- Split View
-
Views
-
Cite
Cite
Miklós Laczi, Dóra Kötél, János Török, Gergely Hegyi, Mutual plumage ornamentation and biparental care: consequences for success in different environments, Behavioral Ecology, Volume 28, Issue 5, September-October 2017, Pages 1359–1368, https://doi.org/10.1093/beheco/arx099
- Share Icon Share
Abstract
According to the good parent and differential allocation models, parental behavior could depend on the individual’s own quality, and it could be adjusted to the coinvestor’s parental care and sexual ornamentation. These investment patterns may interact with environmental conditions and offspring quality in determining reproductive success. Few studies have considered ornament-related own and partner care of both parents and their consequences in relation to environmental conditions. In a brood size manipulation experiment on collared flycatchers (Ficedula albicollis), we measured nestling feeding rates, white patch sizes and plumage reflectance properties of both parents, and quantified nestling growth and reproductive success. We found little relationship between ornamentation and own feeding rate irrespective of manipulation. Parental quality, measured as nestling biomass production per unit feeding effort, was related to male structural plumage brightness in a manipulation-dependent manner. Male wing patch size and the female’s structural plumage brightness were linked to the partner’s feeding rate, and this did not vary with experimental environment. Finally, relationship of prefledging nestling size with male forehead patch size was environment-dependent, and this pattern was apparently due to intrinsic nestling characteristics. Reproductive success only partly reflected these findings. Our results indicate how integrated studies of mutual ornamentation and mutual care with environmental and offspring quality may help us better grasp the selection forces shaping sexual ornaments.
INTRODUCTION
Parental care refers to behaviors which increase the reproductive value of the offspring raised from the current reproductive attempt, and have costs in terms of future reproduction (Williams 1966). Postnatal parental care is one of the central components of life-history and sexual selection studies in iteroparous species (e.g., Székely et al. 2000; Székely et al. 2006). In these species, especially in altricial birds with socially monogamous mating system, the successful rearing of offspring often needs the active care of both parents, and their care contribution is an essential component of reproductive success through lifetime. Because parental care has costs and benefits (Trivers 1972; Clutton-Brock 1991), parental investment is thought to be optimized (Lazarus and Inglis 1986; Daan et al. 1990; Lessells 1993). Optimization means that parents have to divide 1) parental effort among different reproductive bouts (e.g., Stearns 1992), and 2) resources between themselves and their offspring (e.g., Martins and Wright 1993; Garcia et al. 1993).
In species with biparental care, one important question is how parental behavior changes depending on the actual social and environmental conditions. For example, increased predation risk could shape both incubation behavior (Basso and Richner 2015) and nestling provisioning rate (Ghalambor et al. 2013); unfavorable weather conditions may suppress parental investment to ensure the parents’ survival (Öberg et al. 2015) and artificial food supplementation and dispersal status of local breeders may affect the intensity of nest defence (Récapet et al. 2016). Another crucial point for the parent to consider is the degree of the partner’s parental investment. Although both parents acquire benefits by means of offspring, each pair member experiences costs as well (Smiseth et al. 2005; Kosztolányi et al. 2009), so there could be a conflict between parents about how much care they should contribute to the raising of offspring, with each member’s interest being that the partner contributes more (Clutton-Brock 1991).
This dynamic adjustment of investment could be under constraints due to variation in environmental conditions and individual quality. According to the differential allocation hypothesis (Burley 1986; Sheldon 2000), a parent could alter its decision on its current parental investment depending on the partner’s attractiveness. This hypothesis has been supported by many studies, although the results often contradict each other (Ratikainen and Kokko 2010). The principal pattern of variation in the results is that differential allocation could be realized in both negative and positive directions (Ratikainen and Kokko 2010). In the case of positive differential allocation, a parent increases its relative contribution to current reproductive success when mated to an attractive partner (Burley 1986). In contrast to this, in a situation with negative differential allocation, a parent increases its parental contribution when mated to a less preferred partner (Gowaty et al. 2007; Gowaty 2008). Reproductive compensation refers to negative differential allocation (e.g., Harris and Uller 2009), but also to compensation for disadvantageous conditions that are independent of the partner’s attractiveness, for example, handicapped partner, altered rearing environment (e.g., Markman et al. 1995; Ardia 2007; Siefferman and Hill 2007).
Because of the above-mentioned phenomena, it would be beneficial for the individuals to have information about the potential mate’s parental capacity. These abilities are often possible to estimate based on the exaggeration of the secondary sexual characters (Wolf 1997) due to the costs of ornamentation (see Zahavi 1975; Kodric-Brown and Brown 1984). According to the good parent hypothesis (Hoelzer 1989; Price et al. 1993), exaggeration of ornaments is positively related to parental quality, while the parental care–mating trade-off hypothesis predicts the opposite, namely that highly ornamented parents contribute to raising offspring to a lesser degree (Magrath and Komdeur 2003). In birds, one of the most widespread ornaments is the plumage itself, and multiple plumage properties such as patch size or color intensity associate with production or wearing costs (Andersson 1994; McGraw et al. 2002; Hill and McGraw 2006). For example, carotenoid-based ornaments often depend on the environment and advertise the bearer’s physical condition (Olson and Owens 1998; Svensson and Wong 2011), and there are trade-off relationships between physiological functions and ornament expression (Rajasingh et al. 2006). Evidence suggests that melanin-based ornaments could also be costly (Jawor and Breitwisch 2003; McGraw 2003), hence individual variation of these conveys information about, for example, social status (Mennill et al. 2003), ectoparasite load (Fitze and Richner 2002), viability (Emaresi et al. 2014), or health (Galván and Alonso-Alvarez 2008, 2009). Depigmented achromatic feather parts are also linked to several different costs (e.g., Qvarnström 1997; Goldstein et al. 2004), and both the patch size (e.g., Griggio et al. 2010; Purves et al. 2016) and the spectral features (Penteriani et al. 2006; Vágási et al. 2010) may advertise some aspects of quality. Based on the expression of these ornamental traits, a signal receiver could evaluate the potential mate’s parental qualities (reviewed by Møller and Jennions 2001; Hegyi et al. 2015). Most such studies have focused on female or male parental effort in relation to male sexual attractiveness (Hegyi et al. 2015), and the relationships between female ornamentation and female (Linville et al. 1998; Siefferman and Hill 2005; García-Navas et al. 2012) or male parental behavior (Morales et al. 2012; Kötél et al. 2016) are relatively poorely investigated.
A very poorly known phenomenon is the dependence of the ornament-care relationships on environmental conditions. For example, if exaggerated ornamentation indicates individual quality or is accompanied by wearing costs, this may shift the ornament-own care relationship depending on environmental suitability (Hasegawa and Arai 2015). Analogously, if the consequences of parental investment are environment-dependent, this may cause shifts in the direction of differential allocation (Morales et al. 2012). However, we found only 2 studies that have examined parental care in relation to sexual ornamentation in manipulated breeding conditions (Järvistö et al. 2015, Berzins and Dawson 2016).
The collared flycatcher (Ficedula albicollis) is an appropriate model species to study the differential allocation and parental quality indicator hypotheses in both sexes. This species is sexually dichromatic. Males express depigmented breast, collar, forehead, and wing patch, while in females the forehead patch is rare, and there is no collar, furthermore the melanized, brown plumage parts are pheomelanin-dominated in contrast to males, where the melanized parts are black, and contain mostly eumelanin (Calhim et al. 2014). In this species, the information content and life-history correlates of ornamental white plumage patch sizes have been examined in much detail (e.g., Pärt and Qvarnström 1997; Hegyi, Garamszegi, et al. 2008; de Heij et al. 2011), but dark plumage parts have generally been neglected (but see Laczi et al. 2013; Kötél et al. 2016). In our population, roles of both female and male patch sizes in inter- and intrasexual selection are well known (Garamszegi et al. 2006; Hegyi, Garamszegi, et al. 2008; Hegyi, Rosivall, et al. 2008; Hegyi et al. 2010). The degree of exaggeration of white plumage patches has previously been linked to the reproductive investment and body condition during the previous breeding bout (Török et al. 2003; Hegyi, Rosivall, et al. 2008). Nonetheless, recent findings suggest that reflectance attributes (achromatic intensity and relative ultraviolet intensity) of plumage also vary among individuals, and form a partly integrated visual signal system which is related to social mating patterns (Laczi et al. 2011), breeding effort of the previous year (Laczi et al. 2013), and own and partner behavior during incubation (Kötél et al. 2016).
