J. Avian Biol. 42: 309 322, 2011 doi: 10.1111/j.1600-048X.2011.05240.x # 2011 The Authors. J. Avian Biol. # 2011 Nordic Society Oikos Subject Editor: Theunis Rersma. Received 15 May 2010, accepted 3 May 2011
Carotenoid-based plumage coloration in golden-crowned kinglets Regulus satrapa: pigment characterization and relationships with migratory timing and condition
Celia K. S. Chui, Kevin J. McGraw and Ste´phanie M. Doucet
C. K. S. Chui ([email protected]) and S. M. Doucet, Dept of Biol. Sci., Univ. Windsor, 401 Sunset Avenue, Windsor, ON, N9B 3P4, Canada. K. J. McGraw, School of Life Sciences, Arizona State Univ., Tempe, AZ 85287-4501, USA.
Carotenoid-based ornamental coloration has long been proposed to honestly signal quality due to its dependence on individual condition. Because migration can be one of the most stressful periods of an animal’s annual cycle, developing colourful plumage may be particularly challenging for species in which migration and moult periods overlap or occur sequentially. The purpose of this study was to investigate pigmentary and condition-dependent bases of carotenoid colour variation in a small migratory passerine, the golden-crowned kinglet Regulus satrapa (Family Regulidae). We captured 186 male and female kinglets of various ages during fall migration in southwestern Ontario, Canada and recorded arrival date, body condition index, fat and pectoral muscle scores, wing mite infestation, and feather growth rate as measures of condition. We quantified crown coloration using reflectance spectrometry and analyzed feather carotenoids using high- performance liquid chromatography. Yellow crown feathers of female kinglets contained only yellow hydroxycarotenoids, whereas orange feathers of males harboured a suite of eight carotenoid pigments. Males with longer wavelength orange crown hues deposited greater concentrations of ketocarotenoids, especially canthaxanthin. Female kinglets with longer wavelength crown hues and males with longer wavelength crown hues and more saturated crown coloration left for migration earlier in the year. Females with longer wavelength crown hues had fewer feather mites and tended to be in better condition. However, male kinglets with more saturated coloration possessed smaller pectoral muscles. This is the first study to identify plumage carotenoids in this North American bird family and to determine the pigmentary basis for both inter- and intrasexual colour variation. Our results provide further support for the condition-dependence of carotenoid coloration and suggest that ornamental elaboration in both sexes may encode information about fall condition and migratory performance.
Carotenoid coloration serves as a condition-dependent trait adopt a ‘more-is-better’ strategy to produce more colourful that is important in visual signalling and sexual selection in plumage compared to conspecifics. For example, green- a variety of taxa ranging from fishes (Milinski and Bakker finches Carduelis chloris and American goldfinches Spinus 1990, Houde 1997) to reptiles (Kwiatkowski and Sullivan tristis produce deep yellow plumage by depositing high 2002) to birds (reviewed in Hill 2006a, Senar 2006). concentrations of canary xanthophylls (lemon yellow pig- Proximate factors that mediate the expression of carotenoid- ments derived from dietary lutein and zeaxanthin, Saks based plumage pigmentation in birds include dietary et al. 2003, McGraw and Gregory 2004). Other species carotenoid access, immunocompetence, and general nutri- have the capacity to metabolize dietary yellow pigments into tion (reviewed in Hill 2006b, McGraw 2006). Detrimental carotenoids that exhibit redder hues and can selectively effects of endoparasites on individual health and plumage deposit these red pigments to grow orange or red plumage coloration are also well-documented (reviewed in Hill (reviewed in McGraw 2006). In house finches Carpodacus 2006b). However, the effects of ectosymbionts, such as mexicanus, for example, intraspecific variation in carotenoid feather mites, on carotenoid-based colours remain equivocal coloration primarily results from differing proportions of (reviewed in Proctor and Owens 2000). Overall, the red and yellow pigments, rather than their absolute influence of multiple genetic and environmental factors concentrations (Inouye et al. 2001). In the present study, on carotenoid colours enables these ornaments to be we investigated the carotenoid pigment composition and particularly variable, which may enhance their ability to possible condition-dependence of a colourful plumage trait honestly reveal individual quality. in a small, migratory passerine species, the golden-crowned From a biochemical standpoint, the expression of kinglet Regulus satrapa. carotenoid coloration depends on the types and concentra- Migration is a stressful and energetically demanding tions of pigments present (Hill et al. 2002). Some species activity; amongst many other physiological costs, sustained
309 migratory flights require individuals to deplete fat stores, crown patch bordered by yellow feathers (Martens and metabolize muscle protein, and lose body water, thus Pa¨ckert 2006). While these orange feathers are normally necessitating stopovers to refuel (Berthold 2001). It has concealed, male kinglets will expose them during courtship also been shown that strenuous exercise is associated with and antagonistic interactions (Martens and Pa¨ckert 2006), reduced immune response (Ra˚berg et al. 1998), which may, implicating a possible role in sexual selection. There has in turn, affect an individual’s ability to deposit fat (Merila¨ been little research on plumage colour in this family aside and Svensson 1995). Differences in individual health may from carotenoid pigment analyses on goldcrest Regulus therefore be especially apparent during migration, provid- regulus crown feathers (Stradi 1998). Recently, we per- ing a unique opportunity to investigate the condition- formed quantitative analyses of large-scale geographic dependence of plumage ornaments. Relatively little is variation in golden-crowned kinglet coloration (Chui and known about the link between migratory condition and Doucet 2009). We found that the degree of sexual ornamental coloration or their common linkage to factors dichromatism varied with climate, which may be associated such as health or nutrition. In a general sense, a recent with reduced sexual selection pressures for ornamental traits review on red knots Calidris canutus showed that the in harsher climates (Badyaev 1997). We also found that subspecies’ varying annual cycles and environmental con- crown colour variation within a population can be almost as straints result in differential ‘bottlenecks’ that can impact extensive as the overall variation observed throughout the the moult and breeding plumage of these long-distance species range (Chui and Doucet 2009). migrants (Buehler and Piersma 2008). On an individual The purpose of this study was to investigate individual scale, a few studies have investigated the relationship variation in carotenoid coloration in migrating golden- between extent of moult into breeding plumage and crowned kinglets. We sought to characterize the carotenoid individual condition during spring migration in bar- profile of kinglet crown feathers and to quantify tailed godwits Limosa lapponica and great knots Calidris the pigmentary basis for individual variation in colour. tenuirostris, with mixed results (Piersma and Jukema 1993, We predicted that, if variation in crown coloration depends Battley et al. 2004). Additionally, a study on blue tits upon carotenoid content, kinglets with deeper orange Cyanistes caeruleus during fall migration found that birds crowns would deposit greater concentrations of red pig- with larger fat stores exhibited brighter carotenoid colora- ments into their feathers, whereas birds with yellower tion, which may indicate a link between nutritional plumage hues would deposit proportionately more yellow condition at moult and subsequent migratory success pigments. We then sought to investigate the potential