Carotenoid Pigments in Male House Finch Plumage in Relation to Age, Subspecies, and Ornamental Coloration

The Auk 118(4):900-915, 2001

CAROTENOID PIGMENTS IN MALE HOUSE FINCH PLUMAGE IN RELATION TO AGE, SUBSPECIES, AND ORNAMENTAL COLORATION

CARON Y. INOUYE,L6 GEOFFREYE. HILL, 2'5RICCARDO D. STRADI,3 AND ROBERT MONTGOMERIE 4 •Departmentof OrganismicBiology, Ecology and Evolution, University of California,P.O. Box 951606, Los Angeles, California90095-1606, USA; 2Departmentof BiologicalSciences and Alabama Agricultural Experiment Station, 331 FunchessHall, AuburnUniversity, Auburn, Alabama 36849, USA; 3Institutodi ChimicaOrganica, Universita Degli Studidi Milano,Milano, MI 20133,Italy; and 4Departmentof Biology,Queenk University, Kingston, Ontario K7L 3N6, Canada

ABSTRACT.--Likemales of many bird species,male House Finches(Carpodacus mexicanus) havepatches of featherswith ornamentalcoloration that are due to carotenoidpigments. With- in populations,male HouseFinches vary in expressionof ornamentalcoloration from paleyel- low to bright red, which previousresearch suggested was the result of variationin typesand amountsof carotenoidpigments deposited in feathers.Here we usedimproved analytical tech- niquesto describetypes and amountsof carotenoidpigments present in thatplumage. We then used thosedata to make comparisonsof carotenoidcomposition of feathersof male House Finchesat three levels:among individual maleswith different plumagehue and saturation, betweenage groupsof malesfrom the samepopulation, and betweenmales from two subspeciesthat differ in extent of ventral carotenoidpigmentation (patch size): large-patched C. m.frontalis from coastalCalifornia and small-patchedC. m.griscomi from Guerrero,Mexico. In all age groupsand populations,the ornamentalplumage coloration of male HouseFinches resultedfrom the same13 carotenoidpigments, with 3-hydroxyechinenone and luteinbeing the mostabundant carotenoid pigments. The compositionof carotenoidsin featherssuggested that House Finchesare capableof metabolictransformation of dietary forms of carotenoids. The hue of male plumagedepended on componentcarotenoids, their relativeconcentrations, and total concentrationof all carotenoids.Most 4-keto (red) carotenoidswere positivelycorrelated with plumage redness,and most yellow carotenoidpigments were negativelyassoci- ated with plumageredness, although the strengthof the relationshipfor specificcarotenoid pigmentsvaried amongage groupsand subspecies.Using age and subspeciesas factorsand concentrationof eachcomponent carotenoid as dependentvariables in a MANOVA, we found a distinctivepigment profile for eachage group within eachsubspecies. Amongfrontalis males, hatch-yearbirds did not differ from adultsin meanplumage hue, but theyhad a significantly lower proportionof red pigmentsin their plumage,and significantlylower levelsof the red pimentsadonirubin and astaxanthin,but significantlyhigher levels of the yellow pigmentze- axanthin, than adult males. Among griscomimales, hatch-year birds differed from adultsin plumagehue but not significantlyin pigment composition,though in generaltheir feathers had lower concentrationsof red pigmentsand higherconcentrations of yellow pigmentsthan adult males.Both adult and hatch-yearfrontalismales differed from griscomimales in having significantlyhigher levelsof most yellow carotenoidpigments and significantlylower levels of mostred carotenoidpigments. Variation in pigmentprofiles of subspecies and ageclasses may reflectdifferences among the groupsin carotenoidmetabolism, in dietary accessto carotenoids,or in exposureto environmentalfactors, such as parasites, that may affectpigmen- tation. Received18 January1999, accepted 11 June2001.

CAROTENOIDPIGMENTS ARE responsiblefor plumage.Birds obtainthose carotenoids exclu- the bright red, orange,and yellow colorationof sively through their diet. No animal has been shownunequivocally to be capableof in vivo s Addresscorrespondence to this author. synthesisof carotenoids(Goodwin 1984,1986; E-mail: [email protected] Schiedt 1990). In birds, dietary carotenoids 6Present address:Department of BiologicalSci- may either be depositeddirectly into feathers ences, California State University, Hayward, 25800 or chemically changed from ingested forms Carlos Bee Boulevard, Hayward, California 94542, prior to pigment deposition,typically by ad- USA. dition or elimination of oxygen groupsto one

900 October2001] HouseFinch Carotenoids 901 or both end rings of the molecule(Davies 1985, males in expression of these displays (Hill Goodwin 1986, Tyzckowski and Hamilton 1992, 1996a, 2002). 1986a, b; Brush 1990, Schiedt 1990). Feedingexperiments conducted with captive The House Finch (Carpodacusmexicanus) is a House Finches have demonstrated that varia- sexually dichromaticpasserine bird speciesin tion among males in plumage hue and satura- which males display bright, carotenoid-based tion is dependentupon carotenoidaccess dur- patches of color on their crowns, throats, ing molt (Brush and Power 1976, Hill 1992, breasts,and rumps, and male House Finches 1993a).When malesare held in flight cagesand vary in expressionof that ornamental colora- fed a standardized diet, variation in ornamen- tion from a bright red to a dull yellow (Mich- tal plumage colorationis minimized (Hill 1992, ener and Michener 1931, Hill 1990, 1993a). The 1993a).Moreover, after molting in captivity un- carotenoid pigments responsiblefor colorful der conditions of standardized carotenoid ac- plumage in the House Finch and the pigmen- cess,males from populationsthat are typically tary basisfor variationamong males in expres- drab in colorationgrow ornamentalplumage sion of that coloration were first studied by coloration that is indistinguishablefrom that Brushand Power(1976). They attributedplum- grown by malesfrom populationsthat are typ- age colorvariation to differencesin constituent ically bright in coloration (Hill 1993a). How- carotenoids in feathers. Red birds had the most ever, the degree to which accessto dietary ca- complexassemblage of pigments,consisting of rotenoids affects expression of carotenoid- B-carotene,a group of unidentifiedmixed xan- basedplumage colorationin the wild remains thophylls, orange isocryptoxanthin,and red controversial (Hill 1994c, 2002; Hudon 1994a, echinenone;orange birds had the same subset Thompson et al. 1997, Inouye 1999). Birds are of carotenoidswithout echinenone;and yellow capableof endogenousmodification of ingest- birds lacked both echinenoneand isocryptox- ed carotenoidsprior to depositioninto target anthin. Recent analysesof severalcongeneric tissues (Fox et al. 1969, Davies 1985, Schiedt et finch speciesof the PalearcticCarduelinae done al. 1985,Tyczkowski and Hamilton 1986b,c, d; by Stradiet al. (1995a,b; 1996,1997), using new Brush 1990, Hencken 1992). Therefore, mecha- analyticaltechniques, revealed a more complex nisms involved in the digestion, absorption, pattern that differed substantiallyfrom that de- transport, modification,and depositionof die- scribedby Brushand Power (1976). tary carotenoidsmay contribute to plumage The proximate basis of variation in caroten- color variation (Hill 1999, 2002). Furthermore, oid-basedplumage coloration in HouseFinches viral, bacterial, and coccidial infectionsmay is of interestbeyond improved understanding have a significant effect on expressionof or- of the physiologicalcontrol of avian pigmen- namental plumage colorationby male House tation. Plumageredness in House Fincheshas Finches (Thompson et al. 1997, Nolan et al. been shown to be a primary criterion used by 1998, Hill and Brawner 1998, Brawner et al. females in choosingmates (Hill 1990, 1991, 2000). 1994a). In addition, plumage brightness in There is substantial variation in expression male House Finches is correlated with overwin- of ornamental plumage coloration in House ter survival (Hill 1991), nutritional condition Finchesnot just among males within popula- during molt (Hill and Montgomerie1994), par- tions,but alsoamong populations and subspe- asiteload (Thompsonet al. 1997,Brawner et al. cies (Moore 1939, Hill 1993a). There are ap- 2000), and provisioningof femalesduring in- proximately 15 subspeciesof House Finchesin cubation(Hill 1991).It has been proposedthat North America (Moore 1939, Hill 1996b), each male plumagebrightness is an honestsignal of of which has had a unique evolutionaryhistory male condition, becausecarotenoids may be for thousandsof years(Moore 1939).Each sub- scarce resources in the environment and carot- speciesis characterizedby specificplumage enoid-basedcolor displays may be costly to traits, some of which involve carotenoid color- produce(Hill 1994b,1996a, 2002). A thorough ation (Moore 1939,Hill 1996b).Two subspecies understandingof the signal content of carot- are studied in this paper: C. m.frontalis, origi- enoid-basedornamental displays can only be nally native to coastalCalifornia but now in- achieved,however, through an understanding troduced to the Hawaiian Islands and the east- of the proximate control of variation among ern United States and Canada, and C. m. 902 INOUYEET AL. [Auk, Vol. 118