Our aim in this study was to explore how female, male, and total parental effort and their relation to ornamentation are influenced by brood size manipulation in the collared flycatcher. As a measure of parental effort, we used nestling feeding rate. We also quantified the environment-dependent patterns of ornamentation with reproductive success components (mass and size of fledglings, number of young successfully fledged) and with parental quality (fledgling biomass per unit feeding rate). We expected that individuals paired to more ornamented partners would show higher feeding rates, and that condition-dependent ornament expression of the bearer would be positively linked to own parental investment or parental quality (Hegyi et al. 2015). Additionally, we predicted different relationships between ornamentation and own or partner parental care in different brood size manipulation categories, if the environment-dependent costs of care represent a different constraint or a different cost-benefit ratio for parents with different own or partner attractiveness. In our study population, there is evidence for both wearing costs of female and male ornamentation (Garamszegi et al. 2006; Hegyi, Garamszegi, et al. 2008) and context-dependence in the fitness effects of offspring quality (Hegyi, Rosivall, et al. 2011), so it was reasonable to predict environment-dependence in ornament-care correlations. Finally, we examined how ornament-dependent reproductive success is jointly determined by patterns of own and partner care, and whether this picture could be shaped by environmental conditions and further modified by intrinsic nestling attributes potentially linked to parental ornamentation (Kiss et al. 2013).
METHODS
Ethical note
This study was conducted under a long-term research agreement with the Pilis Park Forestry (December 1988 and March 2007), and with research permits from the regional nature conservation authority (KTVF 509-4/2012) and Institutional Animal Care and Use Committee (T-016/2015), and ringing license from the Hungarian Ornithological and Nature Conservation Society (registration number: 128).
Field methods
Data were collected during the breeding seasons of 2015 and 2016 in artificial nest-box plots in deciduous woodland in the Pilis-Visegrád Mountains, Hungary (47°43′N, 19°01′E; for more details, see Török and Tóth 1988). The collared flycatcher is a long-distance migratory, hole-nesting species, which consumes mostly spiders, caterpillars, and flying insects. Females have one clutch (typically 5–8 eggs) per season, which are incubated solely by the female. During the incubation period, males provide supplemental food for their mates (Kötél et al. 2016). After hatching, not only females, but also males take an active part in offspring feeding (Kiss et al. 2013). No extreme weather conditions occurred during the experiment.
We carried out the brood-size manipulation based on the following protocol. We did not use broods of renestings after nest failure, or broods which were involved in other experimental manipulations. Likewise, if we observed partial nest predation on eggs, or not all of the eggs hatched in a nest, we excluded the brood from the experiment. We created trios (10 in 2015, 9 in 2016) of suitable broods with the same clutch size (with 6 or 7 eggs) and hatching date. Nest-boxes were checked daily around the expected hatching in order to determine hatching date exactly. Our experimental nests were randomly located across the central study plots. We used broods of both subadult and adult males. When nestlings were 2 days old, we partially cross-fostered 2 broods in each trio. We transferred 2 randomly selected offspring from one brood (enlarged) into the other, and reciprocally, from the latter one (reduced) we transferred 4 randomly selected offspring into the former. Thus, nestling swapping resulted in reduced and enlarged clutches with approximately half of the offspring in each brood being foreign and half own (Merilä 1997). The reciprocal swapping took no more than 15 min. As a control, we used unmanipulated nests without any cross-fostering. The results suggest no effect of the nestling swapping per se, similarly to a previous, large-scale study using the same protocol (Török et al. 2004). At the time of cross-fostering, we marked nestlings individually by plucking tufts of natal down feathers, and measured their body mass using a Pesola spring balance (to the nearest 0.1 g). There were no differences between the 3 groups of nests in 2-day-old nestling body mass (results not shown here). Eight-day-old nestlings were ringed and weighed using a Pesola spring balance (to the nearest 0.1 g). At 10 days of offspring age, we took video recordings of parental activity and then caught and measured the parents (see below). At 12 days of offspring age, we weighed the nestlings again, and also measured their right tarsus using a sliding calliper (to the nearest 0.1 mm). We finally estimated the number of fledglings from nest checking data after 12 days of age. We excluded data of 1) nests that suffered from predation after manipulation, 2) supposedly uniparental nests (based on video recording data and capturing failure of one parent), 3) nests with a polygynous male, and 4) nests where one of the parents had been brood-size manipulated in the previous year. In total, we used 30 nests from 2015 and 21 from 2016. Missing color data from parents further restricted the datasets in a sex-specific way.
To record the parental activity of females and males, we placed out a digital camera (Panasonic HC-V100EP-K Full HD) 0.5–1 m above ground on a tripod, at least 10 m from the nest-box. The recordings were taken between 0900 and 1830, mostly during the first half of the day, and avoiding midday hours (1200–1400). In our population, there is no difference between morning and afternoon feeding rate of females and males (Kiss et al. 2013). We also avoided taking records during precipitation as rainfall could strongly influence food searching capabilities and thus feeding behavior (e.g., Radford et al. 2001). In order to see if the arriving birds carried any food in their beak, we avoided back-light, and the camera was at an angle to the nest-box. The presence of the camera did not disturb feeding activity as we did not observe enduring presence of signs of disturbance, and actually a notable proportion of birds were seen delivering food for the chicks even during placing out the camera. Although length of each recording was a few (10–15) minutes more than 1 h, we did not use the first 10–15 min in order to exclude potentially disturbed behavior from data extraction. Thus, we analyzed 1-h-long records. Based on a previous correlative study, male and female feeding rate is correlated between 2 consecutive hours of recording (N = 54 broods, r = 0.509 and P < 0.001 for females, r = 0.298 and P = 0.029 for males), so 1 h provides a representative sample of care in this population. When processing the video recordings, we noted the timing of all feeding events (to the nearest 1 s). We calculated the feeding rate of each sex (i.e., number of visits with food delivery per hour), and the total feeding rate summing female and male feeding rates per hour. Feeding frequencies per hour ranged from 10 to 57 in females (mean = 25.52, SD = 10.83) and 5 to 48 in males (mean = 24.75, SD = 9.23).
When nestlings were 10 days old, after the video recording we captured the parents in their nest-box. With respect to age, birds were classified as subadults (yearlings) or adults (>1 year) based on wing patch size and the darkness of primaries in males (Svensson 1992), and ringing data in females (unknown first breeders were classified as yearlings; Hegyi, Rosivall, et al. 2008). Wing patch size of females and males was estimated as the sum of the visible lengths of white areas on the outer vanes of the fourth to eighth primaries on the right wing, using a sliding calliper (to the nearest 0.1 mm) (Török et al. 2003). We measured the maximum height and width of the forehead patch of males (to the nearest 0.1 mm) using a sliding calliper, and calculated forehead patch size by multiplying the 2 measures (Hegyi et al. 2002).