for (Svensson and Merila¨ 1996). condition-dependence of crown coloration in both male Developing colourful plumage may be particularly and female kinglets. We predicted that, if carotenoid challenging for species with overlapping periods of breeding coloration is an honest indicator of quality, crown colour and moult, or moult and migration. In addition, if would be correlated with one or more measures of ornamental coloration is an important criterion for mate condition, including migration arrival date, body condition choice, then investing into fall moult may be especially index, fat stores, pectoral muscle size, prevalence of feather important for species that carry their fall plumage into the mites, and feather growth rate. next breeding season. Research on American redstarts Setophaga ruticilla previously suggested that males breeding late in the season undergo costly moult-migration overlap Methods (Merila¨ 1997, Pe´rez-Tris et al. 2001) and deposit a lower concentration of carotenoids into their plumage, which may We conducted field work at the Holiday Beach Migration affect sexual selection the following year (Norris et al. Observatory (HBMO) in Amherstburg, Ontario, Canada 2004b). However, a recent study countered that these are (42o02?24ƒN, 83o02?24ƒW), an important stopover site for actually high-quality individuals that adventitiously lost and many migratory raptor and passerine species (IBA Canada re-grew feathers (with reduced saturation) during the 2004). We began monitoring the area for migrants on 13 overwintering period, and are not facing reproductive September 2008 with a set-up of mist-nets and continuous trade-offs (Reudink et al. 2008). Furthermore, geographic playback of golden-crowned kinglet songs and calls (Cornell colour variation in redstarts was attributed to differences in Lab of Ornithology, Ithaca, NY, USA). However, because carotenoid availability at their moulting localities rather we encountered most kinglets at their expected peak than individual condition (Norris et al. 2007). migration times (based on previous HBMO data), it was We studied golden-crowned kinglets at a stopover site unclear whether audio playback enhanced our capture during fall migration in southwestern Ontario, Canada. success. From 4 October 9 November, we captured 186 One unique aspect of our study site is the migration golden-crowned kinglets between 07:00 and 15:30 (EST). bottleneck that exists as a result of the local topography and Each bird was assigned a unique band number as issued by the Great Lakes, which could potentially funnel migrating the Bird Banding Laboratory. We sexed birds by crown kinglets here from various breeding areas in Canada (U.S. coloration and aged them as hatching year (HY) or after- Geological Survey 2006). Golden-crowned kinglets are hatching year (AHY) by assessing skull pneumatization and distributed throughout North America and are generally age-specific plumage characters (Pyle 1997). Of the 186 short-distance migrants (Martens and Pa¨ckert 2006). kinglets captured, 16 (8.6%) were AHY females, 36 With the exception of the ruby-crowned kinglet Regulus (19.4%) were AHY males, 54 (29.0%) were HY females, calendula, the other five species within Regulidae exhibit a and 80 (43.0%) were HY males. The majority of kinglets conserved pattern of sexual dichromatism: females have a had completed their prebasic moult, which includes all uniformly yellow crown, whereas males have an orange body feathers (Pyle 1997). In general, golden-crowned
310 kinglets undergo a single moult between July and the number of bands that we scored and the number of September, which may overlap with breeding and/or fall individuals we used in this analysis. On average, we were migration (Ingold and Galati 1997). able to count 16.9592.25 d of growth per feather analyzed. We then recorded the total length of the growth bar section, and calculated feather growth rate by dividing total length Condition variables by number of growth bars. Feather growth rate has been shown to reveal nutritional condition at the time of moult For each kinglet, we recorded six variables that are in various bird species (Grubb 1989, 2006). putatively linked to body condition: migration arrival date, residual body mass, fat and pectoral muscle scores, wing mite infestation, and feather growth rate. We noted Reflectance spectrometry arrival date with the prediction that earlier migrants would be in better condition (Kokko 1999, Ninni et al. 2004). We We collected 10 crown feathers from every kinglet (orange assumed this to be the date of capture for three reasons: 1) for males, yellow for females). We did not collect yellow the Holiday Beach Conservation Area is only 9.0 km2 (IBA feathers from males because we were most interested in the Canada 2004), 2) migrating flocks of kinglets are easily sexual dichromatism of the central crown patch, and also heard, and 3) we rarely recaptured individuals (n 3). We because we would have drastically altered the appearance of used an electronic balance to measure body mass to the the male crown by plucking these sparser feathers. We taped each individual’s feathers in an overlapping fashion to nearest 0.1 g, a wing rule to measure unflattened wing- represent their natural configuration on a piece of chord to the nearest 0.5 mm, and digital callipers to matte black construction paper. We measured crown measure tarsus length to the nearest 0.1 mm. We then plumage reflectance using an Ocean Optics USB4000 calculated a body condition index by log-transforming body reflectance spectrometer and a PX-2 pulsed xenon lamp mass and tarsus length data, then using the residuals from (Ocean Optics, Dunedin, FL, USA). A single fibre-optic an ordinary least squares (OLS) linear regression (r 0.21, probe transmitted full-spectrum light from the lamp and n 185, p 0.004) (Brown 1996, Ardia 2005, Schulte- transferred light reflected from the specimen back to the Hostedde et al. 2005). While these residual indices have spectrometer. The probe was mounted with a rubber sheath been used in numerous bird studies, there are several that extended ca 5 mm past the tip to maintain a fixed underlying biological and statistical assumptions that need distance from the specimen and to exclude external light. to be validated (van der Meer and Piersma 1994, Green o We placed the probe at a 90 angle to the feather surface 2001). In particular, this methodology assumes that body and took five readings, which comprised an average of condition index truly reflects variation in energy stores, and 30 spectra collected sequentially by OOIBase32 software. that size and mass are independent of one another and of All spectral measurements were expressed as the percentage other variables in subsequent analyses. Because we cannot of reflected light relative to a Spectralon white standard directly test all of these assumptions, we should be cautious (Ocean Optics, Dunedin, FL, USA). when interpreting our results involving body condition. We restricted our analyses to wavelengths of 300 We scored subcutaneous fat using an 8-point scale based 700 nm, which comprise the extent of the avian visual on the amount of fat deposited first in the furcular hollow spectrum (Cuthill 2006). We used the software program and subsequently in the wingpits and along the lower CLR (Montgomerie 2008) to calculate three colour vari- abdomen (DeSante et al. 2009). We also scored the size of ables that summarize the variation in reflectance data: the pectoral muscles using a 5-point scale that captures brightness, hue, and saturation (Andersson and Prager sternum prominence and muscle shape (modified from 2006, Montgomerie 2006). Brightness was calculated as Barlein 1995). We recorded the time of day that each bird the average reflectance between 300 and 700 nm; hue as the was captured because fat stores (Lindstro¨m and Piersma wavelength at which the spectrum reached half of its 1993) can change throughout the day, especially in small maximum reflectance; and saturation as the difference birds (Gosler 1996), and we controlled for time of day in