griscomi,found in a relatively small region of sis for differencesbetween subspeciesin carot~ southern Mexico. Male House Finches from the enoid display,we comparedtheir patch sizes frontalispopulation have much more extensive and mean hues, total carotenoid abundances, ventral carotenoidpigmentation (larger patch and carotenoidcomposition of their feathers. size) than thosefrom the griscomipopulation (Moore 1939, Hill 1993a), but some adult male METHODS griscomihave more intense red colorationthan any malefrontalis (Hill 1993a). Samplecollection.--We collected 62 hatch-year(-<1 In contrastto the plasticity of expressionof year) and 69 adult male House Finchesduring 2-13 the color(hue, brightness, and saturation)of ca- August 1992from two locations-12 km apart in San JoseCounty, California (Coyote Creek Riparian Sta- rotenoid-basedplumage coloration, differences tion in Alviso and a private residencein the city of betweenfrontalisand griscomimales in expres- San Jose)and 59 hatch-yearand 32 adult males dur- sion of size of ventral patchesof ornamental ing 9-16 September1992 from varioussites within 20 colorationreflect fixed geneticdifferences be- km of Chilpancingo,Guerrero, Mexico. Individuals tween populations.When they are fed a low- were captured with mist nets or feeder traps. All carotenoid diet, both frontalis and griscomi birds were aged postmortemon the basisof degree males grow drab yellow plumage; when they of ossificationof the skull (Pyle et al. 1987).Hatch- are fed a red-carotenoid-supplementeddiet, year birds had incompletelyossified skulls and had they grow bright red feathers(Hill 1993a).Re- hatched in the same calendar year in which they were collected.Adults had completelyossified skulls gardlessof diet treatmentand plumage color, and were one or more yearsold. The sexof eachbird griscomimales always display a small patchof was confirmedby gonadal examination. ornamental color, and frontalis males always The plumage color of eachbird was quantified by display a relatively large patch of color (Hill visual comparisonof feathers to color plates in the 1993a). Moreover,hybrid males produced by MethuenHandbook of Colour(Kornerup and Wanschef crossinga griscomifemale with afrontalismale 1983;see Hill 1990, 1992, 1998 for a detailed descrip- showeda patchsize intermediate to the parent tion of the colorscoring methods). Birds used in this types (Hill 1993a).Thus, there are somefixed study were undergoing molt when they were col- geneticdifferences between subspecies at least lected. Most hatch-yearmales had only a relatively in distribution of carotenoidpigments in the smallpatch of newly emergedbreast feathers for col- or scoring. For thosebirds with limited ornamental plumage. Whether there are also differences plumage, we scored only the hue of their breast among subspeciesin types and amountsof ca- plumage. For most adult males and a few hatch-year rotenoid pigments that color feathershas not males, molt had progressedfar enough to allow us previouslybeen investigated. to estimate a completeplumage color score,includ- In this study, we identified and quantified ing hue, saturation,and tone (=chroma or bright- the carotenoidsresponsible for male House ness) following Hill (1992). Thus, we analyzed the Finch plumage coloration in the subspecies relationshipbetween plumage hue andplumage pig- frontalisand griscomi.Our objectiveswere, first, mentsfor all males,but we analyzedthe relationship to elucidatethe pigmentary basisfor extreme between saturationand plumage pigmentsfor only 23 adult frontalismales and only 17 adult griscomi variation in expressionof ornamentalplumage males. Tone, the third of the tristimulus color de- colorationamong males observedwithin age scriptors(Hill 1998),varied little amongmales in our classesand subspecies.To do that, we investi- visual assessmentand was not used in comparisons. gated how types and amounts of carotenoid After color scoring, individuals were euthanized, pigments in feathersaffected both hue (red- and each bird's pelt was collected.Pelts were then ness) and saturation of plumage coloration coveredwith sodium chloride and wrapped in alu- among individual males. Alhough we studied minum foil to protect feathersfrom prolonged ex- males in only two subspecies,a wide range of posure to light, which can result in photo-oxidation hues,saturations and patchsizes of ornamental of the pigments. plumage color has been recorded in frontalis Determinationof absoluteand relativecarotenoid concentrations.--Feathershaving yellow, orange,or red and griscomi(Hill 1993a). Second,within both color (indicative of carotenoid pigmentation) were subspecies,we comparedcarotenoid pigments pluckedfrom the pelt, washedwith a commercialde- of hatch-yearversus older malesto analyze ex- tergent solution(0.1% w/v), and air dried. Colored tent to which featherpigmenation was affected barbules were then cut from the feathers and by age.Finally, to investigatethe proximateba- weighed. A 5 mg aliquot of colored barbules was October2001] HouseFinch Carotenoids 903 washed in hexane,then finely ground, using an ul- trasonichomogenizer, in methanol to solubilize the o o pigments. The resulting fluid extract was filtered, evaporatedunder nitrogen,and storedin the dark at -20øC. Carotenoid pigments were isolated via high per- formanceliquid chromatography(HPLC) using two sequentialreverse-phase C18 columns(250 x 4 mm I.D.). The mobile phasewas 70:30 acetonitrile:meth- 09 anol administered at a flow rate of 0.5 ml min •. E1- uents were scannedat wavelengthsbetween 230 and 07 600 nm with a diode array detector.Peak areaswere integrated at 450 nm. Data were recorded as three- 0.5 dimensionalchromatograms using HP Chem Soft- ware (Hewlett-Packard, Palo Alto, California). Quantitative determination of carotenoids was completed using visible-light spectrophotometry. Once a component carotenoid was isolated via