Color measurements
We carried out the spectral measurements in the field. There were no molting individuals in our sample. We took spectra from the following 5 plumage areas: forehead (brownish in most females, white in males), crown (brown in females, black in males), wing patch (white), wing coverts (brown in females and in subadult males, black in adult males), breast (whitish). We performed these measurements using a portable USB2000 UV-VIS spectrometer (sensory range: 179–877 nm; Ocean Optics Europe) with a DH-2000 deuterium-halogen light source (Ocean Optics Europe) and a QR400-7-SR-BX sensor (Ocean Optics Europe). Around the light collecting end of the sensor we fitted a matte black plastic sheath in order to exclude incoming ambient light from the measured area and to standardize measuring distance from the feather surface (3 mm). The sensor was oriented at a 90° angle to the surface. Each plumage area was measured twice consecutively with removing the sensor from the surface between 2 recordings. The detector counted incoming photons (i.e., absolute reflectance, R) at 0.37 nm wavelength steps with 15 ms integration times. Percentage relative intensity scores were calculated for every wavelength by OOIBase32 Spectrometer Operating Software (Ocean Optics Europe) based on the following formula: [(Rsample-Rdark)/(Rwhite-Rdark)] × 100. Measuring of dark signal (i.e., putting the sensor in a spacy box with no incoming light) is necessary (Andersson and Prager 2006) in order to correct for electric noise and for potential sheath reflectance. A WS-1–SS diffuse reflectance standard (Ocean Optics Europe) served as white signal. Both the dark and white signals were remeasured regularly. To achieve better signal-to-noise ratio, we set the “Boxcar” function to 10, and the “Average” function to 10. Before computing any derived spectral variable, it was necessary to remove from each spectrum a significant spectral artifact characteristic to the light source (Dα peak).
We generated 2 objective variables from the reflectance spectra. Brightness was calculated as the average intensity between 320 and 700 nm (Woodcock et al. 2005; Laczi et al. 2011). This variable describes the achromatic intensity of the area. UV chroma or relative UV intensity denotes average UV-A intensity (between 320 and 400 nm) divided by brightness, that is, saturation. Our plumage spectral measurement repeatabilities are high (Pearson correlations calculated for each plumage area separately in the 3 different plumage types, females, yearling males, and older males; brightness: 0.615–0.863; UV chroma: 0.543–0.909). Therefore, we used the average of the measurements of each area of each individual.
Statistical analyses
All analyses were performed in Statistica 8.0 (StatSoft, Inc.). In both females and males, we conducted principal component (PC) analyses separately on brightness variables and UV chroma variables of the different plumage areas (see Laczi et al. 2013). As in previous studies (Laczi et al. 2011, 2013, Kötél et al. 2016), these analyses were done in the whole long-term spectral dataset, because spectral correlation structure is statistically stable among years (Laczi et al. 2011) and long-term data provide more reliable output for multivariate analyses. For the subsequent analyses, we used only PCs with eigenvalues greater than one (the Kaiser criterion). Based on the analyses of age and year effects on the investigated variables, male wing patch size and brightness PC2 were standardized for age (to a mean of 0 and SD of 1), and male UV chroma PC was also standardized for subadults as a whole, and for year in adults. Using general linear models (GLMs) with backward stepwise model selection, we tested the effects of manipulation (reduced, control, enlarged) and year (as well as their interaction) on brood size, fledgling number, 8-day-old and 12-day-old average nestling body mass, 12-day-old average nestling tarsus length, condition of parents, female and male feeding rate and parental quality (see below), and total feeding rate. In the cases of parent condition, body mass was the dependent variable and tarsus length included as an additional predictor variable, furthermore we performed GLMs of feeding rates with and without including within-group centered values of brood size as a covariate (calculated by subtracting the sample mean from all values of the given treatment category). We compiled the models with color variables in accordance with the results of these preliminary models. We investigated the relationships of female and male coloration with own, pair and total feeding rate, fledgling number, 8-day-old and 12-day-old offspring mass, and 12-day-old offspring tarsus length. Feeding rate increases with manipulation in both parents (see Results) and is negatively related to body condition irrespective of sex and even within manipulation groups (results not shown here), but it is unrelated to any aspect of reproductive output in this dataset (not shown here). Therefore, feeding rate is an appropriate measure of parental effort but an inappropriate measure of parental quality in our population. Accordingly, we also used a direct metric of parental quality, calculated as the production of fledgling biomass (total 12-day-old mass of the brood) divided by the feeding rate of the given parent (and log transformed). In models for females, color PCs and wing patch size were used as continuous predictor variables, while one of the other mentioned variables was the dependent variable. Manipulation was used as categorical predictor, and we also included in the models the 2-way interactions of manipulation with color PCs and patch sizes. In the case of GLMs where one of the feeding rates was the dependent variable, we further used the centered values of brood size as an additional independent variable. In the GLMs for males, model structure was the same as above, except that we also included forehead patch size and the treatment × forehead patch size interaction. Post hoc tests were performed by Tukey’s method. Our findings concerning 8-day-old nestling body mass are not shown because these are qualitatively the same as those of the 12-day-old nestling body mass. Results of models with centered brood size are also not shown here, because the results were qualitatively the same as without this variable. We tested the normality of the frequency distributions of data and the GLM model residuals and found them appropriate.
RESULTS
Female brightness and UV chroma variables were compressed into one PC. Females with higher brightness PC or UV chroma PC score expressed higher brightness or UV chroma across the plumage. In the case of males, plumage-level brightness variance was represented in 2 PCs, of which brightness PC1 reflected general brightness while PC2 reflected wing covert brightness (see Table 1). Males with higher values of the single UV chroma PC displayed higher UV chroma across the respective plumage areas.
. | Female . | Male . | |||
---|---|---|---|---|---|
. | Brightness PC . | UV chroma PC . | Brightness PC1 . | Brightness PC2 . | UV chroma PC . |
Forehead | 0.58 | 0.76 | 0.72 | −0.11 | 0.55 |
Crown | 0.76 | 0.74 | 0.57 | 0.35 | 0.61 |
Wing patch | 0.65 | 0.57 | 0.77 | −0.32 | 0.56 |
Wing coverts | 0.49 | 0.58 | 0.14 | 0.91 | 0.59 |
Breast | 0.70 | 0.45 | 0.79 | 0.01 | 0.62 |
. | Female . | Male . | |||
---|---|---|---|---|---|
. | Brightness PC . | UV chroma PC . | Brightness PC1 . | Brightness PC2 . | UV chroma PC . |
Forehead | 0.58 | 0.76 | 0.72 | −0.11 | 0.55 |
Crown | 0.76 | 0.74 | 0.57 | 0.35 | 0.61 |
Wing patch | 0.65 | 0.57 | 0.77 | −0.32 | 0.56 |
Wing coverts | 0.49 | 0.58 | 0.14 | 0.91 | 0.59 |
Breast | 0.70 | 0.45 | 0.79 | 0.01 | 0.62 |
The principal component analyses were conducted separately for sexes and for spectral variable types (brightness, UV chroma).
. | Female . | Male . | |||
---|---|---|---|---|---|
. | Brightness PC . | UV chroma PC . | Brightness PC1 . | Brightness PC2 . | UV chroma PC . |
Forehead | 0.58 | 0.76 | 0.72 | −0.11 | 0.55 |
Crown | 0.76 | 0.74 | 0.57 | 0.35 | 0.61 |
Wing patch | 0.65 | 0.57 | 0.77 | −0.32 | 0.56 |
Wing coverts | 0.49 | 0.58 | 0.14 | 0.91 | 0.59 |
Breast | 0.70 | 0.45 | 0.79 | 0.01 | 0.62 |
. | Female . | Male . | |||
---|---|---|---|---|---|
. | Brightness PC . | UV chroma PC . | Brightness PC1 . | Brightness PC2 . | UV chroma PC . |
Forehead | 0.58 | 0.76 | 0.72 | −0.11 | 0.55 |
Crown | 0.76 | 0.74 | 0.57 | 0.35 | 0.61 |
Wing patch | 0.65 | 0.57 | 0.77 | −0.32 | 0.56 |
Wing coverts | 0.49 | 0.58 | 0.14 | 0.91 | 0.59 |
Breast | 0.70 | 0.45 | 0.79 | 0.01 | 0.62 |
The principal component analyses were conducted separately for sexes and for spectral variable types (brightness, UV chroma).