between the maximum and the minimum reflectance, our statistical analyses. We estimated the number of feather divided by the total brightness. Our measure of saturation mites by eye on the outstretched right wing when backlit. was highly correlated with short-wavelength and carotenoid There was a high prevalence (n 122, 66% of birds saturation variables (both pB0.0001, Montgomerie 2008). captured) and variable number (range: 0 80) of mites on Thus, we simplified the analyses by measuring saturation the flight feathers of golden-crowned kinglets, which that encompasses the entire 300 700 nm spectrum. suggests that wing mite infestation could be a good measure of individual condition. We did not observe any ticks or other ectoparasites that have been reported for this species Carotenoid extraction and chromatography (Ingold and Galati 1997). To calculate feather growth rate, we collected the outer We conducted high-performance liquid chromatography right rectrix whenever possible, where growth bars are the (HPLC) analyses on feathers from a subset of the kinglets most apparent. Each growth bar consists of one dark and captured. In male kinglets, the coloured barbs extend one light band representing a 24 h period of feather growth almost to the base of the crown feather. In females kinglets, (Michener and Michener 1938). We taped each rectrix to however, the coloured portion comprises only about half of black construction paper and counted as many consecutive the feather; therefore, we could not obtain a sufficient bands as possible by eye, marking the centre of each band. sample for carotenoid extraction from each individual We used this approach because we wanted to optimize both female. We thus restricted quantitative HPLC analyses to
311 male crown coloration, and we pooled crown feathers from We identified feather carotenoids by comparing their multiple females for a qualitative assessment of carotenoid retention times, spectral shapes, and absorbance maxima content. Based on our spectrometric measurements of hue, (lmax; Table 1) to published data on known carotenoids we specifically selected 10 males from the shortest- (Stradi et al. 1995, Stradi 1998, Inouye et al. 2001, wavelength hues (range 555.6 561.0 nm, mean9SD Andersson et al. 2007). Peak areas for the pigments were 559.0692.04 nm), 10 from medium-wavelength hues integrated at their respective maximum wavelengths. To (surrounding the population mean; range 568.8 account for chromatographic noise, peaks below 0.01 569.4 nm, mean9SD 569.0290.15 nm), and 10 from absorbance units were not considered. the longest-wavelength hue values (range 576.0 Golden-crowned kinglet chromatograms revealed late- 579.0 nm, mean9SD 577.4291.14 nm). We sampled eluting esterified pigments that were highly repeatable when these males in order to represent the range of wavelengths aliquots of the same samples were run in tandem (n 3, reflected in this population (Fig. 1). As expected from mean repeatability 96.2%). Because we could not con- carotenoid-based coloration (Andersson and Prager 2006), fidently identify the parent carotenoid for all of these crown hue and saturation were highly positively correlated esterified peaks, we categorized them into broader group- in male kinglets (r 0.80, n 116, pB0.0001); therefore, ings according to the carotenoids’ chromophore length and we were also selecting males that generally represented the the functional end group(s) they contain (Table 1) (Stradi population range of colour saturation values. 1998), after determining that ‘free’ pigments and their We extracted feather carotenoids using a mechanical esters were positively correlated within each colour category extraction procedure (McGraw et al. 2003). Briefly, we (all pB0.02). For simplicity, we will subsequently refer to trimmed off the coloured barbs of all 10 crown feathers 4,4?-ketocarotenoids as ‘red’ ketocarotenoids due to their collected in the field and weighed each sample to the nearest longer chromophore, and 4-ketocarotenoids as ‘orange’ 0.0001 g (ca 0.0010 g per sample). We finely ground the ketocarotenoids; ‘yellow’ hydroxycarotenoids possess the feathers in 1 ml methanol using 440C stainless steel balls shortest chromophores (Britton et al. 2004). and a SPEX 8000D Mixer/Mill (SPEX CertiPrep, Metu- Because reference standards for other carotenoids were chen, NJ, USA) at 60 Hz for 2 h to solubilize the pigments. unavailable, we quantified pigment concentrations using a We then centrifuged the solution for 5 min at 10 000 RPM, calibration curve generated from an external lutein standard (Toomey and McGraw 2009). Carotenoid concentration in transferred the supernatant, and evaporated the latter to 1 dryness under a stream of nitrogen. the standard (mg g ) was calculated with the following We resuspended the extracts in 200 ml HPLC mobile formula: (A volume of extract [ml] 10)/E, where A was phase (42:42:16, v/v/v, methanol:acetonitrile:dichloro- the absorbance of the sample at lmax (448 nm for methane), and injected 50 ml of each sample into a Waters xanthophylls) using a Beckman DU530 UV/Vis spectro- Alliance 2695 Separations Module (Waters Corporation, photometer, and E is the extinction coefficient at 1% per Milford, MA, USA) installed with a Waters YMC Car- centimetre of the relevant carotenoid at lmax (2550 for otenoid 5 mm column (4.6 250 mm) with column heater lutein, Britton 1985). We calculated total carotenoid set to 308C. Mobile phase was administered at a constant concentration in each kinglet sample by summing up the flow rate of 1.2 ml min 1. Data were collected from 300 to HPLC peak areas. However, because we did not measure 550 nm using a Waters Alliance 2998 Photodiode Array absorbance by spectrophotometry and because we instituted Detector and analyzed with Waters Empower 2 Software. the aforementioned peak exclusion threshold, these values may not perfectly represent actual total carotenoid content. Nevertheless, total carotenoid concentrations should be 2.8 consistent for inter-sample comparisons.
Statistical analyses 2.4 Data analyses were conducted using JMP 6.0 (SAS Inst., Cary, NC, USA). Due to our uneven sampling of males and 2.0 females, and of HY and AHY birds, we used a general linear
Crown saturation model incorporating sex, age, and sex-by-age interactions to investigate sex- and age-specific differences in coloration and condition. For our coloration vs condition analyses, we 1.6 constructed general linear multivariate models using five of our six measures of condition as predictor variables, and we 510 525 540 555 570 585 assigned the three colour metrics as dependent variables. Crown hue (nm) While multicollinearity was not an issue in our models (all Figure 1. Scatterplot of crown saturation vs crown hue for variance inflation factors [VIF]B5), we opted to include male (black circles; n 116) and female (grey circles; n 68) only body condition index in our analyses because fat scores golden-crowned kinglets, illustrating the colour variation within were highly correlated with body condition index in both and between sexes. Open circles indicate the subset of males sexes (Table 2). We verified that an all-inclusive model (n 30) that were used in HPLC analyses: left cluster ‘short- produced similar results (data not shown). Since golden- wavelength hues’, centre cluster ‘medium-wavelength hues’, crowned kinglets exhibit substantial sexual dichromatism in right cluster ‘long-wavelength hues’. crown plumage, we conducted coloration vs condition
312 analyses split by sex to investigate whether male and female ornaments signal different kinds of quality. For our HPLC analyses, we lost one sample from the long-wavelength hue group during the extraction procedure, so we excluded this sample from further analyses of pigment concentration. We compared the absolute and relative concentrations of each pigment as well as carotenoid groups (red, orange, yellow pigments; Table 1) between the three categories of male crown hue (long-, medium-, and short-wavelength; de- scribed above) using analysis of variance (ANOVA) and post hoc Tukey tests.