HPLC, it was evaporatedto drynessand redissolved '• 1200 c- in methanol, then the visible spectrum of the pigment was recorded. The carotenoid concentration in feathers(micrograms carotenoids per gramsfeather) •,• ,•oo was calculatedaccording to the formula (A.... X vol- ume of extract [ml] X 104] / (E X feathermass [g]), 400 where Amax is the absorbancerecorded at the maxi- mum wavelength(Xmdx) of the pigment sample,and o E is the extinctioncoefficient at 1% per centimeterof AHY HY AHY HY the relevant carotenoid in the solvent (methanol). frontalis griscomi Values for E for most carotenoidsare published in Davies (1976) but an E of 2200 was used when the FIG. 1. Comparison of adult male House Finches precisevalue was unknown (e.g.for 3-hydroxy-echi- from the subspeciesfrontalis (California) and gris- nenone).Areas under the peaks recordedat the time corni(Mexico). (A) Breasthue (scoredby comparison with color chips, see text) where yellow is 4-5, or- of HPLC analysiswere determined and used to cal- culate both absolute carotenoid concentrations and angeis 6-8 and red is 9-11). (B) Patchsize (propor- relative abundances of individual carotenoids in tion of ventral surface with carotenoid pigmentation). (C) Total carotenoidpresent in plumage (see each sample. text for methods). Plotted are the medians, 10th, All biochemicalextractions and analysesof carot- 25th, 75th percentiles,90th percentiles,and outlying enoidswere performedby R.D.S. and coworkersat data points. the University of Milan. A detailed descriptionof the methods is given in Stradi et al. (1995a). All stan- RESULTS dards, except 4-oxo-rubixanthin, were generously donatedby Hoffman-LaRoche(Basel, Switzerland). Patchsize, plumage hue, and age effects.--As was Statisticalanalyses.--According to Hill (1998), cor- relation betweenhue scoresderived by visual com- found in a previousstudy comparinggriscomi parison to the plates in Kornerup and Wanscher and frontalis populations (Hill 1993a), males (1983) and hues measuredwith a spectrometerare from those two subspeciesdiffered in both significantlypositive (n = 55, r = 0.96, P = 0.0001) mean proportion of ventral surfacecovered by and linear, so hue scores were considered continuous carotenoidpigmentation (patch size) and mean variables (rather than ranks) in all analyses. That plumagecolor (hue;Fig. 1). Adult malesfrom permitted use of parametric statistical analyses the griscomisubspecies had significantlysmall- when relevantassumptions could be met. For anal- er patchesof color (Mann-Whitney U-test, z: ysesinvolving the surveyof many repetitionsof the 6.73, P = 0.0001, n: 69, 39; Fig. 1) but signif- same statistical test across several variables, we used icantly redder breasthues (z = 7.81,P = 0.0001, the sequentialBonferroni correction to assessstatis- n = 65, 30; Fig. 1) than adult males from the tical significanceof P-values (tablewide o• = 0.05; frontalissubspecies. Similarly, hatch-year males Rice 1989). from the griscomisubspecies had significantly 904 INOUYEET AL. [Auk,Vol. 118

TABLE1. Commonand structuralnames of carotenoidpigments isolated from the feathersof HouseFinch of two subspecies,Carpodacus mexicanus frontalis and C. m. griscomi.Main absorptionmaxima in nanometers in methanol are indicated for each carotenoid.

Commonname Structure Absorptionmaxima (nm) RED PIGMENTS Astaxanthin 3,3'-dihydroxy-fi,•-carotene-4, 4'-dione 474 Adonirubin 3-hydroxy-•, •-carotene-4,4'-dione 474 Canthaxanthin •, fi-carotene-4, 4'-dione 473 4-Oxo-rubixanthin 3-hydroxy:•, q•-caroten-4-one 438 47O 49O 3-Hydroxy-echinenone 3-hydroxy-•, •-caroten-4-one 465 Echinenone • fi-caroten-4-one 461 YELLOW PIGMENTS Canary xanthophylls e,e-caroten-3, 3'-dione 416 440 470 3'-hydroxy-e,e-caroten-3-one 416 440 470 3'-Dehydro-lutein 3-hydroxy-g,e-caroten-3'-one 424 445 474 Lutein •, e-carotene-3, 3'-diol 421 445 473 Zeaxanthin •, •-carotene-3, 3'-diol 422 450 481 fi-Cryptoxanthin •, fi-caroten-3-ol 428 449 473 fi-Carotene [•, [•-carotene 427 449 475 redderbreast plumage than hatch-yearfrontalis agreedwith standards.A typical three-dimen- males (z = 4.29, P = 0.0001,n = 65, 59; Fig. 1). sional chromatogramof a red male selected Fewhatch-year males had grownsufficient new from theSan Jose, California population offron- plumagewhen they were scoredto allow an es- talis is shown in Figure 2 and illustrateschar- timationof patchsize; yet for the small sample acteristicretention times and spectraof plum- available,hatch-year males of the griscomisub- age carotenoids.In both subspecies,the most specieshad significantlysmaller patchesthan abundantcarotenoids were lutein, appearing at frontalismales (z = 2.24,P = 0.03,n = 3, 5; Fig. a retention time of 12.0 min, and 3-hydroxy- 1). echinenone, with a retention time of 21.5 min. Unlikefrontalis males, griscomi males exhibit For the purposeof later analyses,we classi- delayedplumage maturation,in which hatch- fied thesecarotenoids into (1) red pigments(4- year males grow only small patchesof orna- keto-carotenoids)and (2) yellowpigments (Ta- mentalplumage (Hill 1996b).Not surprisingly, ble 1). We make that distinction and use the therefore,adult maleshad significantlyredder terms"yellow" and "red" pigmentsbecause it breast patchesthan hatch-yearmales in gris- aids in understandingthe combinedcontribu- comi(z = 3.51, P = 0.0004,n = 32, 59; Fig. 1), tion of different classesof pigments to the as well as significantlylarger mean patch sizes plumage coloration of House Finches,even (z = 2.62, P = 0.009,n = 5, 35; Fig. 1). Among thoughsome yellow carotenoids can produce a frontalismales, there were no significantdiffer- range of huesfrom yellow to orangeand some encesin mean breasthue or mean patch sizes red carotenoidscan appearorange depending between age classes(hue: z = 0.09, P = 0.93, n on their concentration(see Stradi 1998). The = 62, 69; patch size: z = 0.19, P = 0.85, n = 3, componentcarotenoids we classedas red ap- 69; Fig. 1). pearedas red bandsand thoseclassed as yel- Componentcarotenoids and individualvariation low appearedas yellow bands on thin-layer in plumagehue. In both subspecies,the same chromatographyplates (C. Y. Inouye pers. 13 carotenoids were extracted from feathers: obs.). See also Stradi (1998) for characteristic astaxanthin,adonirubin (=phoenicoxanthin), huesof carotenoidpigments. canthaxanthin,4-oxo-rubixanthin, 3-hydroxy- For both hatch-year and adult males from echinenone, echinenone, e,e-caroten-3,3'-dione, both subspecies,the concentrationof specific 3'-hydroxy-e,e-caroten-3-one,3'-dehydrolutein, red carotenoidpigments tended to be positive- lutein, zeaxanthin,[3-cryptoxanthin, and [3-car- ly correlatedwith plumageredness, whereas otene (Table1). All 13 componentcarotenoids concentrationof specificyellow carotenoid pig- were identifiedby their retentiontimes during ments tended to be negativelycorrelated with HPLC and spectral characteristicsat )kmax that plumageredness (Table 2). Although few of October2001] HouseFinch Carotenoids 905

(A)

I•'lutein

13 : 12 11 -' 10

(B)

FIG. 2. Three-dimensionalchromatograms exhibiting the typical carotenoidpatterns of a red male House Finch from San Jose,California. Retentiontime during HPLC is plotted along the x-axis and is separated horizontally into (A) 0-16 min and (B) 16-30 min. Absorbance(unitless) is plotted along the z-axis, and its rangediffers for (A) 0-250 and (B) 0-500. Wavelength(nanometers) is shownalong the y-axis.The two most prominent peaks are indicated:the yellow carotenoid,lutein, with a retention time of 12 rain, and the red carotenoid,3-hydroxy-echinenone, with a retention time of 21.5 min. 906 INOUYEET AL. [Auk, Vol. 118

TABLE2. Pearsoncorrelation coefficients between the concentrationof specificplumage carotenoids and the hue of feathersfrom which theywere extractedfor hatch-year(HY) and adult (AHY) malesfrom two House Finch subspecies.