Brood size after manipulation significantly differed between groups (see details in Table 2), that is, reduced nests contained less offspring, while enlarged ones contained more offspring (all post hoc P < 0.0001), but we did not observe year effect or interaction with manipulation. We detected similar patterns for fledgling number (manipulation: all post hoc P < 0.001). Twelve-day-old nestling mass differed only between reduced and enlarged nests with lesser mass in the latter (post hoc P = 0.003, nonsignificant (NS) for other comparisons). Control nests were intermediate between the two. Female, male, and total feeding rates were elevated in enlarged broods in comparison to reduced and control ones (all post hoc P < 0.007, NS for other comparisons). In females, treatment significantly influenced condition, with females rearing reduced and enlarged broods significantly differing from each other (post hoc P = 0.002, NS for other comparisons). Females with enlarged brood were in worse condition. There was also a weaker effect of year on female condition but no interaction between year and manipulation. Finally, our measure of parental quality did not vary with year, manipulation or their interaction in either sex.
. | Manipulation . | Year . | . | Manipulation × year . | Reduced . | . | Control . | . | Enlarged . | . | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | Mean . | SE . | Mean . | SE . | Mean . | SE . |
Brood size | 2, 48 | 164.40*** | 1, 47 | 1.06 | 2, 45 | 1.56 | 4.47 | 0.16 | 6.33 | 0.15 | 8.56 | 0.16 |
Female feeding rate | 2, 45 | 15.81*** | 1, 44 | 1.88 | 2, 42 | 0.92 | 17.50 | 2.12 | 24.75 | 2.12 | 34.31 | 2.12 |
Male feeding rate | 2, 45 | 15.36*** | 1, 44 | 2.70 | 2, 42 | 0.54 | 19.31 | 1.82 | 22.13 | 1.82 | 32.81 | 1.82 |
Total feeding rate | 2, 45 | 23.71*** | 1, 44 | 3.55 | 2, 42 | 1.17 | 36.81 | 3.17 | 46.88 | 3.17 | 67.13 | 3.17 |
Female condition | 2, 44 | 6.57** | 1, 44 | 4.40* | 2, 42 | 0.51 | 3.42 | 1.44 | 1.05 | 1.31 | −3.61 | 1.39 |
Male condition | 2, 47 | 2.36 | 1, 46 | 0.00 | 2, 44 | 0.73 | 1.57 | 1.41 | 0.64 | 1.37 | −2.60 | 1.45 |
Female parental quality | 2, 40 | 1.83 | 1, 39 | 1.76 | 2, 37 | 1.07 | 1.56 | 0.05 | 1.50 | 0.05 | 1.44 | 0.04 |
Male parental quality | 2, 40 | 0.84 | 1, 39 | 1.02 | 2, 37 | 1.01 | 1.54 | 0.04 | 1.51 | 0.04 | 1.47 | 0.04 |
12-day-old body mass | 2, 42 | 5.11* | 1, 41 | 0.14 | 2, 39 | 2.49 | 138.20 | 2.50 | 133.51 | 2.50 | 126.95 | 2.50 |
12-day-old tarsus length | 2, 42 | 3.46* | 1, 41 | 0.58 | 2, 39 | 1.87 | 180.22 | 0.99 | 178.16 | 0.99 | 176.52 | 0.99 |
Fledgling number | 2, 42 | 52.54*** | 1, 41 | 0.76 | 2, 39 | 1.04 | 4.27 | 0.23 | 5.73 | 0.23 | 7.60 | 0.23 |
. | Manipulation . | Year . | . | Manipulation × year . | Reduced . | . | Control . | . | Enlarged . | . | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | Mean . | SE . | Mean . | SE . | Mean . | SE . |
Brood size | 2, 48 | 164.40*** | 1, 47 | 1.06 | 2, 45 | 1.56 | 4.47 | 0.16 | 6.33 | 0.15 | 8.56 | 0.16 |
Female feeding rate | 2, 45 | 15.81*** | 1, 44 | 1.88 | 2, 42 | 0.92 | 17.50 | 2.12 | 24.75 | 2.12 | 34.31 | 2.12 |
Male feeding rate | 2, 45 | 15.36*** | 1, 44 | 2.70 | 2, 42 | 0.54 | 19.31 | 1.82 | 22.13 | 1.82 | 32.81 | 1.82 |
Total feeding rate | 2, 45 | 23.71*** | 1, 44 | 3.55 | 2, 42 | 1.17 | 36.81 | 3.17 | 46.88 | 3.17 | 67.13 | 3.17 |
Female condition | 2, 44 | 6.57** | 1, 44 | 4.40* | 2, 42 | 0.51 | 3.42 | 1.44 | 1.05 | 1.31 | −3.61 | 1.39 |
Male condition | 2, 47 | 2.36 | 1, 46 | 0.00 | 2, 44 | 0.73 | 1.57 | 1.41 | 0.64 | 1.37 | −2.60 | 1.45 |
Female parental quality | 2, 40 | 1.83 | 1, 39 | 1.76 | 2, 37 | 1.07 | 1.56 | 0.05 | 1.50 | 0.05 | 1.44 | 0.04 |
Male parental quality | 2, 40 | 0.84 | 1, 39 | 1.02 | 2, 37 | 1.01 | 1.54 | 0.04 | 1.51 | 0.04 | 1.47 | 0.04 |
12-day-old body mass | 2, 42 | 5.11* | 1, 41 | 0.14 | 2, 39 | 2.49 | 138.20 | 2.50 | 133.51 | 2.50 | 126.95 | 2.50 |
12-day-old tarsus length | 2, 42 | 3.46* | 1, 41 | 0.58 | 2, 39 | 1.87 | 180.22 | 0.99 | 178.16 | 0.99 | 176.52 | 0.99 |
Fledgling number | 2, 42 | 52.54*** | 1, 41 | 0.76 | 2, 39 | 1.04 | 4.27 | 0.23 | 5.73 | 0.23 | 7.60 | 0.23 |
General linear models with stepwise backward selection and reintroduction. *P < 0.05, **P < 0.01, ***P < 0.001.