Results nce maxima were obtained by high-performance Carotenoid pigment analyses
Yellow crown feathers of female kinglets contained only yellow xanthophylls, the majority being lutein and 3?-dehydrolutein, with lesser amounts of canary xanthophyll A and zeaxanthin (Table 1, Fig. 2A). In contrast, at least eight carotenoid pigments were responsible for producing the orange colour of male crown feathers, including four ketocarotenoids (canthaxanthin, astaxanthin, a-doradexanthin, and small amounts of echinenone) and four hydroxycarotenoids (lutein, zeaxanthin, 3?-dehydrolutein, and small quantities of canary xanthophyll A; Table 1, Fig. 2B). Males with long-wavelength crown hues deposited greater absolute concentrations of ketocarotenoids (AN- OOO 5.1 6.5 8.0 446 447 453 M and F M and F M and F O 4.6 444 M and F O 10.6 468 M O 5.9 455, 474 M O 8.8 476 M O 7.5OVA: 479 both F2,26 10.7, M p 0.0004), contributing to greater levels of carotenoids in general (F2,26 8.63, p 0.001), than individuals whose crown hues exhibited intermediate and shorter wavelengths (Fig. 3A). However, the relative concentration of ‘red’ ketocarotenoids was not significantly higher in kinglets with long-wavelength crown hues (F2,27 2.20, p 0.13). Rather, males with long- wavelength hues deposited higher proportions of ‘orange’ ketocarotenoids than those with short-wavelength hues (F2,27 5.90, p 0.008), while the opposite was true for ‘yellow’ hydroxycarotenoids (F2,27 10.4, p 0.0005; Fig. 3B). Birds with medium- and short-wavelength hues were not significantly different for any of the above analyses (Tukey tests, p 0.05). Looking at pigment variation in more detail, we found that feathers exhibiting long-wavelength hues had higher absolute concentrations of all ketocarotenoids than feathers with short-wavelength hues, and we found a similar pattern for long- vs medium-wavelength hues except for a-doradexanthin (Fig. 4A). Echinenone was the only ketocarotenoid that differed between medium- and short- Carotenoid Chromophore length * Functional end group(s) ** Retention time (min) Absorbance maxima (nm) Present in which sex? wavelength hues; feathers from the latter group did not -doradexanthin 11 1 C LuteinZeaxanthin 11 10 0 C 0 C a possess echinenone at all (Fig. 4A). Comparisons of relative pigment concentrations revealed alternative patterns of variation: differences in ketocarotenoids were no longer as prominent between hue categories, although canthaxanthin O) groups that increase the chromophore length by adding additional conjugated double bond(s) (Britton et al. 2004). concentration remained significantly higher in feathers displaying long- vs short-wavelength hues (F2,27 6.68, p 0.004; Fig. 4B). Among the yellow xanthophylls, absolute concentrations -ketocarotenoids Astaxanthin 13 2 C ? of lutein and 3?-dehydrolutein did not differ among * number of conjugated** double carbonyl bonds (C that generate a carotenoid pigment’s colour (Britton et al. 2004). ‘yellow’ pigments 3’-dehydrolutein 10 0 C Hydroxycarotenoids Canary xanthophyll A 9 0 C ‘orange’ pigments Echinenone 12 1 C 4-ketocarotenoids Table 1. Characteristics of the carotenoid pigments found in the crown feathers of golden-crowned kinglets. Data on retention times and main absorba liquid chromatography (HPLC) analyses in this study. ‘red’ pigments Canthaxanthin 13 2 C 4,4 birds as a function of crown hue. As may be expected,
313 Table 2. Pearson correlation coefficients between condition variables used to investigate the condition-dependence of crown coloration in golden-crowned kinglets. Females shown in lower diagonal (n 70), males in upper diagonal (n 116).
Arrival date Body condition index Fat score Pectoral muscle score Wing mite load Feather growth rate
Arrival date 0.17 0.19 * 0.05 0.04 0.25 ** Body condition index 0.005 0.60 *** 0.16 0.10 0.06 Fat score 0.01 0.58 *** 0.14 0.15 0.05 Pectoral muscle score 0.11 0.02 0.19 0.10 0.18 Wing mite load 0.30 * 0.16 0.17 0.11 0.03 Feather growth rate 0.03 0.06 0.10 0.05 0.14
*pB0.05, ** pB0.01, *** pB0.0001. feathers from male kinglets with short-wavelength hues time of day, we found that AHY birds had slightly greater 2 possessed significantly greater concentrations of zeaxanthin fat stores than HY birds (model: R 0.11, F4,179 5.34, than feathers with medium-wavelength hues (F2,26 4.97, p 0.0004; age: F1,182 3.34, p 0.07), which suggests p 0.01; Fig. 4A). Interestingly, individuals with short- that AHY birds are in better condition. However, AHY wavelength hues deposited a lower concentration of canary individuals were also infested with significantly more wing 2 xanthophyll A, which absorbs shorter wavelengths of light mites (model: R 0.14, F3,180 9.47, pB0.0001; age: than other yellow plumage carotenoids (Stradi 1998), than F1,182 20.1, pB0.0001). birds with either long- or medium-wavelength hues. Patterns of variation in pigment concentrations may be attributed to the intercorrelations between the various Crown coloration vs condition feather carotenoids (Table 3). Analyses of relative pigment concentrations showed that male kinglets exhibiting short- Among female kinglets, only variation in crown hue was wavelength crown hues deposited more 3?-dehydrolutein, explained by differences in individual condition (Table 4). lutein, and zeaxanthin than individuals with long- Females that migrated earlier and had fewer feather mites wavelength hues, and birds with medium-wavelength hues and higher body condition indices exhibited deeper yellow had intermediate levels (Fig. 4B). crown hues (i.e. shifted toward longer wavelengths; Correlations between condition variables and absolute Table 4). Among male kinglets, variation in both crown pigment concentrations in male kinglets showed that earlier saturation and hue were predicted by condition variables migrants deposited a higher concentration of canthaxanthin (Table 5). Males that arrived earlier had more saturated and (r 0.38, n 29, p 0.04) and echinenone (r 0.43, deeper orange crown coloration. Contrary to expectation, n 29, p 0.02) into feathers, and also tended to possess however, males that had smaller pectoral muscles also had higher concentrations of canary xanthophyll A (r 0.33, more saturated crown coloration (Table 5). Intercorrela- n 29, p 0.08) and lower concentrations of zeaxanthin tions among the condition variables do not appear to be (r 0.34, n 29, p 0.07; Fig. 5). These results are responsible for these patterns (Table 2). Feather growth rate consistent with the carotenoid content of feathers exhibiting was not related to crown coloration in males or females long-wavelength hues, and confirm that crown pigmenta- (Table 4, 5). tion is associated with arrival date. Discussion Sex- and age-specific differences in plumage coloration, morphology, and condition We found that crown coloration in golden-crowned king- lets is determined by the types and concentrations of We found no difference in crown brightness between the carotenoid pigments deposited into feathers, and is corre- sexes; however, male kinglets