C. m.frontalis C. m. griscomi Carotenoid HY (n = 39) AHY (n = 19) HY (n = 28) AHY (n = 15) Red Carotenoids Astaxanthin 0.51' 0.08 0.14 -0.01 Adonirubin 0.43 0.15 0.44 -0.15 Canthaxanthin 0.19 0.64 0.46 -0.04 4-Oxo-rubixanthin 0.26 0.51 0.51 -0.32 3-Hydroxy-echinenone 0.57• 0.18 0.53 -0.20 Echinenone 0.18 0.25 0.22 0.23

Yellow Carotenoids e, e-caroten-3, 3'-dione 0.05 -0.03 -0.19 0.51 3'-hydroxy-e,e-caroten-3-one -0.08 -0.45 -0.21 0.45 3'-Dehydro-lutein -0.08 -0.46 -0.23 0.08 Lutein -0.10 -0.46 -0.33 0.09 Zeaxanthin 0.09 0.002 -0.41 0.22 [3-Cryptoxanthin 0.11 0.11 0.06 -0.09 [g-Carotene -0.05 -0.39 -0.22 -0.13

0.05 after Bonferroni correction.

those correlationswere significant after Bon- ferroni correction,overall trend is highly significant with 19 of 24 correlations between hue and red pigments positive and 17 of 28 with yellow pigmentsnegative (Fisher'sexact test, P ---•-• • AHY ,lll,l,l,l,l,11[11 I = 0.005). Note, however, that relative contri- 12 bution of various component carotenoidsvar- griscomi © o ied among age groups and subspecies.For in- 10 ooc)•l•.c• stance, 3-hydroxy-echinenone was the best predictor of male breast patch hue in hatch- ß o year frontalismales but was a relatively poor predictorof breasthue for adultfrontalismales, 0...-" 0 0 0 though both correlationswere positive (Table o 2). Likewise, 3-hydroxy-echinenonewas the best predictor of male breast patch hue in ß i . i . i i hatch-yeargriscomi males, but was a poor pre- frontalis dictor of breast hue for adult griscomimales, 10 and in the oppositedirection (Table2). Propor- tion of red pigmentsin the plumagewas a good predictor of breast patch hue for both hatch- oeoc year and adultfrontalis(r = 0.60, P = 0.0001,n = 39, and r; 0.48, P = 0.04, n = 19, respec- tively) and hatch-yeargriscomi (r; 0.63, P = oo o 0.0004,n = 28) males (Fig. 3). Thus, in general, redder males had a higher proportion of 4- ' i ß i ß i , i ß i , i ß i ' i , i , 0 0.2 0.4 0.6 0.8 1.0 keto-carotenoids,but the relationshipdid not hold for adult griscomimales (r = -0.19, P = Proportionof red pigments 0.49, n = 15). Overall, relationshipbetween hue FIG. 3. The relationshipbetween breast hue and and proportionof red pigmentswas significant proportion of 4-keto (red) carotenoidpigments for and positive (ANCOVA, F = 24.7, df = 4 and hatch-yearand adult male House Finchesfrom the 96, P < 0.0001) with no significantdifferences subspecies(A) C. m.frontalis and (B) C. m. griscomi. October2001] HouseFinch Carotenoids 907

betweenages (F = 0.03,df = 1 and 96, P = 0.86) or subspecies(F = 1.68, df -- 1 and 96, P = I Mexico 0.20), and no significantvariation in slopes(F CAROTENOID []Californ,a I g-carotene = 0.52, df = 1 and 96, P = 0.47). lg-cryptoxanthin * 1 Componentcarotenoids and individualvariation Zeaxauthin I in plumagesaturation.--By visual assessment, Lutein I intensity (saturation)of breastpatch coloration 3'-dehydro-lutein I I varied modestly among males with 21 of 23 3'-hydroxyq•.œ--caroten-3-one . frontalisand 12 of 17 griscomimales assigned a œ.œ-caroten-3.3'-dmne• * color saturation of 6 or 7. Because few males Echinenone• had lower or higher intensity scores,we clas- 3-hydroxy-echinenone sified males into two groups with high-satu- 4-oxo-rubixanthin• ration males having scoresof sevenor higher Canth.... thinI and low-saturation males with six or lower. We ^.•xanthin had saturation scoresfor too few hatch-year o l; 2.0 3.0 d0 5.0 d0 *.0 8'0 do l•0 males in either subspeciesto allow us to ex- % of population plore differencesbetween age classes,so the following analysesare basedon data for both FIc. 4. Prevalenceof carotenoidsin plumagesof male House Finchesin the subspeciesgriscomi (N age classescombined. In each case,the same 43) andfrontalis(N = 58). Valuesshown are percent- trends were found within adult males alone. agesof malesdisplaying the respectivecarotenoid in Saturationof plumagecoloration is thoughtto their plumages.Data for adults and yearlingswere be a functionof the concentrationof pigmentsin pooled. An asterisk(*) indicatesa significantdiffer- plumage(Ryan et al. 1994),so we comparedthe encebetween the subspecies(P < 0.05, Fisher'sexact meanpigment concentration of malesin the two tests, see text). saturationcategories. High-saturation frontalis maleshad a significantlyhigher concentration of oids differed between the subspecies.Lutein, red (t = 2.29,P = 0.03,n = 10,13) but not yellow 3'-dehydro-lutein, and 3-hydroxy-echinenone (t = 0.64, P = 0.53) or total (t = 1.39, P = 0.18) appearedin plumagesof all malessampled in pigmentsthan low-saturationmales and a sig- both populations(Fig. 4). Zeaxanthin, E,E-car- nificantly higher proportion of pigmentsthat oten-3,3'-dione,and 3'-hydroxy-E,•-caroten-3- were red (t = 2.89, P = 0.008). However,high- one also occurredin all frontalismales (Fig. 4). saturationgriscomi males did not have a higher Proportion of males with [3-cryptoxanthin,3'- concentrationof red pigments (t = 1.09, P = hydroxy-•,•-caroten-3-one,•,E-caroten-3,3'-di- 0.30,n = 10, 7), but they did havea significantly one, 4-oxo-rubixanthin, adonirubin, and astax- lower concentrationof yellow pigments (t = anthin in their plumagesdiffered significantly 3.34, P = 0.005) and a significantlyhigher pro- betweenthe subspecies(Fisher's exact tests, all portion of red pigmentsin their plumage (t = P < 0.05 after Bonferronicorrection; Fig. 4). In 2.36, P = 0.03). As in frontalismales, high- and contrast,there were no significantdifferences low-saturationgriscomi males did not differ sig- in proportion of adult and hatch-year males nificantlyin total pigments(t = 0.06,P = 0.95). that had eachof the carotenoidsin their plum- Thus,color saturation in thebreast patch offron- age in eithersubspecies (Fisher's exact tests, all talis,but not griscomi,males was positively re- P > 0.05 after Bonferroni correction). lated to concentrationof red pigments.The sat- Mean (+SE) total carotenoid concentration uration of griscomibreast patch was negatively was more than twice as high in griscomi(885.4 relatedto concentrationof the yellow pigments. +--33.5 •g carotenoids/gfeather) than in fron- Despite the assumptionthat saturationreflects talis (405.5 + 17.7) males, and this difference is pigmentconcentration, there appears to be only highly significant(ANOVA, F = 185.5, df = 1 a weak link betweenvarious measures of pig- and 97, P < 0.0001;Fig. 1) with no significant ment concentrationand plumage saturation. age effect (F = 2.96, df = 1 and 97, P = 0.09). Subspeciescomparisons.--Although the same Using both subspeciesand age as factors,we 13 carotenoid pigments were found in the alsofound significant differences betweenfron- plumage of both frontalisand griscomimales, talis and griscomimales (ANOVA, all P < 0.05 the prevalenceof various componentcaroten~ after Bonferroni correction, no interaction 908 INOUYEET AL. [Auk,Vol. 118

Carotenoid stituting 69 + 2.3% (mean __+95% CI) of carot- B-caroten• enoids by mass in the plumage of frontalis B-cryptoxanthin I Mexico

. malesand 58 + 1.9%of carotenoidsby massin

Zeaxanthin the plumage of griscomimales. The difference betweensubspecies, but not betweenage clas- Lutein ses,was significant(ANOVA: subspecies,F = 3'-dehydro-lutein 53.2, df = 1 and 97, P < 0.0001;age, F = 1.30,