. | Manipulation . | Year . | . | Manipulation × year . | Reduced . | . | Control . | . | Enlarged . | . | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | Mean . | SE . | Mean . | SE . | Mean . | SE . |
Brood size | 2, 48 | 164.40*** | 1, 47 | 1.06 | 2, 45 | 1.56 | 4.47 | 0.16 | 6.33 | 0.15 | 8.56 | 0.16 |
Female feeding rate | 2, 45 | 15.81*** | 1, 44 | 1.88 | 2, 42 | 0.92 | 17.50 | 2.12 | 24.75 | 2.12 | 34.31 | 2.12 |
Male feeding rate | 2, 45 | 15.36*** | 1, 44 | 2.70 | 2, 42 | 0.54 | 19.31 | 1.82 | 22.13 | 1.82 | 32.81 | 1.82 |
Total feeding rate | 2, 45 | 23.71*** | 1, 44 | 3.55 | 2, 42 | 1.17 | 36.81 | 3.17 | 46.88 | 3.17 | 67.13 | 3.17 |
Female condition | 2, 44 | 6.57** | 1, 44 | 4.40* | 2, 42 | 0.51 | 3.42 | 1.44 | 1.05 | 1.31 | −3.61 | 1.39 |
Male condition | 2, 47 | 2.36 | 1, 46 | 0.00 | 2, 44 | 0.73 | 1.57 | 1.41 | 0.64 | 1.37 | −2.60 | 1.45 |
Female parental quality | 2, 40 | 1.83 | 1, 39 | 1.76 | 2, 37 | 1.07 | 1.56 | 0.05 | 1.50 | 0.05 | 1.44 | 0.04 |
Male parental quality | 2, 40 | 0.84 | 1, 39 | 1.02 | 2, 37 | 1.01 | 1.54 | 0.04 | 1.51 | 0.04 | 1.47 | 0.04 |
12-day-old body mass | 2, 42 | 5.11* | 1, 41 | 0.14 | 2, 39 | 2.49 | 138.20 | 2.50 | 133.51 | 2.50 | 126.95 | 2.50 |
12-day-old tarsus length | 2, 42 | 3.46* | 1, 41 | 0.58 | 2, 39 | 1.87 | 180.22 | 0.99 | 178.16 | 0.99 | 176.52 | 0.99 |
Fledgling number | 2, 42 | 52.54*** | 1, 41 | 0.76 | 2, 39 | 1.04 | 4.27 | 0.23 | 5.73 | 0.23 | 7.60 | 0.23 |
. | Manipulation . | Year . | . | Manipulation × year . | Reduced . | . | Control . | . | Enlarged . | . | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | Mean . | SE . | Mean . | SE . | Mean . | SE . |
Brood size | 2, 48 | 164.40*** | 1, 47 | 1.06 | 2, 45 | 1.56 | 4.47 | 0.16 | 6.33 | 0.15 | 8.56 | 0.16 |
Female feeding rate | 2, 45 | 15.81*** | 1, 44 | 1.88 | 2, 42 | 0.92 | 17.50 | 2.12 | 24.75 | 2.12 | 34.31 | 2.12 |
Male feeding rate | 2, 45 | 15.36*** | 1, 44 | 2.70 | 2, 42 | 0.54 | 19.31 | 1.82 | 22.13 | 1.82 | 32.81 | 1.82 |
Total feeding rate | 2, 45 | 23.71*** | 1, 44 | 3.55 | 2, 42 | 1.17 | 36.81 | 3.17 | 46.88 | 3.17 | 67.13 | 3.17 |
Female condition | 2, 44 | 6.57** | 1, 44 | 4.40* | 2, 42 | 0.51 | 3.42 | 1.44 | 1.05 | 1.31 | −3.61 | 1.39 |
Male condition | 2, 47 | 2.36 | 1, 46 | 0.00 | 2, 44 | 0.73 | 1.57 | 1.41 | 0.64 | 1.37 | −2.60 | 1.45 |
Female parental quality | 2, 40 | 1.83 | 1, 39 | 1.76 | 2, 37 | 1.07 | 1.56 | 0.05 | 1.50 | 0.05 | 1.44 | 0.04 |
Male parental quality | 2, 40 | 0.84 | 1, 39 | 1.02 | 2, 37 | 1.01 | 1.54 | 0.04 | 1.51 | 0.04 | 1.47 | 0.04 |
12-day-old body mass | 2, 42 | 5.11* | 1, 41 | 0.14 | 2, 39 | 2.49 | 138.20 | 2.50 | 133.51 | 2.50 | 126.95 | 2.50 |
12-day-old tarsus length | 2, 42 | 3.46* | 1, 41 | 0.58 | 2, 39 | 1.87 | 180.22 | 0.99 | 178.16 | 0.99 | 176.52 | 0.99 |
Fledgling number | 2, 42 | 52.54*** | 1, 41 | 0.76 | 2, 39 | 1.04 | 4.27 | 0.23 | 5.73 | 0.23 | 7.60 | 0.23 |
General linear models with stepwise backward selection and reintroduction. *P < 0.05, **P < 0.01, ***P < 0.001.
Detailed results of feeding rate, parental quality, and success in relation to male and female ornamentation are shown in Tables 3 and 4 respectively. There were no significant relationships between female or male feeding rate and own ornamentation regardless of manipulation. Our parental quality metric showed no robust pattern with female ornamentation, but a marginally nonsignificant (P = 0.062) interaction with male brightness PC1. The effect of brightness PC1 was marginally positive in the reduced group (F1,12 = 4.52, P = 0.055, effect size r = 0.52, Lower 95% confidence interval (CIL) = −0.01, Upper 95% confidence interval (CIU) = 0.82) while it was nonsignificantly positive in the control group (F1,10 = 1.56, P = 0.240, effect size r = 0.34, CIL = −0.29, CIU = 0.76) and nonsignificantly negative in the enlarged group (F1,12 = 2.05, P = 0.178, effect size r = −0.38, CIL = −0.76, CIU = 0.19). That is, the relationship decreased with increasing manipulation. When looking at partner feeding rates, female feeding rate showed a marginally nonsignificant (P = 0.050) positive relationship with male wing patch size (effect size r = 0.30, CIL = 0.008, CIU = 0.55; Figure 1), while male feeding rate associated significantly and negatively to female brightness PC (P = 0.041, effect size r = −0.32, CIL = −0.56, CIU = −0.02; Figure 2). Ornamentation showed no significant pattern with 12-day-old nestling body mass irrespective of manipulation. Concerning 12-day-old nestling tarsus length, we found a significant interaction between treatment and male forehead patch size (Figure 3). The effect of forehead patch size was marginally negative in the reduced group (F1,13 = 4.38, P = 0.056, effect size r = −0.50, CIL = −0.81, CIU = 0.01), nonsignificant in the control group (F1,11 = 0.35, P = 0.566, effect size r = 0.16, CIL = −0.42, CIU = 0.66), and significantly positive in the enlarged group (F1,12 = 5.79, P = 0.033, effect size r = 0.57, CIL = 0.06, CIU = 0.85). Repeating the analysis on the own and the foreign nestlings of the cross-fostered broods (reduced and enlarged) showed that the manipulation-dependent forehead patch size effect was strong in the own nestlings (interaction F1,25 = 11.01, P = 0.003, forehead patch size: reduced F1,13 = 7.89, P = 0.01, effect size r = −0.61, CIL = −0.86, CIU = −0.15, enlarged F1,12 = 2.87, P = 0.12, effect size r = 0.44, CIL = −0.12, CIU = 0.79) but nonsignificant in the foster nestlings (interaction F1,25 = 2.51, P = 0.13). Finally, the number of young fledged was related negatively to female wing patch size (P = 0.008, effect size r = −0.43, CIL = −0.65, CIU = −0.13; Figure 4) and marginally (P = 0.066) positively to male forehead patch size (effect size r = 0.29, CIL = −0.01, CIU = 0.55).