had significantly more lated with migration timing and other measures of 2 saturated (model: R 0.88, F3,180 460.5, pB0.0001; individual condition during migration. Male kinglets sex: F1,182 1021.2, pB0.0001) and deeper orange produced deeper orange crown coloration by depositing (longer-wavelength) crown coloration than females (model: greater concentrations of carotenoids into their feathers, 2 R 0.97, F3,180 1927.9, pB0.0001; sex: F1,182 especially ketocarotenoids that give orange and red colours 4201.9, pB0.0001; Fig. 1, 6). Males also exhibited greater to avian integuments (Stradi 1998, McGraw 2006). This variation in these colour metrics than did females (satura- strategy is consistent with research conducted on house tion: 2.4390.14 vs 1.7390.08, Levene’sF1,182 16.00, finches (Inouye et al. 2001), albeit within a smaller range of pB0.0001; hue: 569.0795.02 nm vs 518.3492.91 nm; natural hues. In theory, animals can display redder Levene’sF1,182 17.50, pB0.0001). Age and sex-by-age carotenoid coloration by increasing either absolute or interactions did not significantly influence crown colour relative ketocarotenoid concentrations. This means that variables. Among morphological variables, male kinglets male kinglets with ‘redder’ crowns could theoretically were significantly larger than females in terms of wing- deposit greater amounts of yellow xanthophylls than those 2 chord (model: R 0.37, F3,180 34.6, pB0.0001; sex: with ‘yellower’ crown feathers, yet maintain a higher ratio 2 F1,182 68.6, pB0.0001) and tarsus length (model: R of red carotenoids. Our results showed that kinglets with 0.08, F3,180 5.41, p 0.0014; sex: F1,182 4.51, p longer-wavelength (‘redder’) crown hues have reduced 0.04), and AHY individuals also tended to have longer concentrations as well as lower proportions of yellow tarsi (age: F1,182 3.45, p 0.07). After controlling for carotenoids in their feathers.
314 (A) 0.045
0.035
24
0.025
0.015 Absorbance units
0.005
1 6 300 350 400 ngth (nm) 450 500 Wavele 3.0 4.0 5.0 6.0 7.0 8.09.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 Minutes (B)
0.24 4 7 3 0.20
0.16
0.12
0.08 Absorbance units (3) 5 0.04
2 (3) 1 0.00 6 300 8 350 (nm) 400 450 length 500 Wave 3.0 4.0 5.0 6.0 7.0 8.09.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 Minutes
Figure 2. Representative three-dimensional HPLC chromatograms showing the carotenoid pigments found in the crown feathers of (A) female and (B) male golden-crowned kinglets. Bolded numbers above the peaks indicate identified carotenoids: 1) canary xanthophyll A, 2) 3?-dehydrolutein, 3) a-doradexanthin [peaks labelled ‘(3)’ are a-doradexanthin isomers], 4) lutein, 5) astaxanthin, 6) zeaxanthin, 7) canthaxanthin, 8) echinenone. Note the difference in scale of the y-axes.
Although the relative proportions of ketocarotenoids Furthermore, individuals may selectively incorporate revealed different patterns compared to their concentra- canthaxanthin over other carotenoids into their crown tions, levels of canthaxanthin were consistently higher for feathers (McGraw 2006). male kinglets with deeper orange crowns, suggesting that Males exhibiting a more yellow crown (i.e. shorter- canthaxanthin could be an important pigment for orange wavelength hues) incorporated a lower concentration of colour generation (McGraw and Toomey 2010). For carotenoids overall, as well as greater proportions of yellow insectivores such as members of the Regulidae, red feather hydroxycarotenoids. The transformation of dietary carote- colorants are believed to result from metabolic conversions noids into red pigments is thought to be energetically of yellow pigments harboured by prey items (Stradi 1998), expensive (Hill 1996, 2002, Olson and Owens 1998), and as opposed to the direct ingestion of red ketocarotenoids by research in house finches suggests that only individuals in aquatic birds (Fox 1962). It would be interesting to good nutritional condition are capable of extensive carote- determine whether increased levels of canthaxanthin in noid metabolism (Hill 2000). Since total carotenoid kinglets are due to greater access to the dietary precursor, content was similar for feathers with medium- and short- b-carotene (Stradi 1998, confirmed in Euplectes spp. by wavelength hues, perhaps the higher levels of plumage Prager et al. 2009) (e.g. better foraging skills or preferen- zeaxanthin in the latter group are the result of inadequate tially choosing prey items rich in this pigment). Alterna- carotenoid metabolism of this dietary precursor. Further tively, high-quality males may have more efficient research into the biochemical pathways of both plasma and absorption, circulation, or metabolism of b-carotene. feather carotenoids in this species would be fruitful for
315 a (A) 3.0 Long-wavelength hues Medium-wavelength hues 2.5 Short wavelength hues
) b –1 b 2.0
1.5 a a 1.0 b b b Mean concentration (mg g b 0.5
0 4,4’-ketocarotenoids 4-ketocarotenoids Hydroxycarotenoids TOTAL CAROTENOIDS (“red” pigments) (“orange” pigments) (“yellow” pigments)
(B) 50 a
40 ab
a ab b 30 b
20
Mean relative percentage (%) 10
0 4,4’-ketocarotenoids 4-ketocarotenoids Hydroxycarotenoids (“red” pigments) (“orange” pigments) (“yellow” pigments)
Figure 3. Comparisons of (A) concentrations and (B) relative percentages of carotenoid groups (mean9SE) in the feathers of male golden-crowned kinglets exhibiting different crown hues. Sample sizes are n 10 for all categories except long-wavelength hue concentration (n 9). Bars with different letters above indicate significant differences (pB0.05) from ANOVA with post hoc Tukey tests. understanding why pigment composition varies among (McGraw 2006). The conversion of dietary lutein and individuals. zeaxanthin into canary xanthophylls and 3?-dehydrolutein, Interestingly, crown pigmentation in golden-crowned which are yellow pigments that maximally absorb shorter kinglets differed from that of the closely-related goldcrest in wavelengths of light than their precursors (by 3 5 nm), has a number of important respects. In the latter species, lutein previously been reported in other passerine species with red/ and zeaxanthin produce the yellow coloration, and orange plumage (Hudon 1991, Inouye et al. 2001, a-doradexanthin, astaxanthin, and adonirubin comprise McGraw et al. 2004, Andersson et al. 2007, Hudon et al. the red pigments contributing to the orange crown 2007). coloration of males (Stradi 1998). Thus, most notably, Regarding the possible condition-dependence of carote- canthaxanthin was not isolated in goldcrest feathers, and we noid pigmentation in kinglets, arrival date was significantly did not find adonirubin in golden-crowned kinglet feathers associated with crown coloration in both sexes. While the in this study. Future research should focus on elucidating advantages of early spring migration are well-established whether the differences in feather carotenoids are due to (Lozano et al. 1996, Hasselquist 1998, Kokko 1999, dietary or enzymatic/genetic differences between these and Forstmeier 2002), the importance of migration timing in other Regulus species. In addition, we identified in golden- the fall is more difficult to ascertain, especially for a species crowned kinglets several other carotenoids that were absent such as the golden-crowned kinglet, which does not hold in the goldcrest: echinenone, canary xanthophyll A, and winter territories (Ingold and Galati 1997). Avoidance of 3?-dehydrolutein. Since echinenone is only present in small migrating predatory birds, such as northern shrikes Lanius amounts and shares a very strong positive relationship with excubitor (Cade and Atkinson 2002) and northern saw-whet canthaxanthin (Table 3), this orange ketocarotenoid is likely owls Aegolius acadicus (Rasmussen et al. 2008), as well as present as the incomplete transformation of b-carotene inclement weather later in the season, could be important
