
.
2004
.
Habitat selection, acoustic adaptation, and the evolution of reproductive isolation
.
Evolution
 .
58
:
2144
2155
.
Pearson
SF
Rohwer
S
.
2000
.
Asymmetries in male aggression across an avian hybrid zone
.
Behav Ecol
 .
11
:
93
101
.
Price
T
.
1998
.
Sexual selection and natural selection in bird speciation
.
Proc R Soc Lond B Biol Sci
 .
353
:
251
260
.
Price
T, editor
.
2008
.
Hybrid zones
. In:
Speciation in birds
 .
Greenwood Village (CO)
:
Roberts & Company Publishers
.
Price
T
, editor.
2008
.
Social selection and the evolution of song
. In:
Speciation in Birds. Greenwood Village (CO)
:
Roberts & Company Publishers
.
Qvarnström
A
Haavie
J
Saether
SA
Eriksson
D
Pärt
T
.
2006
.
Song similarity predicts hybridization in flycatchers
.
J Evol Biol
 .
19
:
1202
1209
.
R Core Development Team
2008
.
R: a language and environment for statistical computing
 .
Vienna, Austria
:
R Foundation for statistical computing
[Computer program]
Rowley
I
Russell
EM
.
1997
.
Fairy-wrens and grasswrens: Maluridae
 .
New York
:
Oxford University Press
.
Sánchez-Guillén
RA
Wellenreuther
M
Cordero-Rivera
A
Hansson
B
.
2011
.
Introgression and rapid species turnover in sympatric damselflies
.
BMC Evol Biol
 .
11
:
210
.
Sattler
GD
Sawaya
P
Braun
MJ
.
2007
.
An assessment of song admixture as an indicator of hybridization in Black-capped Chickadees (Poecile atricapillus) and Carolina Chickadees (P carolinensis)
.
Auk
 .
124
:
926
944
.
Searcy
WA
Nowicki
S
Hughes
M
Peters
S
.
2002
.
Geographic song discrimination in relation to dispersal distances in song sparrows
.
Am Nat
 .
159
:
221
230
.
Secondi
J
Bordas
P
Hipsley
C
Bensch
S
.
2011
.
Bilateral song convergence in a passerine hybrid zone: genetics contribute in one species only
.
Evol Biol
 .
38
:
441
452
.
Secondi
J
Bretagnolle
V
Compagnon
C
Faivre
B
.
2003
.
Species-specific song convergence in a moving hybrid zone between two passerines
.
Biol J Linn Soc
 .
80
:
507
517
.
Seddon
N
Tobias
JA
.
2007
.
Song divergence at the edge of Amazonia: an empirical test of the peripatric speciation model
.
Biol J Linn Soc
 .
90
:
173
188
.
Slabbekoorn
H
Smith
TB
.
2002
.
Bird song, ecology and speciation
.
Proc R Soc Lond B Biol Sci
 .
357
:
493
503
.
Sorenson
MD
Sefc
KM
Payne
RB
.
2003
.
Speciation by host switch in brood parasitic indigobirds
.
Nature
 .
424
:
928
931
.
Stein
AC
Uy
JA
.
2006
.
Unidirectional introgression of a sexually selected trait across an avian hybrid zone: a role for female choice?
Evolution
 .
60
:
1476
1485
.
Uy
JA
Moyle
RG
Filardi
CE
.
2009
.
Plumage and song differences mediate species recognition between incipient flycatcher species of the Solomon Islands
.
Evolution
 .
63
:
153
164
.
Varian-Ramos
CW
Webster
MS
.
2012
.
Extra-pair copulations reduce inbreeding for female red-backed fairy-wrens
.
Anim Behav
 .
83
:
857
864
.
Ward
JH
Jr
1963
.
Hierarchical grouping to optimize an objective function
.
J Am Stat Assoc
 .
58
:
236
244
.
Webster
MS
Karubian
J
Schwabl
H
.
2010
.
Dealing with uncertainty: flexible reproductive strategies by a tropical passerine bird in an unstable ecological and social environment
. In: Macedo R, editor.
Advances in the study of behavior: behavioral ecology of tropical animals
 .
42
:
123
153
.
Webster
MS
Tarvin
KA
Tuttle
EM
Pruett-Jones
S
.
2004
.
Reproductive promiscuity in the splendid fairy-wren: effects of group size and auxiliary reproduction
.
Behav Ecol
 .
15
:
907
915
.
Webster
MS
Varian
CW
Karubian
J
.
2008
.
Plumage color and reproduction in the red-backed fairy-wren: why be a dull breeder?
Behav Ecol
 .
19
:
517
524
.

Author notes

Handling editor: Shinichi Nakagawa

Supplementary data