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12d . | Nestling mass 12d . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 42 | 16.29*** | 2, 42 | 14.40*** | 2, 36 | 3.17† | 2, 39 | 5.23** | 2, 39 | 61.02*** |
Forehead patch size | 1, 41 | 0.14 | 1, 41 | 0.01 | 1, 36 | 0.08 | 1, 38 | 1.48 | 1, 38 | 3.58 |
Wing patch size | 1, 41 | 0.77 | 1, 41 | 4.08† | 1, 35 | 0.31 | 1, 38 | 1.78 | 1, 38 | 0.00 |
Brightness PC1 | 1, 41 | 0.22 | 1, 41 | 0.00 | 1, 35 | 0.48 | 1, 38 | 3.10 | 1, 38 | 0.16 |
Brightness PC2 | 1, 41 | 2.51 | 1, 41 | 0.13 | 1, 35 | 2.03 | 1, 38 | 0.25 | 1, 38 | 2.08 |
UV PC1 | 1, 41 | 1.52 | 1, 41 | 2.04 | 1, 35 | 1.41 | 1, 38 | 0.60 | 1, 38 | 0.22 |
Manipulation × forehead patch size | 2, 39 | 0.67 | 2, 39 | 2.23 | 2, 36 | 4.20* | 2, 36 | 1.40 | 2, 36 | 2.21 |
Manipulation × wing patch size | 2, 39 | 1.28 | 2, 39 | 0.99 | 2, 33 | 0.18 | 2, 36 | 1.40 | 2, 36 | 0.27 |
Manipulation × brightness PC1 | 2, 39 | 1.88 | 2, 39 | 1.03 | 2, 33 | 0.17 | 2, 36 | 0.33 | 2, 36 | 0.10 |
Manipulation × brightness PC2 | 2, 39 | 0.07 | 2, 39 | 2.02 | 2, 33 | 0.26 | 2, 36 | 0.17 | 2, 36 | 0.67 |
Manipulation × UV PC1 | 2, 39 | 0.37 | 2, 39 | 0.90 | 2, 33 | 0.84 | 2, 36 | 1.00 | 2, 36 | 0.57 |
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12d . | Nestling mass 12d . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 42 | 16.29*** | 2, 42 | 14.40*** | 2, 36 | 3.17† | 2, 39 | 5.23** | 2, 39 | 61.02*** |
Forehead patch size | 1, 41 | 0.14 | 1, 41 | 0.01 | 1, 36 | 0.08 | 1, 38 | 1.48 | 1, 38 | 3.58 |
Wing patch size | 1, 41 | 0.77 | 1, 41 | 4.08† | 1, 35 | 0.31 | 1, 38 | 1.78 | 1, 38 | 0.00 |
Brightness PC1 | 1, 41 | 0.22 | 1, 41 | 0.00 | 1, 35 | 0.48 | 1, 38 | 3.10 | 1, 38 | 0.16 |
Brightness PC2 | 1, 41 | 2.51 | 1, 41 | 0.13 | 1, 35 | 2.03 | 1, 38 | 0.25 | 1, 38 | 2.08 |
UV PC1 | 1, 41 | 1.52 | 1, 41 | 2.04 | 1, 35 | 1.41 | 1, 38 | 0.60 | 1, 38 | 0.22 |
Manipulation × forehead patch size | 2, 39 | 0.67 | 2, 39 | 2.23 | 2, 36 | 4.20* | 2, 36 | 1.40 | 2, 36 | 2.21 |
Manipulation × wing patch size | 2, 39 | 1.28 | 2, 39 | 0.99 | 2, 33 | 0.18 | 2, 36 | 1.40 | 2, 36 | 0.27 |
Manipulation × brightness PC1 | 2, 39 | 1.88 | 2, 39 | 1.03 | 2, 33 | 0.17 | 2, 36 | 0.33 | 2, 36 | 0.10 |
Manipulation × brightness PC2 | 2, 39 | 0.07 | 2, 39 | 2.02 | 2, 33 | 0.26 | 2, 36 | 0.17 | 2, 36 | 0.67 |
Manipulation × UV PC1 | 2, 39 | 0.37 | 2, 39 | 0.90 | 2, 33 | 0.84 | 2, 36 | 1.00 | 2, 36 | 0.57 |
General linear models with stepwise backward selection and reintroduction. Most of the ornament measures are standardized, see Methods for details. *P < 0.05, **P < 0.01, ***P < 0.001, †P < 0.06.
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12d . | Nestling mass 12d . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 42 | 16.29*** | 2, 42 | 14.40*** | 2, 36 | 3.17† | 2, 39 | 5.23** | 2, 39 | 61.02*** |
Forehead patch size | 1, 41 | 0.14 | 1, 41 | 0.01 | 1, 36 | 0.08 | 1, 38 | 1.48 | 1, 38 | 3.58 |
Wing patch size | 1, 41 | 0.77 | 1, 41 | 4.08† | 1, 35 | 0.31 | 1, 38 | 1.78 | 1, 38 | 0.00 |
Brightness PC1 | 1, 41 | 0.22 | 1, 41 | 0.00 | 1, 35 | 0.48 | 1, 38 | 3.10 | 1, 38 | 0.16 |
Brightness PC2 | 1, 41 | 2.51 | 1, 41 | 0.13 | 1, 35 | 2.03 | 1, 38 | 0.25 | 1, 38 | 2.08 |
UV PC1 | 1, 41 | 1.52 | 1, 41 | 2.04 | 1, 35 | 1.41 | 1, 38 | 0.60 | 1, 38 | 0.22 |
Manipulation × forehead patch size | 2, 39 | 0.67 | 2, 39 | 2.23 | 2, 36 | 4.20* | 2, 36 | 1.40 | 2, 36 | 2.21 |
Manipulation × wing patch size | 2, 39 | 1.28 | 2, 39 | 0.99 | 2, 33 | 0.18 | 2, 36 | 1.40 | 2, 36 | 0.27 |
Manipulation × brightness PC1 | 2, 39 | 1.88 | 2, 39 | 1.03 | 2, 33 | 0.17 | 2, 36 | 0.33 | 2, 36 | 0.10 |
Manipulation × brightness PC2 | 2, 39 | 0.07 | 2, 39 | 2.02 | 2, 33 | 0.26 | 2, 36 | 0.17 | 2, 36 | 0.67 |
Manipulation × UV PC1 | 2, 39 | 0.37 | 2, 39 | 0.90 | 2, 33 | 0.84 | 2, 36 | 1.00 | 2, 36 | 0.57 |
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12d . | Nestling mass 12d . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 42 | 16.29*** | 2, 42 | 14.40*** | 2, 36 | 3.17† | 2, 39 | 5.23** | 2, 39 | 61.02*** |
Forehead patch size | 1, 41 | 0.14 | 1, 41 | 0.01 | 1, 36 | 0.08 | 1, 38 | 1.48 | 1, 38 | 3.58 |
Wing patch size | 1, 41 | 0.77 | 1, 41 | 4.08† | 1, 35 | 0.31 | 1, 38 | 1.78 | 1, 38 | 0.00 |
Brightness PC1 | 1, 41 | 0.22 | 1, 41 | 0.00 | 1, 35 | 0.48 | 1, 38 | 3.10 | 1, 38 | 0.16 |
Brightness PC2 | 1, 41 | 2.51 | 1, 41 | 0.13 | 1, 35 | 2.03 | 1, 38 | 0.25 | 1, 38 | 2.08 |
UV PC1 | 1, 41 | 1.52 | 1, 41 | 2.04 | 1, 35 | 1.41 | 1, 38 | 0.60 | 1, 38 | 0.22 |
Manipulation × forehead patch size | 2, 39 | 0.67 | 2, 39 | 2.23 | 2, 36 | 4.20* | 2, 36 | 1.40 | 2, 36 | 2.21 |
Manipulation × wing patch size | 2, 39 | 1.28 | 2, 39 | 0.99 | 2, 33 | 0.18 | 2, 36 | 1.40 | 2, 36 | 0.27 |
Manipulation × brightness PC1 | 2, 39 | 1.88 | 2, 39 | 1.03 | 2, 33 | 0.17 | 2, 36 | 0.33 | 2, 36 | 0.10 |
Manipulation × brightness PC2 | 2, 39 | 0.07 | 2, 39 | 2.02 | 2, 33 | 0.26 | 2, 36 | 0.17 | 2, 36 | 0.67 |
Manipulation × UV PC1 | 2, 39 | 0.37 | 2, 39 | 0.90 | 2, 33 | 0.84 | 2, 36 | 1.00 | 2, 36 | 0.57 |
General linear models with stepwise backward selection and reintroduction. Most of the ornament measures are standardized, see Methods for details. *P < 0.05, **P < 0.01, ***P < 0.001, †P < 0.06.