316 (A) 0.8 Long-wavelength hues Medium-wavelength hues a a Short wavelength hues
) 0.6 -1 ab b
0.4 b b a
Mean concentration (mg g (mg concentration Mean 0.2 b b a ab b a a b a c b 0 ne o ry a anthin Lutein x hinen c Can Astaxanthin Zea E xanthophyllA Canthaxanthin α-doradexanthin 3'-dehydrolutein (B) 25 a a ab
20 ab
b 15 b
10
a Mean relative percentage (%) percentage relative Mean
5 ab a b b b a a a a b b 0 in n n hin ei hi hyllA t lute Lu hop ro axant Canary e Astaxant hyd Z Echinenone xant Canthaxanthin '-de α-doradexanthin 3
Figure 4. Comparisons of (A) concentrations and (B) relative percentages of all identified carotenoids (mean9SE) in the feathers of male golden-crowned kinglets exhibiting different crown hues. Sample sizes are n 10 for all categories except long-wavelength hue concentration (n 9). Bars with different letters above indicate significant differences (pB0.05) from ANOVA with post hoc Tukey tests. factors in promoting early migration (Berthold 2001, and may thus be in better nutritional condition. If crown Martens and Pa¨ckert 2006). Moreover, high-quality in- coloration plays a role in competition and mate choice, dividuals may be able to accumulate sufficient energy stores more colourful males may be able to secure better breeding for earlier departure from the breeding grounds. In our territories or attract high-quality females earlier in the study, earlier male migrants grew their tail feathers faster season, which could advance their breeding and thus enable
Table 3. Pearson correlation coefficients between carotenoid pigment concentrations in the crown feathers of male golden-crowned kinglets (n 29).
Pigment Astaxanthin Canthaxanthin a-doradexanthin Echinenone Canary 3’-dehydrolutein Lutein Zeaxanthin xanthophyll A
Astaxanthin 0.54 ** 0.28 0.56 ** 0.30 0.21 0.07 0.16 Canthaxanthin 0.79 **** 0.85 **** 0.68 **** 0.39 * 0.2 0.06 a-doradexanthin 0.58 ** 0.85 **** 0.69 **** 0.54 ** 0.14 Echinenone 0.60 *** 0.16 0.09 0.19 Canary 0.53 ** 0.23 0.12 xanthophyll A 3’-dehydrolutein 0.52 ** 0.31 Lutein 0.55 ** Zeaxanthin
*pB0.05, ** pB0.01, *** pB0.001, **** pB0.0001.
317 (A) earlier fall migration. Indeed, different periods of the avian 10.0 annual cycle are likely to be inextricably linked (Marra et al. 1998, Norris et al. 2004a). 8.0 ) Contrary to our predictions for the condition- -1 dependence of crown coloration, male kinglets that pos- sessed smaller pectoral muscles had more richly saturated 6.0 crown coloration. This result could be interpreted multiple ways: firstly, crown coloration may not be causally linked to 4.0 condition or muscle mass in male kinglets. Secondly, pectoral muscles of migrating birds can vary in size relatively quickly, on the scale of days (Lindstrom et al.
Canthaxanthin (mg g ¨ 2.0 2000); therefore, pectoral score may not be an ideal measure of longer-term individual quality. A better measure of 0 quality might be an individual’s rate of rebuilding fat stores and muscle protein (Woodrey and Moore 1997, Jones et al.
01-Oct 08-Oct 15-Oct 22-Oct 29-Oct 05-Nov 12-Nov 2002); however, because we rarely recaptured kinglets on Arrival date the same or subsequent days, we could not collect these (B) measurements. Thirdly, it is possible that individuals with 2.0 more colourful crowns are in better condition and capable of rigorous, faster migratory flights, leading to depleted pectoral muscle size upon earlier arrival at our stopover site
) 1.5 (Schwilch et al. 2002). -1 While crown colour was more variable in male compared to female kinglets, the condition-dependence of this trait 1.0 at least for the condition measures we considered here was surprisingly more apparent for females. Female kinglets with deeper yellow crown hues were earlier migrants, with
Zeaxanthin (mg g slightly higher body condition indices and reduced wing 0.5 mite infestation. As a measure of size-adjusted body mass, a body condition index is commonly used to deduce an individual’s current nutritional state (Schulte-Hostedde 0 et al. 2005). However, the use of indices based on regression residuals requires satisfying several underlying assumptions 01-Oct 08-Oct 15-Oct 22-Oct 29-Oct 05-Nov 12-Nov (van der Meer and Piersma 1994, Green 2001), some of Arrival date which we were unable to test with our dataset. Thus, the Figure 5. Correlations between the arrival date of male golden- correlation between body condition index and female crown crowned kinglets to our study site during fall migration and hue may be over- or underestimated. It is unknown whether two major crown feather carotenoids: canthaxanthin (A) and the mites that we observed on the wing feathers of golden- zeaxanthin (B). crowned kinglets exist as parasitic or commensal symbionts (Proctor and Owens 2000). Although wing mites are unlikely to directly affect crown coloration, infestation status may be another correlate of general body condition. A previous study of nine passerine species found that birds 80 Males infested with higher mite loads also moulted into duller 70 Females plumage, grew shorter wings, and had smaller pectoral muscle scores, suggesting that feather mites are indeed 60 detrimental (Harper 1999). Feather mites can be trans- mitted between individuals in close proximity, such as 50 members of communal roosts, mates, and offspring
40 (Proctor and Owens 2000). Assuming that relative mite infestation is stable among individuals in the long term, it
Reflectance (%) 30 may be particularly unfavourable for females to harbour ectosymbionts that could be vertically transmitted to 20 offspring in their nests. The differences we documented between male and 10 female kinglets may provide some insight into the intensity
0 of sexual selection in this species. Studies on the quality- 300 400 500 600 700 indicating functions of ornamental traits have traditionally Wavelength (nm) focused on males (reviewed in Johnstone 1995). However, Figure 6. Reflectance spectra (mean9SE) for the crown evidence is now mounting that females can also possess coloration of male (n 116) and female (n 70) golden-crowned condition-dependent characters (reviewed in Amundsen kinglets. and Pa¨rn 2006), and our results provide further support