3'-hydroxy-œ,œ-caroten-3-one df = 1 and 97, P = 0.26). The concentrationsof thosetwo carotenoidswere significantlyneg- œ,œ-caroten-3,3'-dione atively related in griscomi(r = -0.58, P < Echinenone 0.0001,n = 43), but not infrontalis(r = 0.07, P = 0.59, n = 58). Four of the six most abundant carotenoids (>-4%of total carotenoids,by microgramper 4-oxo-rabixanthin• gramfeather) found in theplumages offrontalis Canthaxanthin• males were yellow (lutein, 3'-dehydro-lutein, Adonirabin• zeaxanthin,and 3'-hydroxy-e,e-caroten-3-one, Astaxanthin•l•,•,, i .... •.... s.... i ....i , ,, in decreasingabundance), whereas three of the 0 100 200 300 400 500 five mostabundant pigments in theplumages [Carotenoid](pg/g feather) of griscomimales were red (3-hydroxy-echine- FIG. 5. Mean (_+95%CL) concentrations(micro- none,adonirubin, and 4-oxo-rubixanthin;Fig. gramspigment per gram feather)of carotenoidpig- 5). ments extracted from the feathers of male House To further analyzedifferences in total carot- Finchesin the subspeciesgriscomi and frontalis.An enoidcomposition between subspecies and age asterisk(*) indicatesa significantdifference between classes,we performeda multivariateanalysis the two populations(P < 0.05, t-tests,see text). of variance (MANOVA) with all 13 carotenoid concentrationsas dependentvariables and both terms significant)in the mean concentrationsof age and subspeciesas factors.Only five of the carotenoid concentrations had normal distri- all carotenoidsexcept [3-carotene, zeaxanthin, lutein,and 3'-dehydro-lutein(Fig. 5). Levelsof butions(Shapiro-Wilks' tests, P > 0.05)in each [3-cryptoxanthin,3'hydroxy-E,E-caroten-3-one, age classwithin subspeciessample and the and E,e-caroten-3,3'-dionewere significantly others could not be normalized with transfor- higher infrontalismales, whereas levels of echi- mations.MANOVA is robustto somedeparture henone, 4-oxo-rubixanthin, canthaxanthin, from normality,so we ran that analysisusing adonirubin,astaxanthin, and most notably 3- data for all 13 carotenoids.MANOVA using hydroxy-echinenonewere significantlyhigher only the five carotenoidsthat had normal dis- in griscomimales. In frontalis,adult males had tributionsresulted in exactlythe sameconclu- significantly higher concentrationsof the red sionsreported below. pigmentadonirubin than did hatch-yearmales, The completeMANOVA (Pillai's trace,F = and significantly lower concentrationsof the 4.3, df = 39 and 261, P < 0.0001)revealed sig- pigments astaxanthin, lutein, and zeaxanthin nificant age (F = 2.4, df = 13 and 85, P = 0.009) (ANOVA, all P < 0.05 after Bonferroni correc- and subspecies(F = 21.1, df = 13 and 85, P < tion). That is an interestingresult given that 0.0001) effects, and the interaction term was not plumage of adult and hatch-yearmales from significant (F = 1.4, df = 13 and 85, P = 0.18). the frontalispopulation did not differ signifi- We then performedseparate MANOVA on each cantly in hue or saturation.Among griscomi subspeciesand age classesto further explore males, only the concentrationof adonirubin variationin plumage carotenoidcomposition. differed significantlybetween hatch-year and Those additional MANOVA revealed that adult adult males. and hatch-yearmales were significantlydiffer- The carotenoidspresent in the highestcon- ent in frontalis (F = 2.6, df = 13 and 29, P < centrationsin male House Finch plumages 0.0001),but not griscomi(F = 1.4, df = 13 and were the red carotenoid3-hydroxy-echinenone, 29, P = 0.21) and that frontalisand griscomi and the yellow carotenoidlutein, together con- males differed significantlyin the carotenoid October2001] HouseFinch Carotenoids 909

:.:../ '• griscomi, 0 froatalis•

. o o .o.