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12 days . | Nestling mass 12 days . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 40 | 13.68*** | 2, 41 | 12.77*** | 2, 37 | 5.03* | 2, 37 | 3.82* | 2, 36 | 46.67*** |
Wing patch size | 1, 39 | 1.59 | 1, 40 | 0.48 | 1, 36 | 0.67 | 1, 36 | 0.95 | 1, 36 | 7.99** |
Brightness PC1 | 1, 39 | 4.48* | 1, 40 | 0.07 | 1, 36 | 0.00 | 1, 36 | 0.04 | 1, 35 | 1.82 |
UV PC1 | 1, 39 | 0.21 | 1, 40 | 0.29 | 1, 36 | 0.03 | 1, 36 | 0.00 | 1, 35 | 2.03 |
Manipulation × wing patch size | 2, 37 | 1.67 | 2, 38 | 1.12 | 2, 34 | 2.23 | 2, 34 | 1.35 | 2, 33 | 2.18 |
Manipulation × brightness PC1 | 2, 37 | 0.31 | 2, 38 | 1.48 | 2, 34 | 0.80 | 2, 34 | 2.14 | 2, 33 | 2.13 |
Manipulation × UV PC1 | 2, 37 | 0.29 | 2, 38 | 0.07 | 2, 34 | 1.09 | 2, 34 | 1.75 | 2, 33 | 0.49 |
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12 days . | Nestling mass 12 days . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 40 | 13.68*** | 2, 41 | 12.77*** | 2, 37 | 5.03* | 2, 37 | 3.82* | 2, 36 | 46.67*** |
Wing patch size | 1, 39 | 1.59 | 1, 40 | 0.48 | 1, 36 | 0.67 | 1, 36 | 0.95 | 1, 36 | 7.99** |
Brightness PC1 | 1, 39 | 4.48* | 1, 40 | 0.07 | 1, 36 | 0.00 | 1, 36 | 0.04 | 1, 35 | 1.82 |
UV PC1 | 1, 39 | 0.21 | 1, 40 | 0.29 | 1, 36 | 0.03 | 1, 36 | 0.00 | 1, 35 | 2.03 |
Manipulation × wing patch size | 2, 37 | 1.67 | 2, 38 | 1.12 | 2, 34 | 2.23 | 2, 34 | 1.35 | 2, 33 | 2.18 |
Manipulation × brightness PC1 | 2, 37 | 0.31 | 2, 38 | 1.48 | 2, 34 | 0.80 | 2, 34 | 2.14 | 2, 33 | 2.13 |
Manipulation × UV PC1 | 2, 37 | 0.29 | 2, 38 | 0.07 | 2, 34 | 1.09 | 2, 34 | 1.75 | 2, 33 | 0.49 |
General linear models with stepwise backward selection and reintroduction. *P < 0.05, **P < 0.01, ***P < 0.001.
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12 days . | Nestling mass 12 days . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 40 | 13.68*** | 2, 41 | 12.77*** | 2, 37 | 5.03* | 2, 37 | 3.82* | 2, 36 | 46.67*** |
Wing patch size | 1, 39 | 1.59 | 1, 40 | 0.48 | 1, 36 | 0.67 | 1, 36 | 0.95 | 1, 36 | 7.99** |
Brightness PC1 | 1, 39 | 4.48* | 1, 40 | 0.07 | 1, 36 | 0.00 | 1, 36 | 0.04 | 1, 35 | 1.82 |
UV PC1 | 1, 39 | 0.21 | 1, 40 | 0.29 | 1, 36 | 0.03 | 1, 36 | 0.00 | 1, 35 | 2.03 |
Manipulation × wing patch size | 2, 37 | 1.67 | 2, 38 | 1.12 | 2, 34 | 2.23 | 2, 34 | 1.35 | 2, 33 | 2.18 |
Manipulation × brightness PC1 | 2, 37 | 0.31 | 2, 38 | 1.48 | 2, 34 | 0.80 | 2, 34 | 2.14 | 2, 33 | 2.13 |
Manipulation × UV PC1 | 2, 37 | 0.29 | 2, 38 | 0.07 | 2, 34 | 1.09 | 2, 34 | 1.75 | 2, 33 | 0.49 |
. | Male feeding rate . | Female feeding rate . | Nestling tarsus 12 days . | Nestling mass 12 days . | Number of fledglings . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | df . | F . | df . | F . | df . | F . | df . | F . | df . | F . |
Manipulation | 2, 40 | 13.68*** | 2, 41 | 12.77*** | 2, 37 | 5.03* | 2, 37 | 3.82* | 2, 36 | 46.67*** |
Wing patch size | 1, 39 | 1.59 | 1, 40 | 0.48 | 1, 36 | 0.67 | 1, 36 | 0.95 | 1, 36 | 7.99** |
Brightness PC1 | 1, 39 | 4.48* | 1, 40 | 0.07 | 1, 36 | 0.00 | 1, 36 | 0.04 | 1, 35 | 1.82 |
UV PC1 | 1, 39 | 0.21 | 1, 40 | 0.29 | 1, 36 | 0.03 | 1, 36 | 0.00 | 1, 35 | 2.03 |
Manipulation × wing patch size | 2, 37 | 1.67 | 2, 38 | 1.12 | 2, 34 | 2.23 | 2, 34 | 1.35 | 2, 33 | 2.18 |
Manipulation × brightness PC1 | 2, 37 | 0.31 | 2, 38 | 1.48 | 2, 34 | 0.80 | 2, 34 | 2.14 | 2, 33 | 2.13 |
Manipulation × UV PC1 | 2, 37 | 0.29 | 2, 38 | 0.07 | 2, 34 | 1.09 | 2, 34 | 1.75 | 2, 33 | 0.49 |
General linear models with stepwise backward selection and reintroduction. *P < 0.05, **P < 0.01, ***P < 0.001.
DISCUSSION
Own care in relation to ornaments and environment
We may expect environment-dependence in the relationship between parental performance and ornamentation for various reasons. For example, some environments may provide particularly good opportunities for alternative reproductive tactics, thereby specifically reducing the parental contribution of ornamented males (Qvarnström et al. 2003). Alternatively, parental quality differences among individuals may be differently expressed in different environments (García-Navas and Sanz 2011). We failed to find environment-dependent variation in the ornamentation-feeding rate relationship for any of the female or male ornamental traits we measured. We are aware of 2 previous, relevant brood size manipulation studies. In one of these, parental care was not examined, but relationships of ornamentation with several reproductive success measures were environment-dependent (tawny owl Strix aluco, Emaresi et al. 2014). In the other study, relationship of one studied male ornament (pied flycatcher Ficedula hypoleuca forehead patch size) and feeding rate did not depend on the experimental environment (Järvistö et al. 2015). None of these studies (including ours) concerned ornaments with direct dependence on limiting nutrients (e.g., carotenoid-based traits, Olson and Owens 1998), although environment-dependence of parental quality indication may be most likely in such traits. In fact, in our population, there was no evidence for any relationship between ornamentation and own parental contribution, although such correlations have been detected in the incubation period in this same population (Kötél et al. 2016). This pattern is consistent with the literature, where relatively few studies have detected consistent relationships between ornamentation and own nestling care (Hegyi et al. 2015).
In our population, nestling feeding rate is a reliable measure of feeding effort but apparently not of parental quality (see Mänd et al. 2013 for similar results from pied flycatchers). Expressing parental quality as total biomass output per unit effort made little difference to our conclusions. Only a marginally manipulation-dependent pattern emerged between male brightness PC1 (structural brightness) and parental quality. The relationship was positive in reduced broods but it faded to nonsignificantly negative with increasing manipulation. This pattern likely reflects the interaction between the effects of color on feeding and vice versa. That is, parental quality indication by structural coloration may be destroyed by unpredictable reverse effects of experimentally increased overall feeding effort on coloration (see also results for female brightness PC and male feeding in the next section). In conclusion, there is little evidence that plumage coloration indicates either own parental effort or parental quality in this population.