318 Table 4. Relationships between crown coloration and condition variables in female golden-crowned kinglets.
Dependent variable Model effects F b’ DF p
Crown brightness whole model (R2 0.09) 0.84 7, 59 0.56 age 0.0006 0.004 1, 65 0.98 time released 0.53 0.11 1, 65 0.47 arrival date 0.09 0.04 1, 65 0.77 body condition index 2.68 0.22 1, 65 0.11 pectoral 0.54 0.10 1, 65 0.47 wing mites 0.002 0.006 1, 65 0.97 feather growth rate 0.61 0.10 1, 65 0.44 Crown saturation whole model (R2 0.18) 1.81 7, 59 0.10 age 0.01 0.01 1, 65 0.92 time released 3.30 0.25 1, 65 0.07 arrival date 2.59 0.21 1, 65 0.11 body condition index 3.50 0.24 1, 65 0.07 pectoral 2.39 0.19 1, 65 0.13 wing mites 1.65 0.18 1, 65 0.20 feather growth rate 2.66 0.20 1, 65 0.11 Crown hue whole model (R2 0.23) 2.53 7, 59 0.02 age 3.36 0.24 1, 65 0.07 time released 0.98 0.13 1, 65 0.33 arrival date 4.20 0.26 1, 65 0.04 body condition index 3.20 0.22 1, 65 0.08 pectoral 0.01 0.01 1, 65 0.92 wing mites 5.17 0.30 1, 65 0.03 feather growth rate 1.19 0.13 1, 65 0.28 for this idea. Golden-crowned kinglets are serially mono- colourful males exhibit some aspects of quality that our gamous and both sexes are involved in breeding duties: study could not identify, such as better parental care. females are solely responsible for incubating and brooding, Furthermore, song and behaviour may also be important males are involved in nest/territory defence and mate- sexually selected traits in this species (Ingold and Galati feeding, and both sexes participate in nest-building and 1997, Martens and Pa¨ckert 2006). chick-feeding (Ingold and Galati 1997). This type of Our study is unique in terms of sampling a potentially breeding system would favour mutual mate choice wide-ranging population of migrating kinglets in varying (Andersson 1994, Johnstone et al. 1996). Our findings states of condition; however, uncertainty regarding the suggest that it may benefit male kinglets to select females origins and destinations of these birds constitutes a caveat. based on their condition-dependent crown coloration. Some of the condition variables we included, such as arrival As for female preferences, perhaps early-arriving, more date and fat and pectoral scores, may depend on the
Table 5. Relationships between crown coloration and condition variables in male golden-crowned kinglets.
Dependent variable Model effects F b’ DF p
Crown brightness whole model (R2 0.10) 1.70 7, 108 0.12 age 0.12 0.03 1, 114 0.73 time released 0.40 0.06 1, 114 0.53 arrival date 2.15 0.14 1, 114 0.15 body condition index 1.33 0.12 1, 114 0.25 pectoral 5.14 0.22 1, 114 0.03 wing mites 0.02 0.01 1, 114 0.90 feather growth rate 0.0009 0.003 1, 114 0.98 Crown saturation whole model (R2 0.15) 2.62 7, 108 0.02 age 0.17 0.04 1, 114 0.68 time released 0.60 0.08 1, 114 0.44 arrival date 9.21 0.29 1, 114 0.003 body condition index 0.008 0.009 1, 114 0.93 pectoral 5.97 0.23 1, 114 0.02 wing mites 0.69 0.08 1, 114 0.41 feather growth rate 0.22 0.04 1, 114 0.64 Crown hue whole model (R2 0.16) 2.93 7, 108 0.008 age 0.04 0.02 1, 114 0.84 time released 0.21 0.05 1, 114 0.65 arrival date 15.26 0.37 1, 114 0.0002 body condition index 0.008 0.009 1, 114 0.93 pectoral 1.57 0.12 1, 114 0.21 wing mites 0.005 0.007 1, 114 0.94 feather growth rate 0.03 0.02 1, 114 0.87
319 location of the breeding grounds, distances and rates of mechanisms and measurements. Harvard Univ. Press, migration, and time spent at stopover sites (Merila¨ and pp. 41 89. Svensson 1995, Lindstro¨m et al. 2000, Kelly 2006). In Andersson, S., Prager, M. and Johansson, E. I. A. 2007. addition, ectosymbiont load, body condition, feather Carotenoid content and reflectance of yellow and red nuptial growth rate, and even crown coloration may differ between plumages in widowbirds (Euplectes spp.). Funct. Ecol. 21: populations, depending on the ecology of each habitat and 272 281. Ardia, D. R. 2005. Super size me: an experimental test of the the available resources (Strong and Sherry 2000, Simon factors affecting lipid content and the ability of residual body et al. 2005). Nevertheless, this study has revealed some mass to predict lipid stores in nestling European starlings. interesting correlations between carotenoid coloration and Funct. Ecol. 19: 414 420. our migratory measurements, and the various possible Badyaev, A. V. 1997. Altitudinal variation in sexual dimorphism: a interpretations of these results deserve further attention. new pattern and alternative hypotheses. Behav. Ecol. 8: 675 In summary, our study is the first to identify the feather 690. carotenoid composition in a North American regulid Barlein, F. 1995. Manual of field methods, revised ed. species. We determined how carotenoid pigments contri- European-African Songbird Migration Network. bute to sexual dichromatism and intrasexual colour varia- Battley, P. F., Piersma, T., Rogers, D. I., Dekinga, A., Spaans, B. tion in golden-crowned kinglets. Our findings provide and van Gils, J. A. 2004. Do body condition and plumage further support for the condition-dependent nature of during fuelling predict northwards departure dates of great carotenoid-based coloration and the quality-signalling knots Calidris tenuirostris from north-west Australia? Ibis potential of female ornamental traits. Our data also suggest 146: 46 60. that differences in migration timing in the fall may be Berthold, P. 2001. Bird migration: a general survey. Oxford correlated with individual fitness. In conjunction with data Univ. Press. Britton, G. 1985. General carotenoid methods. In: Law, J. H. on large-scale geographic colour variation in kinglets (Chui and Rilling, H. C. (eds), Methods in enzymology. Vol. 111: and Doucet 2009), our study provides a more comprehen- steroids and isoprenoids, part B. Academic Press, pp. 113 sive and mechanistic view of how plumage coloration can 149. vary intraspecifically. Because carotenoid coloration is an Britton, G., Liaaen-Jensen, S. and Pfander, H. 2004. Carotenoids important trait used in sexual selection, especially mate handbook. Birkha¨user. choice (Hill 2006a), future research on variation in Brown, M. E. 1996. Assessing body condition in birds. Curr. carotenoid coloration should provide a thorough investiga- Ornithol. 13: 67 135. tion of environmental and genetic factors that serve to Buehler, D. M. and Piersma, T. 2008. Travelling on a budget: maintain signal honesty. predictions and ecological evidence for bottlenecks in the annual cycle of long-distance migrants. Phil. Trans. R. Soc. B 363: 247 266. Acknowledgements Feather collection was approved by the Cade, T. J. and Atkinson, E. C. 2002. Northern shrike (Lanius Canadian Wildlife Service Migratory Bird Regulations (Scientific excubitor). In: Poole, A. (ed.), The birds of North America Permit CA 0209) and the Univ. of Windsor Animal Care online. Cornell Lab of Ornithology,