First Oanonioal Variate FKs.6. Canonicalvatlares plots showing variation between the subspeciesgriscomi and,frontalis within eachage class and variation between age classes HY (hatch-year)and AHY (adult)males within each subspecies.Canonical scores calculated from MANOVA using concentrations of all 13 componentcarotenoids; 95% confidenceellipses are shown.Axis unitsare not relevantin this studyand havebeen removed for clarity. compositionsof both adult (F = 10.4,df = 13 variatecomparing subspecies was significantly and 20, P < 0.0001)and hatch-yearmales (F = positivelycorrelated with most red pigments 13.5, df = 13 and 53, P < 0.0001). (sixin hatch-yearmales, four in adults)and sig- Plots of the canonical variate scores from the nificantlynegatively correlated with someyel- MANOVA (Fig. 6) clearlyshow that subspecies low pigments(three in hatch-yearmales, two are well separatedwithin age classeswith the in adults; all P < 0.05 after sequentialBonfer- discriminant function calculated from the ca- roni correction).Similarly, the first canonical rotenoidconcentrations correctly predicting all variatecomparing age classes was significantly adult males (n = 19 frontalisand 15 griscomi), correlatedwith both red (two positive in gris- all hatch-yearfrontalis (n = 39), and 89% of comi,two negativeinfrontalis) and yellow pig- hatch-yeargriscomi (n = 28).The first canonical ments(one negative in griscomi,three positive 910 INOUYEET AL. [Auk, Vol. 118 in frontalis). Thus, the first canonical variate enoid) and echinenone (a red carotenoid) as (Fig. 6) nicely illustratesage classesand sub- well as relative concentration of lutein--indi- speciesalong a continuumfrom yellow to red viduals with higher levels of lutein appeared pigments,though in reverseorder when com- more orange(Troy and Brush1983). Similarly, paring age classesinfrontalis males. the color polymorphism exhibited by the Sooty-cappedBush Tanager(Chlorospingus pi- DISCUSSION leatus)is due to differing concentrationsof lutein in the feathers (Johnsonand Brush 1972). Pigmentarybasis for individual variation in 3-Hydroxy-echinenonehas also been found to plumagecoloration.--For male House Finchesof be the primary carotenoidpigment responsible the subspeciesC. m.frontalis and C. m. griscomi, for red plumage color in severalother Carpo- the hue of ornamentalplumage dependedpri- dacusfinches, for example,C. roseusand C. rub- marily on proportion of red (4-keto-caroten- ricilloides(Stradi et al. 1995a, b; 1997), as well as oids) versus yellow pigments deposited in the plumagesof Pine Grosbeak(Pinicola eunu- feathers. Relationship between proportion of cleator;Stradi et al. 1996). red pigmentsand plumage rednesswas strong The samebasic relationship between redness and significantfor hatch-yearand adult fron- of integumentary display and proportion of talis males and for hatch-yeargriscomi males, red carotenoid pigments has also been ob- but therewas no relationshipbetween plumage served in fish. In the stickleback (Gasterosteus redness and proportion of red pigments in aculeatus),red males had primarily the red ca- adult griscomimales. It seemslikely that pro- rotenoid pigment astaxanthin in their skin portion of red pigments was not correlated whereas yellow males had primarily yellow with plumage rednessin adult griscomimales, pigments tunaxanthin and lutein in their skin because there was little detectable variation in (Wedekind et al. 1998). hue. That is, all adult males were bright red The color saturation (intensity) of ornamen- with a hue of 9 or 10. Lack of variation in plum- tal plumage appearedto dependprimarily on agehue amongadult griscomimales would also concentration of carotenoids in feathers, but re- explain weak relationshipsbetween hue and lationship between saturation and carotenoid virtually all componentcarotenoids (Table 2), concentrationwas different for frontalis and the weak negativerelationships between plum- griscomimales. Among frontalis males, total age rednessand concentrationof severalred concentrationof red pigmentswas the best pre- pigments,and the weak positiverelationships dictor of plumage color saturation,but among betweenplumage redness and concentrationof griscomimales, total concentrationof yellow severalyellow pigments(Table 2). pigmentswas the bestpredictor of plumagein- Contributionsof specificcarotenoids to var- tensity. Most griscomimales had abundantred iation in plumage hue were different for the pigments in their plumage, and variation in varioussubspecies and age classesof males.3- concentrationof red pigmentsapparently had Hydroxy-echinenonewas the most abundant little influenceon plumagecolor saturation. On red pigmentin the plumageof malesfrom both the other hand, abundanceof yellow pigments subspeciesand age classes,but its concentra- was variable,and that variationin yellow-pig- tion was not alwaysthe best predictor of plum- ment concentration appeared to determine age redness.Likewise, lutein was the most plumage intensity in griscomimales. Converse- abundant yellow pigment in the plumage of ly, most frontalis males had abundant yellow both subspeciesand age classesthat we sam- pigments but variable concentrationsof red pled, but its contributionto plumage redness pigments.In that subspecies,the concentration varied amonggroups. Total carotenoidconcen- of red pigmentsdetermined plumage color sat- tration in feathersalso had a weak but signifi- uration. Overall, the relationshipbetween color cant effect on plumage hue. saturationand pigment concentrationwas not Our observationsare similar to thosereport- as strong as expected.However, color satura- ed for Common(Carduelis fiammea) and Hoary tion was difficult to assessby the visualscoring (C. hornemanni)redpolls, in which hue differ- methods used in this study. Hence, error in encesamong individuals were attributedto the scoringsaturation may haveobscured patterns. concentrationsof both lutein (a yellow carot- Future studiesthat quantify plumage satura- October2001] HouseFinch Carotenoids 911 tion with a spectrophotometermay better re- and hatch-yearbirds in accessto dietary carot- solvethe issueof what pigment propertiesde- enoids, levels of parasite infection, immuno- termine plumage saturation. competence,or generalnutrition during molt. Ageeffects.--We found significant differences Implicationsfor potentialpathways of carotenoid in plumage color and ventral patch size be- metabolism.Birds are known to be capableof tween adult and hatch-year males in griscomi metabolically altering ingested carotenoids but not frontalis.The smaller patchesof color (Fox et al. 1969,Schiedt et al. 1985,Tyczkowski and drabber plumage of hatch-yeargriscomi and Hamilton 1986a, b, c, d; Brush 1990, males is not surprising,because males from Schiedt 1990). Much of the transformation and that subspecieshave been shown to havea dis- modificationof dietary carotenoidsoccurs by tinctive and female-like first-year plumage the introduction of oxo- or hydroxy-groups (Hill 1996b). In frontalis populations, hatch- into the main b-iononering (Schiedt1990), by year maleshave generallybeen observed to be the alterationof end rings, for example,a b- less colorful than adults (Michener and Mich- into an e-ring (Davies 1985), or both. Stradi et ener 1931, Gill and Lanyon 1965, Hill 1992, al. (1996)indicated that many carduelinefinch- 1993b),but hatch-yearand adult males exhibit es are capableof convertingcarotenoids by this the samerange of colors.What was surprising pathway,for example,converting dietary zea- was that, despitedelayed plumage maturation, xanthin into astaxanthinthat is depositedinto adult and hatch-yeargriscomi did not differ sig- feathers. nificantly in pigment composition of their Resultsfrom the presentstudy suggestthat feathers. Clearly, many hatch-year griscomi HouseFinches are also capableof the addition males had the same carotenoidcomposition in of a keto group at the C-4 position,C-4' posi- their feathersas adult males(Fig. 6), but others tion, or both, thusexplaining presence of red 4- had moreand higherconcentrations of yellow keto-carotenoidsin the feathers.Keto groupsat pigments.In contrast,despite their similarity the C-4(') position functionally extend the in appearance,adult and hatch-yearfrontalis chain of conjugateddouble bonds in the carot- finches had significantly different pigment enoid molecule,causing a bathochromicshift, compositions,suggesting that differentcarot- that is, shift of hue towards red (Hudon 1994b). enoid combinationsmay result in the same Thus, higherlevels of 4-keto-carotenoidsmay plumagecoloration. intensifyredness of feathers,accounting for the Differencesin the pigment compositionof observation that redder House Finches had the feathers of adult and hatch-yearfrontalis males may reflect differencesbetween those higher proportions of 4-keto-carotenoidsin age groups in physiologicalmechanisms in- their plumages. volved in feather pigmentation.Higher levels Many vertebrateshave the ability to convert of zeaxanthin and lutein and the lower levels of b- into e-endrings, which shortens the chainof adonirubin and astaxanthinin hatch-yearcom- conjugateddouble bonds and alterscarotenoid pared to adult .frontalismales suggestthat color to bright yellow (Matsuno et al. 1985, hatch-yearmales may have a greatertendency Miki et al. 1985). In birds, that may be done to to depositunmodified dietary carotenoidsdi- produce "canary xanthophylls"(Stradi et al. rectly into the feathers.Thus, ability to convert 1995a, b; 1997). House Finchesmay also be ca- dietaryprecursors into 4-keto-carotenoidsmay pableof thoseconversions, as demonstrated by increasewith age. Similar results have been presenceof the yellowcarotenoids, E,E-caroten- documentedfor femaleRed-winged Blackbirds 3,3'-dione and 3'-hydroxy-E,e-caroten-3-one,in (Agelaiusphoeniceus) in which epaulet color feathers.Those pigments may be produced changedfrom yellowin juvenilesto orangein from dietary sourcesof lutein or zeaxanthinbe- adults (Miskimen 1980).Such changes could be causeHouse Finch diets probably do not con- regulatedby sex hormones(Stoehr and Hill tain e,e-carotenoids(Inouye 1999).Many birds 2001), such that ability to convert dietary ca- havebeen shown to oxidizehydroxy- into keto- rotenoids into the redder 4-keto-carotenoids is groupsat the C-3(') position(Stradi et al. 1996), enhancedby onsetof sexualmaturity. The age- and that mechanismmay accountfor the oc- related differences in pigment composition currenceof 3'-dehydro-lutein in House Finch may alsobe due to differencesbetween adult feathers,converted from dietary lutein. 912 INOUYEET AL. [Auk, Vol. 118

