Plumage coloration and social context influence male investment in song

Abstract

Animals use multiple signals to attract mates, including elaborate song, brightly coloured ornaments and physical displays. Female birds often prefer both elaborate male song and intense carotenoid-based plumage coloration. This could lead less visually ornamented males to increase song production to maximize their attractiveness to females. We tested this possibility in the highly social and non-territorial house finch (Haemorhous mexicanus), in which females discriminate among males based on both song and on the intensity of red carotenoid-based plumage coloration. We manipulated male plumage coloration through carotenoid supplementation during moult, so that males were either red or yellow. Males were then housed under three social environments: (i) all red birds, (ii) all yellow birds or (iii) a mixture of red/yellow birds. We recorded song after presentation of a female. Red males produced more song than yellow males. But when yellow males were housed with red conspecifics, they produced more song relative to yellow males housed with equally unattractive yellow males. This study provides novel evidence that a male's plumage coloration and the plumage colour of his social competitors influence investment in song.

1. Introduction

Multimodal signalling is common in sexually selected animal communication systems [1]. In birds, the intensity of sexually selected carotenoid-based plumage is determined by individuals' physiological condition and dietary carotenoid availability during plumage replacement (moult) [2]. Therefore, individuals' plumage coloration, and relative attractiveness compared with conspecifics, can vary between years. Since females often prefer both elaborate song and intense carotenoid-based plumage coloration [3], male birds could compensate for inferior plumage by adjusting investment in song. In this case, less attractive males should show enhanced song (amount, complexity), especially when competing with more colourful males. However, male songbirds may be unable to modulate certain properties of song—such as repertoire size and song complexity—that are linked to their age, cognition or developmental conditions [4], and therefore, may primarily adjust the total number of songs they produce [5].

House finches (Haemorhous mexicanus) are songbirds with sexually dichromatic plumage and elaborate male song, both of which are used by females of the species to assess mates [6,7]. Male plumage ranges from yellow to vibrant red, with females preferring red males [7], and males produce song with longer bouts that contain more unique syllables during the breeding season [8]. The extreme variation in plumage colour and their highly social nature make house finches ideal to investigate how investment in a signal in one dimension (plumage colour) affects males' investment in another (song) as a function of perceived competition for mates. We manipulated male plumage coloration in captive house finches through carotenoid supplementation during moult, so that 50% of birds produced red plumage and 50% produced yellow plumage. Males were then housed under three social conditions: (i) all red birds, (ii) all yellow birds or (iii) a mixture of red and yellow birds. We then recorded male song after presentation of a female and measured female attentiveness towards males during song recordings. We predicted that if males can compensate for less attractive plumage, yellow males housed with red males would increase investment in song, while males in groups with similar plumage may not differ in song investment. However, if females show more attentive behaviours towards red males, and/or red males are in superior physiological condition due to the carotenoid manipulation, we would also predict that red males would produce more song than yellow males.

2. Material and methods

(a). Birds and plumage colour manipulation

In summer 2010 and 2011, house finches were captured around Davis, CA, USA and sexed by plumage [7]. Birds were held in aviaries at the University of California, Davis on naturally changing photoperiods. All birds were maintained on a low-carotenoid base diet of Roudybush and sunflower seeds. Prior to moult in June 2012, male birds were randomly assigned to one of two diet treatments (all ad libitum access): (i) low-carotenoid base diet (n = 15, lutein : zeaxanthin, 4 : 1, less than 1 µg g⁻¹) and (ii) a supplemented base diet with ORO GLO® (lutein : zeaxanthin, 15 : 1, 30 µg g⁻¹), and water supplemented with canthaxanthin (35 µg ml⁻¹) (n = 15). The supplemented diet (75 µg g⁻¹) is within the upper range of carotenoid concentrations found in the gut contents of wild house finches [9]. Males were housed in pairs during moult, until the beginning of the experiment, with each pair receiving the same diet. Pair housing standardized access to carotenoids and thus plumage colour, as well as social interactions between birds before the start of the experiment. All birds were returned to the base diet within a month of moult completion and maintained on this diet until song recordings in spring 2013. The diet manipulation resulted in two distinct plumage morphs, hereon referred to as yellow (Y) and red (R) (figure 1a).