320 Hasselquist, D. 1998. Polygyny in great reed warblers: a long-term mass changes during flight, fasting and fuelling. J. Exp. Biol. study of factors contributing to male fitness. Ecology 79: 203: 913 919. 2376 2390. Lozano, G. A., Perreault, S. and Lemon, R. E. 1996. Age, arrival Hill, G. E. 1996. Redness as a measure of the production cost of date and reproductive success of male American redstarts ornamental coloration. Ethol. Ecol. Evol. 8: 157 175. Setophaga ruticilla. J. Avian Biol. 27: 164 170. Hill, G. E. 2000. Energetic constraints on expression of Marra, P. P., Hobson, K. A. and Holmes, R. T. 1998. Linking carotenoid-based plumage coloration. J. Avian Biol. 31: winter and summer events in a migratory bird by using stable- 559 566. carbon isotopes. Science 282: 1884 1886. Hill, G. E. 2002. A red bird in a brown bag: the function and Martens, J. and Pa¨ckert, M. 2006. Family Regulidae (kinglets and evolution of colorful plumage in the house finch. Oxford firecrests). In: del Hoyo, J., Elliot, A. and Christie, D. (eds), Univ. Press. Handbook of the birds of the world. Vol. 11: Old World Hill, G. E. 2006a. Female mate choice for ornamental coloration. flycatchers to Old World warblers. Lynx Edicions, pp. 330 In: Hill, G. E. and McGraw, K. J. (eds), Bird coloration. 349. Vol. II: function and evolution. Harvard Univ. Press, pp. 137 McGraw, K. J. 2006. Mechanics of carotenoid-based coloration. 200. In: Hill, G. E. and McGraw, K. J. (eds), Bird coloration. Hill, G. E. 2006b. Environmental regulation of ornamental Vol. I: mechanisms and measurements. Harvard Univ. Press, coloration. In: Hill, G. E. and McGraw, K. J. (eds), Bird pp. 177 242. coloration. Vol. I: mechanisms and measurements. Harvard McGraw, K. J. and Gregory, A. J. 2004. Carotenoid pigments in Univ. Press, pp. 507 560. male American goldfinches: what is the optimal biochemical Hill, G. E., Inouye, C. Y. and Montgomerie, R. 2002. Dietary strategy for becoming colourful? Biol. J. Linn. Soc. 83: 273 carotenoids predict plumage coloration in wild house finches. 280. Proc. R. Soc. B 269: 1119 1124. McGraw, K. J. and Toomey, M. B. 2010. Carotenoid accumula- Houde, A. E. 1997. Sex, color, and mate choice in guppies. tion in the tissues of zebra finches: predictors of integumentary Princeton Univ. Press. pigmentation and implications for carotenoid allocation Hudon, J. 1991. Unusual carotenoid use by the western tanager strategies. Physiol. Biochem. Zool. 83: 97 109. (Piranga ludoviciana) and its evolutionary implications. Can. McGraw, K. J., Hill, G. E. and Parker, R. S. 2003. Carotenoid J. Zool. 69: 2311 2320. pigments in a mutant cardinal: implications for the genetic Hudon, J., Ancia˜es, M., Bertacche, V. and Stradi, R. 2007. and enzymatic control mechanisms of carotenoid metabolism Plumage carotenoids of the pin-tailed manakin (Ilicura in birds. Condor 105: 587 592. militaris): evidence for the endogenous production of rhodox- McGraw, K. J., Wakamatsu, K., Clark, A. B. and Yasukawa, K. anthin from a colour variant. Comp. Biochem. Physiol. B 2004. Red-winged blackbirds Agelaius phoeniceus use carote- 147: 402 411. noid and melanin pigments to color their epaulets. J. Avian IBA Canada 2004. Important bird areas of Canada: Holiday Biol. 35: 543 550. Beach/Big Creek Conservation Areas, Windsor, Ontario. Merila¨, J. 1997. Fat reserves and moult-migration overlap in IBA Canada,
321 Pe´rez-Tris, J., de la Puente, J., Pinilla, J. and Bermejo, A. 2001. Senar, J. C. 2006. Color displays as intrasexual signals of Body moult and autumn migration in the barn swallow aggression and dominance. In: Hill, G. E. and McGraw, Hirundo rustica: is there a cost of moulting late? Ann. Zool. K. J. (eds), Bird coloration. Vol. II: function and evolution. Fenn. 38: 139 148. Harvard Univ. Press, pp. 87 136. Piersma, T. and Jukema, J. 1993. Red breasts as honest signals of Simon, A., Thomas, D. W., Bourgault, P., Blondel, J., Perret, P. migratory quality in a long-distance migrant, the bar-tailed and Lambrechts, M. M. 2005. Between-population differences godwit. Condor 95: 163 177. in nestling size and hematocrit level in blue tits (Parus Prager, M., Johansson, E. I. A. and Andersson, S. 2009. caeruleus): a cross-fostering test for genetic and environmental Differential ability of carotenoid C4-oxygenation in yellow effects. Can. J. Zool. 83: 694 701. and red bishop species (Euplectes spp.). Comp. Biochem. Stradi, R. 1998. The colour of flight: carotenoids in bird plumage. Physiol. B 154: 373 380. Solei Gruppo Editoriale Informatico. Proctor, H. and Owens, I. 2000. Mites and birds: diversity, Stradi, R., Celentano, G. and Nava, D. 1995. Separation and parasitism and coevolution. Trends Ecol. Evol. 15: 358 364. identification of carotenoids in bird’s plumage by high- Pyle, P. 1997. Identification guide to North American birds. performance liquid chromatography diode-array detection. Part I: Columbidae to Ploceidae. Slate Creek Press. J. Chromatogr. B 670: 337 348. Raberg, L., Grahn, M., Hasselquist, D. and Svensson, E. 1998. ˚ Strong, A. M. and Sherry, T. W. 2000. Habitat-specific effects of On the adaptive significance of stress-induced immunosup- food abundance on the condition of ovenbirds wintering in pression. Proc. R. Soc. B 265: 1637 1641. Jamaica. J. Anim. Ecol. 69: 883 895. Rasmussen, J. L., Sealy, S. G. and Cannings, R. J. 2008. Northern Svensson, E. and Merila¨, J. 1996. Molt and migratory condition in saw-whet owl (Aegolius acadicus). In: Poole, A. (ed.), The birds of North America online. Cornell Lab of Ornithology, blue tits: a serological study. Condor 98: 825 831.
322