Partner care in relation to ornaments and environment
Contrary to own care results, we found patterns in both male and female feeding in relation to partner ornamentation. The literature contains a higher percentage of significant relationships concerning nestling feeding in relation to partner ornaments than for own ornaments, and the overall direction of the relationships tends to be positive (Hegyi et al. 2015). Female feeding rate was positively related to male wing patch size in our study, which is in line with the literature, but this may easily reflect territory quality differences, as wing patch size of male collared flycatchers has a role in intrasexual competition for territories and possibly for mates during courtship period (Garamszegi et al. 2006), and it may also signal the quality of the defended breeding area, as judged from the nonrandom spatial distribution of wing patch size across the study site (Hegyi, Rosivall, et al. 2008). Alternatively, male wing patch size may inversely associate with the male’s own parental investment (trade-off hypothesis), because displaying a more conspicuous patch is associated with increased wearing costs (Martin and Ghalambor 1999; Garamszegi et al. 2006), so a more attractive male provides less care (Mitchell et al. 2007; Diniz et al. 2015). In this case, females might increase reproductive success if they try to compensate the attractive mate’s reduced parental contribution. Kötél et al. (2016) found in this very same population that the intensity of male parental behavior during incubation was negatively related to the size of the wing patch (i.e., males with larger wing patches invested less in feeding their incubating mates), and probably as a response females altered their own incubation behavior (they left the nest less often when the partner’s patch size was larger). In the present study we found no reduction in the nestling feeding of males with large wing patches, but increased female nestling feeding may still reflect the experience with male care during incubation, when male care is directly visible to females, being targeted to them and not to the offspring.
We also found that male feeding rate tended to relate inversely to female brightness PC. As suggested by the pattern of factor loadings of the different plumage areas, this color axis probably represented the microstructure-based individual variation of feather intensity. Males seemed to increase their nestling feeding rate when paired to a less ornamented female. It is possible that more frequent foraging trips or nest visits result in greater feather abrasion and reduced structural brightness in females. This latter idea is supported by the findings of a study carried out in another population of the collared flycatcher, where it was demonstrated that experimental manipulation of parental effort generated notable differences with respect to feather abrasion, but only in females (Merilä and Hemborg 2000). Feather wear is supposedly responsible for changing of brightness in other species, but in different directions (decreasing brightness: Surmacki et al. 2011, increasing brightness: Örnborg et al. 2002). Additionally, in the sister species pied flycatcher Alt et al. (2015) found that increased parental effort had a negative effect on females’ feather sanitation as suggested by free-living bacterial numbers on feathers, which could also cause color changes (Shawkey et al. 2007). Additionally, Shawkey et al. (2007) found a positive relationship between the presence of bacteria and brightness, and a negative association with UV chroma, which patterns are partly in accordance with our results. Therefore, we cannot exclude that the relationship of male feeding and female brightness reflected a response of male care to perceived female plumage deterioration associated with breeding. Indeed, increased male feeding of nestlings when female color expression is poor has been reported previously in other species (Limbourg et al. 2013a, 2013b). In conclusion, we have little evidence for sexual selection on the female structural brightness axis we used here (Laczi et al. 2011) and the present results also indicate that this trait may rather be used as an indicator of general plumage state and poor values may elicit compensatory care response from males.
To the very best of our knowledge, only one previous study has tested whether partner feeding investment in relation to an ornamental trait changed with environmental quality. This study experimentally manipulated both female ornamentation and brood size but found no preferential investment by males in relation to the manipulated female ornament, regardless of brood size (Berzins and Dawson 2016). Our results were only partly similar to theirs, as we did find ornament-dependent partner care but this did not vary with manipulated brood size. More studies are needed to determine whether differential or compensatory partner investment patterns in relation to ornamentation generally show environment-dependence.
Nestling attributes in relation to ornaments and environment
There are a few studies suggesting that nestlings originating from differently ornamented parents have intrinsically different potential to cope with environmental stress. For example, in our study population, when cross-fostered to another nest, broods of nestlings from females with larger wing patches elicited higher feeding rates from the rearing parents (Kiss et al. 2013). In barn owls (Tyto alba), nestling growth and maternal pheomelanism were positively related under ad libitum feeding conditions but negatively related under simultaneous food restriction and immune challenge (Piault et al. 2009). In feral pigeons (Columba livia domestica), plumage darkness of the biological father was positively related to fledgling body mass under food limitation (Jacquin et al. 2012). However, in pied flycatchers, forehead patch size of the original father did not predict the body mass of offspring under different temperature conditions or different brood size manipulation groups (Järvistö et al. 2015). These studies, together with results on nestling growth from cross-fostering studies without brood size manipulation (Hegyi, Rosivall, et al. 2011; Remeš and Matysioková 2013) provide evidence that parental ornamentation may indicate intrinsic nestling attributes in some systems.
Our present experimental study yielded opposite patterns as Piault et al. (2009), that is, offspring of more ornamented males (in terms of forehead patch size) fared better under food stress, but worse under unexpectedly good food conditions. Analyzing own and foreign nestlings separately, we confirmed that this forehead patch size dependent nestling size pattern was restricted to own nestlings, so it reflected an interaction between intrinsic nestling traits and environmental quality. Previous studies of our population detected positive relationships between various aspects of nestling growth and paternal forehead patch size (Szöllősi et al. 2009; Hegyi, Rosivall, et al. 2011). No study of our population detected any relationship between forehead patch size and either maternal effects (e.g., Hargitai et al. 2005; Török et al. 2007) or parental care (Kiss et al. 2013; Kötél et al. 2016, this study), so the various forehead patch size dependent patterns of nestling growth, including our present findings, may reflect genetic relationships between ornamentation and the resource acquisition or assimilation capacity of the nestlings.
Reproductive success patterns and future directions
Concerning the number of offspring fledged from our experimental broods, there was a robust negative effect of female wing patch size, which provides further evidence for the widespread pleiotropic effects of this trait, from territoriality and polygyny (Hegyi et al. 2007; Hegyi, Garamszegi, et al. 2008) to maternal effects (Hegyi, Herényi, et al. 2011), parental behavior (Kötél et al. 2016) and offspring food acquisition capacity (Kiss et al. 2013). Further studies are needed to integrate these several lines of evidence into a coherent picture on the adaptive value and evolutionary significance of this female ornament. In addition, we found a positive tendency between fledgling number and male forehead patch size, irrespective of manipulation. In light of the nestling size and growth patterns discussed above (see also Szöllősi et al. 2009; Hegyi, Rosivall, et al. 2011), this pattern may reflect intrinsic nestling characteristics.
Overall, few of the above discussed patterns of ornamentation with own care, parental quality, partner care, or nestling morphology were manifested in fledgling number. It is possible that recruitment would show a number of additional patterns (Hegyi, Herényi, et al. 2011; Szász et al. 2017) and should therefore be analyzed as the data become available. However, the lack of the expected reproductive success patterns may largely be due to the heterogeneity of the effects we detected. Our focus on multiple different ornaments of both parents provided a unique perspective compared with previous experimental studies of the environment-dependence of ornament information content. The overall conclusion from the findings must be a word of caution. Examining single success measures in relation to single ornaments may provide information concerning the proximate mechanisms and immediate consequences of ornament expression. However, integrated studies of well known model species that combine the known ornamental traits of both sexes and examine multiple reproductive parameters under experimentally manipulated environmental conditions are important if we want to see whether, why, and how single ornaments exert fitness effects and thereby shape the evolution of sexual ornamentation systems.
FUNDING
This work was supported by Hungarian Scientific Research Fund (OTKA) (grant numbers K101611, K115970), and the Eötvös Loránd University (PhD scholarship) to D.K.
The authors are grateful to N. Boross, R. Hargitai, M. Jablonszky, K. Krenhardt, G. Markó, G. Nagy, B. Rosivall, E. Szász, and S. Zsebők for help with the fieldwork.
Data accessibility: Analyses reported in this article can be reproduced using the data provided by Laczi et al. (2017).
REFERENCES
Author notes
Handling editor: Naomi Langmore