Figure 1.

(a) The two plumage colours produced by carotenoid manipulation. (b) Influence of plumage colour and social group upon male song production. (c) Song complexity was unaffected by social group or plumage colour. **p < 0.001. Graphs show mean ± s.e. Shaded bars indicate birds housed under mixed conditions (RM and YM).

(b). Social environment

A month before song recordings (February 2013), all males were moved to flight aviaries (3 m × 2 m × 3 m (h × w × d)) and housed under three social environments; (i) all red (RR, n = 10), (ii) all yellow (YY, n = 10) or (iii) both red and yellow birds (n = 10, 5 yellow (YM) and 5 red (RM)). Social groups were visually but not acoustically isolated from each other. Females were housed in a flight cage in view of the males, so both males and females could see and hear each other prior to song recordings.

(c). Song recording and analysis

Song recordings were conducted in March 2013. A male and female were placed in separate cages immediately beside each other, within a sound attenuated chamber (Industrial Acoustics, Inc.). The following morning at lights-on, we recorded each male's song for 2 h. For 1 h, the female was present and her behaviour recorded; for another hour, she was removed to stimulate male song production [10]. We recorded song using a Sennheiser omnidirectional microphone ME62 suspended above the centre of the male's cage, and a Marantz PMD660 digital recorder. Total duration of song, number of song bouts and song complexity (number of unique syllables per bout) were measured. As the same females were used for multiple males (n = 8), each female was randomly assigned either a red or yellow male for her first recording, and the male plumage colour was alternated for subsequent recordings. Male body condition, fat score and testis mass did not differ between plumage colours during song recordings. For full details, see electronic supplementary material, S1.

(d). Female behaviour

Female behaviour was recorded while she was placed beside the focal male for song recordings, using a Sony Handycam HDR-CX330. Female attentiveness was quantified in two ways. First, we recorded the number of times females performed circular flights [6], defined as when the female perched directly onto the side of the cage closest to the male and then flew in the opposite direction while looking back towards the male. Circular flights were counted for 15 min, after 10 min of acclimatization. Second, we measured the side preference of the female by recording the female's location every 30 s for the same 15 min, to calculate the proportion of time spent in the half of the cage closest to the male.

(e). Statistical analysis

To investigate whether individual plumage colour or social group influenced song production and female behaviour we used LMMs, with female ID as a random factor. Log transformations were used to normalize bout number and song duration data. Bout number and total song duration were significantly correlated (R² = 0.97, p < 0.001) and results were consistent using either variable, therefore, only song duration is reported. Multiple comparison tests (Tukey) were conducted to compare social groups. Some males did not produce any song (n: total = 10, RR: 3, YY: 5, RM: 1, YM: 1), but probability to produce song was unaffected by an individual's plumage colour (binomial GLMM: z = −0.77, p = 0.44) or social group (binomial GLMM: z = 0.41, p = 0.68). Arcsine transformation was used to normalize the proportion of time spent by females in the side of the cage closest to the male. As circular flights were a count, a GLMM with a negative binomial error structure was used. Data were analysed in R (v. 3.3.0, R Core Team, 2016).

3. Results

Red birds sang for significantly longer than yellow birds (χ² = 19.33, p < 0.001). Social group also predicted male song production (χ² = 45.46, p < 0.001). Both RR and RM birds produced more song than YY birds (figure 1b, RR: z = 6.60, p < 0.001, RM: z = 4.46, p = 0.002), but RR and RM birds did not differ from YM birds (figure 1b, RR: z = −1.82, p = 0.26, RM: z = −0.59, p = 0.93). YM birds sang significantly more than YY birds (figure 1b, z = 3.40, p = 0.004). Plumage colour (figure 1c, χ² = 2.14, p = 0.14) and social group (figure 1c, χ² = 3.24, p = 0.36) did not influence song complexity. Females performed more circular flights towards red compared with yellow males (figure 2a, χ² = 5.26, p = 0.02), and females performed fewer circular flights towards YM males compared with RM males (z = −2.85, p = 0.02), but the number of circular flights performed by females did not differ between the other social groups (p > 0.1). The proportion of time a female spent in the side of the cage nearest to the male was not influenced by male plumage colour (figure 2b, χ² = 2.37, p = 0.12) or social group (χ² = 3.08, p = 0.38).