Subspeciescomparisons.--We found substan- energy required for carotenoid utilization is tial differences in amounts and kinds of carot- great enough to constrainexpression of orna- enoidsin both hatch-yearand adult malesfrom mental plumage colorationis unknown (Hill the two subspecies.Those differences were 1996a,Inouye 1999).There is evidencethat nu- large and consistentenough that we were able tritionally stressedHouse Finches produce less to usecarotenoid concentrations to classifycor- red plumage than birds that are not stressed rectly to subspeciesall adult malesand 96% of evenwhen they have accessto the samecarot- hatch-yearmales on the basisof discriminant enoid pigments (Hill 2000). Furthermore,par- function analysis.Infrontalis males, the yellow asitespotentially play a large role in determin- 3'-hydroxy-e,e-caroten-3-oneand e,e-caroten- ing expressionof carotenoids;coccidia may 3,3'-dione were more likely to be found in inhibit intestinalabsorption of carotenoidsand plumage than the other componentcaroten- have been shownto decreaseplumage redness oids. Thoseyellow carotenoidsare most likely upon molt in male House Finches (Hill and derived from dietary sourcesof lutein or zea- Brawner 1998, Brawner et al. 2000). Other dis- xanthin. On the other hand, the 4-keto-carot- eases that affect the overall health of finches, in- enoids, echinenone, 4-oxo-rubixanthin, cantha- cluding pox and Mycoplasmagalliceptum, cause xanthin, adonirubin, and astaxanthin, in the males to grow a lessred plumage (Thompson plumage occurred in a greater percentage of et al. 1997, Nolan et al. 1998, Brawner et al. the population of griscomimales. With the ex- 2000). Carefully controlledstudies will be re- ception of astaxanthin,those carotenoidsare quired to determinethe relative contributionof probably not lutein- or zeaxanthin-derived. those various factors to the expressionof or- Thoseresults indicate (a) that there may be dif- namental plumage coloration in wild House Finches. ferencesbetween the two subspeciesin levels of dietary lutein and zeaxanthin, and (b) that The differencesin plumage colorationand griscomimales demonstrateincreased capaci- plumage pigment compositionthat we observed between age classes,between subspe- ties for adding keto- functionsat C-4('), where- cies,and amongmales within an age classand asfrontalis males show increasedcapacities for subspeciesmay be the result of any of the fac- converting p- into e-end rings. tors listed above acting alone or in combina- Carotenoid-basedcoloration as an honestsig- tion. The different mix of carotenoidpigments nal.--In this paper,we describethe pigmentary in the plumagesof hatch-yearversus adultfron- basis for variation in color expressionamong talis House Finches suggeststhat hatch-year male House Finches.The data, however,pro- malesmay utilize dietary carotenoidpigments vide no direct information about what causes differently than adult males,and suchage-spe- somemales to have more red pigmentsin their cific carotenoidutilization has potentially im- feathersthan other males. Diet may play a role portant implicationsfor signal content of or- in determining plumage colorationin males namental plumage coloration in this species (Hill 1992). Male House Finchesmust either in- (Hill 1990, 1994a, 2002). In addition, differenc- gest red pigments(which are rare in the diet; es in the carotenoidcomposition of feathersof Inouye 1999) or specificprecursors to the red malesfrom the two subspeciesthat we sampled pigmentsthat are ultimately depositedin the suggestthat those subspeciesutilize different feathers.However, the role of dietary accessto metabolicpathways in modifying dietary pig- carotenoid pigments in determining variation ments. Whether those differencesin pigment in expression of plumage coloration among physiologyevolved under sexualselection for wild birds remains controversial (Hill 1994c, brighter colordisplay (Hill 1994c)and what the 1999; Hudon 1994a, Thompsonet al. 1997, In- differencemeans for signal honestyin those ouye 1999). To convertprecursors in their diet two subspeciescan perhaps be answered to red pigments in their feathers,male House through a comparativestudy looking at pig- Finchesmust establish and maintain complex ment compositionof other subspeciesas well physiologicalsystems for carotenoid absorp- as the evolutionaryrelationships of thosetaxa. tion, transport, and deposition (Allen 1987, ACKNOWLEDGMENTS Erdman et al. 1993, Parker 1996, Furr and Clark 1997, Hill 2002). Those systemsrequire energy We extend specialthanks to the individuals and (Hill 1996a,2002), althoughwhether or not the families who provided logistical support and hos- October2001] HouseFinch Carotenoids 913 pitality during collectingtrips, particularly Fran Me- FURR, H. C., AND R. M. CLARK. 1997. Intestinal ab- waldt, Jorge Serrano, Sr., Jorge Serrano, Jr., the Ser- sorption and tissue distribution of carotenoids. rano family, and Mike Rigney and the staff at Coyote Nutritional Biochemistry8:364-377. Creek Riparian Station. Jim Dale assistedwith the GILL, D. E., AND W. E. LANYON. 1965. Establishment, field work in California. Birds were collected under growth, and behavior of an extralimital popu- a joint U.S. Fish and Wildlife and California Fish and lation of House Finches at Huntington, New Game ScientificCollecting Permit in California and York. Bird-Banding39:1-14. a scientificcollecting permit from La Secretariade GOODWIN,T. W. 1984. The Biochemistryof the Ca- RelacionesExteriores, Mexico. This study was ap- rotenoids, vol. II: Animals, 2nd ed. Chapman proved by the Chancellor'sAnimal ResearchCom- and Hall, New York. mittee at University of California Los Angeles and GOODWIN, T. W. 1986. Metabolism, nutrition, and the Animal Care Committee at Queen'sUniversity. function of carotenoids. Annual Review of Nu- Funding for this researchwas provided in part by the trition 6:273-297. American Museum of Natural History (Frank M. HENCKEN,H. 1992. Chemical and physiologicalbe- Chapman Memorial Fund), American Ornitholo- havior of feed carotenoids and their effects on gists'Union, Los AngelesAudubon Society, and Sig- pigmentation.Poultry Science71:711-717. ma Xi grants to C.I., and Natural Sciencesand En- HILL, G. E. 1990. Female House Finchesprefer col- gineering Research Council of Canada (NSERC) ourful males: Sexual selection for a condition- research and equipment grants to R.M.G.H. was dependenttrait. Animal Behaviour40:563-572. supportedduring field work by an NSERC Interna- HILL, G. E. 1991.Plumage coloration is a sexuallyse- tional PostdoctoralFellowship and during write-up lectedindicator of male quality.Nature 350:337- by NSF grant IBN9722172.Hoffman-La Roche,Basel, 339. Switzerland generously donated carotenoid stan- HILL, G. E. 1992. Proximate basis of variation in ca- dardsused in this study.Kevin McGraw,Jocelyn Hu- rotenoid pigmentation in male House Finches. don, David Chapman, Ken Nagy, and two anony- Auk 109:1-12. mous reviewers provided comments on the HILL, G. E. 1993a.Geographic variation in the carot- manuscript. enoid plumage pigmentation of male House Finches (Carpodacusmexicanus). Biological Jour- LITERATURE CITED nal of the Linnean Society49:63-86. HILL, G. E. 1993b.House Finch (Carpodacusmexican- ALLEN, P. C. 1987. Effect of Eimeria acervulina infec- us). In Birds of North America, no. 46. (A. Poole tion on chick (Gallusdomesticus) high densityli- and E Gill, Eds.). Academy of Natural Sciences, poprotein composition. Comparative Biochem- Philadelphia, and American Ornithologists' istry and PhysiologyB 87:313-319. Union, Washington,D.C. BRAWNER,W. R. III., G. E. HILL, AND C. A. SUNDER- HILL, G. E. 1994a. Geographicvariation in male or- MAN. 2000. The effects of coccidial and myco- namentationand female mate preferencein the plasmal infectionon carotenoid-basedplumage House Finch: A comparative test of models of pigmentation in male House Finches. Auk 117: sexual selection.Behavioral Ecology 5:64-73. 952-963. HILL, G. E. 1994b. Trait elaboration via adaptive BRUSH,A. H. 1990. Metabolism of carotenoid pig- mate choice: Sexual conflict in the evolution of ments in birds. FASEB Journal 4:2969-2977. signals of male quality. Ethology Ecology and BRUSH, A. H., AND D. M. POWER. 1976. House Finch Evolution 6:351-370. pigmentation: Carotenoid metabolism and the HILL, G. E. 1994c.House Finchesare what they eat: effects of diet. Auk 93:725-739. A reply to Hudon. Auk 111:221-225. DAVIES, B. H. 1976. Carotenoids. Pages 38-165 in HILL, G. E. 1996a.Redness as a measureof the pro- Chemistry and Biochemistryof Plant Pigments duction costs of ornamental traits. Ethology (T. W. Goodwin, Ed.). Academic Press, London. Ecologyand Evolution 8:157-175. DAVIES, B. H. 1985. Carotenoid metabolism in ani- HILL, G. E. 1996b. Subadult plumage in the House mals: A biochemist'sview. Pure and Applied Finch and tests of models for the evolution of de- Chemistry 57:679-684. layed plumage maturation. Auk 113:858-874. ERDMAN, J. W., JR., T. L. BIERER,AND E. T. GUGGER. HILL, G. E. 1998. An easy, inexpensive method to 1993. Absorption and transport of carotenoids. quantify plumage coloration. Journal of Field Annals of the New York Academy of Sciences Ornitholology69:353-363. 691:76-85. HILL, G. E. 1999. Mate choice,male quality, and ca- Fox, D. L., A. A. WOLFSON, AND J. W. MCBETH. 1969. rotenoid-basedplumage coloration.Pages 1654- Metabolism of •3-carotenein the American Fla- 1668 in Proceedings22nd International Ornitho- mingo, Phoenicopterustuber. Comparative Bio- logical Congress (N. Adams and R. Slotow, chemistryand Physiology29:1223-1229. Eds.).University of Natal, Durban, SouthAfrica. 914 INOUYEET AL. [Auk, Vol. 118