Figure 2.

Influence of male plumage colour upon (a) number of circular flights performed by females and (b) percentage of time spent by female in the half of cage closest to male. *p < 0.05. Graphs show mean ± s.e., n = 8.

4. Discussion

This study provides novel evidence that a male's plumage coloration and the plumage colour of his social competitors influence investment in song. House finch males with red plumage sang for significantly longer than yellow males. Red males may have been responding to female attentiveness, as females performed more circular flights towards red males compared with yellow males, and towards RM males compared with YM males. Female behaviour plays an important role in male song production in many avian species [11–13]. Ultimately, if a female appears to be receptive, it may be worthwhile elevating investment in song to secure a mating. Females spent the majority of their time in the side of the cage closest to the male regardless of male social group or plumage colour, suggesting that this is a less informative measure of female attentiveness in this highly social species. As birds were not acoustically isolated, the increase in circular flights by females could have been caused by the greater song production by red males. Therefore, further work is required to establish the causational direction of this result.

Alternatively, as yellow males received a restricted carotenoid diet during moult, they may have been in inferior physiological condition. Moult is metabolically costly [14] and low carotenoid availability during moult could divert resources from additional cellular processes and cause elevated oxidative stress [2,15]. Recent studies suggest elevated oxidative stress is linked to reduced song output in male birds [16]. This points to the intriguing possibility that carotenoid diet during moult could have carry-over effects upon song performance in the following breeding season.

Males' relative attractiveness also influenced their investment in song. Yellow males housed with more attractive red males increased song production compared with yellow males housed with equally unattractive yellow males. As yellow males experienced similar costs of moult and similar female behaviour, this suggests that it was the males' social group specifically that influenced song production. Therefore, males with less attractive plumage can, at least to a limited extent, upregulate song to optimize their mating opportunities. The higher song production by red males under both social conditions could also be a response to the attractiveness of their competitors. We used a canthaxanthin supplement to produce relatively uniform plumage coloration at the top end of that produced in the wild [7]. Therefore, for males in both RR and mixed conditions, the relative attractiveness of competitors was high. Song complexity, however, was not affected by either plumage colour or social conditions. Male birds of other species also increase total song production, but not song complexity, when exposed to elevated conspecific song competition [5]. Together, these results suggest that some components of song may be less flexible than others.

Social status can modulate song output in male birds, with socially dominant birds, in general, showing greater song investment [17]. In house finches, the intensity of red carotenoid plumage colour is negatively correlated with social dominance, with less ornamented males having a tendency to dominate brighter males [18]. Less ornamented house finches also show greater conspecific aggression [18]. Consequently, yellow males may have experienced more aggressive encounters, causing reduced physiological condition [19] and preventing them from investing maximally in song [20]. Including observations of social interactions between males in future studies would be informative.

Our study suggests that males may flexibly allocate singing effort by responding to their own plumage colour, and to the plumage colour of competing conspecifics. Improving our knowledge of the interactions between multiple signals used for mate attraction, and the mechanisms that regulate them, will improve our understanding of the evolution of multimodal signalling [1].

Ethics

Animals' care was in accordance with institutional guidelines at the University of California-Davis under approved Animal Care Protocol no. 13209.

Data accessibility

Data are available to download from Dryad (http://dx.doi.org/10.5061/dryad.66v3d46) [21].

Authors' contributions

L.J.H. carried out data collection. All authors participated equally in the study design, data analysis and drafting the manuscript. All authors gave final approval for publication and agree to be held accountable for published content.

Competing interests

The authors do not have any competing interests.

Funding

This work was supported by a grant awarded by the National Science Foundation to T.P.H. (award no. IOS-0744705).