HILI•,G. E. 2000. Energetic constraints onexpression American Passerines. Slate Creek Press, Bolinas, of carotenoid-basedplumage coloration. Journal California. of Avian Biology 31:559-566. RICE,W. 1989. Analyzing tables of statisticaltests. HILL, G. E. 2002. A Red Bird in a Brown Bag: The Evolution 43:223-225. Function and Evolution of Colorful Plumage in RYAN, P. G., C. L. MALONEY, AND J. HUDON. 1994. the House Finch. Oxford University Press,New Color variation and hybridization among Neos- York. piza buntings on InaccessibleIsland, Tristan da HILL, G. E., AND W. R. BRAWNERIII. 1998. Melanin- Cunha. Auk 111:314-327. basedplumage colorationin the House Finchis SCHAEFFER,J. L., J. K. TYCZKOWSKI,C. P. PARKHURST, unaffectedby coccidialinfection. Proceedings of AND P. B. HAMILTON. 1987. Alterations in carot- the Royal Societyof London, SeriesB 265:1-5. enoid metabolism during ochratoxicosis in HILL, G. E., AND R. MONTGOMERIE.1994. Plumage young broiler chickens.Poultry Science66:318- color signalsnutritional conditionin the House 324. Finch. Proceedingsof the Royal Societyof Lon- SCHIEDT,K. 1990.New aspectsof carotenoidmetab- don, Series B 258:47-52. olism in animals.Pages 247-268 in Carotenoids: HUDON,J. 1994a.Showiness, carotenoids, and cap- Chemistry and Biology (N. I. Krinsky,M. M. Ma- tivity: A comment on Hill (1992). Auk 111:218- thews-Roth, and R. E Taylor, Eds.). Plenum 221. Press, New York. HUDON,J. 1994b.Biotechnological applications of re- SCHIEDT, K., E J. LEUENBERGER,M. VECCHI, AND E. search on animal pigmentation. Biotechnical GL1NZ.1985. Absorption,retention, and meta- Advances 12:49-69. bolic transformations of carotenoids in rainbow INOUYE,C. Y. 1999. The physiologicalbases for ca- trout, salmon, and chicken. Pure and Applied rotenoid color variation in the House Finch, Car- Chemistry 57:685-692. podacusmexicanus. Ph.D. dissertation,University STOEHR, A.M., AND G. E. HILL. 2001. The effects of of California, Los Angeles. elevated testosteroneon plumage hue in male JOHNSON,N. K., AND A. H. BRUSH.1972. Analysis of HouseFinches. Journal of Avian Biology32:153- polymorphism in the Sooty-cappedBush Tana- 158. ger. SystematicZoology 21:245-262. STRADI,R., G. 1998. The Colour of Flight: Caroten- KORNERUP,A., AND J. H. WANSCHER. 1983. Methuen oids in Bird Plumages. University of Milan Handbook of Colour, 3rd ed. Methuen, London. Press,Milan, Italy. MATSUNO, T., M. KATSUYAMA,T. MAOKA, T. HIRONO, STRADI, R., G. CELENTANO, M. BOLES, AND E MER- AND T. KOMORI.1985. Reductive metabolic path- CATO.1997. Carotenoidsin bird plumage:The ways of carotenoidsin fish (3S,3'S)-astaxanthin pattern in a seriesof red-pigmentedCarduelinae. to tunaxanthin A, B and C. Comparative Bio- Comparative Biochemistry and Physiology B chemistryand PhysiologyB 80:779-789. 117:85-91. MICHENER, H., AND J. R. MICHENER. 1931. Variation in color of male House Finches. Condor 33:12- STRADI,R., G. CELENTANO,AND D. NAVA. 1995a.Sep- aration and identification of carotenoids in 19. MIKI, W., K. YAMAGUCHI, S. KONOSU, R. TAKANE, M. bird's plumage by high-performance liquid SATAKE, T. FUJITA, H. KUWABARA, S. SHIMENO, chromatography-diode-arraydetection. Journal ANDM. TAKEDA.1985. Origin of tunaxanthinin of ChromatographyB 670:337-348. the integument of Yellowtail ($eriolaquinquera- STRADI, R., G. CELENTANO, E. Rossi, G. ROVATI, AND diata).Comparative Biochemistryand Physiolo- M. PASTORE.1995b. Carotenoidsin bird plum- gy B 80:195-201. age-I. The carotenoidpattern in a seriesof Pa- MISKIMEN,M. 1980. Red-winged Blackbirds:II. Pig- learcticCarduelinae. Comparative Biochemistry mentationin epauletsof females.Ohio Journalof and PhysiologyB 110:131-143. Science 80:236-239. STRAD[, R., E. ROSSI, G. CELENTANO, AND B. BELLAR- MOORE, R. T. 1939. A review of the House Finches of DL 1996. Carotenoidsin bird plumageMi. The the subgenusBurfica. Condor 41:177-205. carotenoidpattern in three Loxiaspecies and in NOLAN, P.M., G. E. HILL, AND A.M. STOEHR. 1998. Pinicola enucleator.Comparative Biochemistry Sex, size, and plumage rednesspredict House and PhysiologyB 113:427-432. Finch survival in an epidemic. Proceedingsof THOMPSON, C. W., N. HILLGARTH, M. LEU, AND H. E. the Royal Society of London, SeriesB 265:961- MCCLURE.1997. High parasite load in House 965. Finches(Carpodacus mexicanus) is correlatedwith PARKER,R. S. 1996. Absorption, metabolism,and reduced expressionof a sexuallyselected trait. transport of carotenoids.FASEB Journal 10:542- American Naturalist 149:270-294. 551. TROY, D., AND A. H. BRUSH.1983. Pigments and PYLE, 12.,S. N. G. HOWELL, R. P. YUNICK, AND D. E feather structure of the redpolls, Carduelisfiam- DESANTE. 1987. Identification Guide to North mea and C. hornemanni. Condor 85:443-446. October2001] HouseFinch Carotenoids 915

TYCZKOWSKI,J. K., AND P. B. HAMILTON. 1986a. Evi- TYCZKOWSKI,J. K., AND P. B. HAMILTON. 1986d. Ab- dencefor differential absorptionof zeacarotene, sorption,transport, and depositionin chickens cryptoxanthin, and lutein in young broiler of lutein diester, a carotenoid extracted from chickens.Poultry Science65:1137-1140. marigold (Taseteserecta) petals. Poultry Science TYCZKOWSKI,J. K., AND P. B. HAMILTON. 1986b. Lu- 65:1526-1531. tein as a model dihydroxycarotenoidfor the WEDEKIND, C., P. MEYER, M. FRISCHKNECHT,U. A. study of pigmentationin chickens.Poultry Sci- NIGGLI, AND H. PFANDER.1998. Different carot- ence 65:1141-1145. enoidsand potentialinformation content of red TYCZKOWSKI,J. K., AND P. B. HAMILTON. 1986c. Can- coloration of male three-spined stickleback. thaxanthinas a modelfor the studyof utilization Journalof Chemical Ecology24:787-801. of oxycarotenoidsby chickens.Poultry Science 65:1350-1356. AssociateEditor: C. Bosque