The effects of preen oils and soiling on the UV visible reflectance of carotenoid-pigmented feathers

Transcription

1 ehv Ecol Sociobiol DOI 1.17/s y ORIGINL PPER The effects of preen oils nd soiling on the UV visible reflectnce of crotenoid-pigmented fethers Lorenzo Pérez-Rodríguez & Frncois Mougeot & Gry R. ortolotti Received: 7 October 21 / Revised: 14 December 21 / ccepted: 26 Jnury 211 # Springer-Verlg 211 bstrct Plumge colortion, prticulrly when crotenoidbsed, is importnt in socil signling in birds. lthough fether color is reltively stble trit, individuls my modify it with cosmetic substnces such s preen oils. In ddition, dirt ccumultion my influence plumge colortion nd further ffect signl perception by receivers. Here, we nlyze the seprte potentil effects of preen oils nd soil ccumultion on the reflectnce properties of crotenoidpigmented fethers cross the visul rnge of most bird species, which includes the ultrviolet (UV). Using the yellow Communicted by E. Fernndez-Juricic Electronic supplementry mteril The online version of this rticle (doi:1.17/s y) contins supplementry mteril, which is vilble to uthorized users. L. Pérez-Rodríguez (*) Deprtmento de Ecologí Evolutiv, Museo Ncionl de Ciencis Nturles (CSIC), José Gutiérrez bscl 2, 286 Mdrid, Spin e-mil: lorenzo.perez@mncn.csic.es L. Pérez-Rodríguez e-mil: lorenzoperezrodriguez@gmil.com F. Mougeot Estción Experimentl de Zons Árids (CSIC), Crreter de Scrmento s/n, L Cñd de Sn Urbno, 412 lmerí, Spin L. Pérez-Rodríguez : F. Mougeot Instituto de Investigción en Recursos Cinegéticos, IREC (CSIC, UCLM, JCCM), Rond de Toledo s/n, 135 Ciudd Rel, Spin G. R. ortolotti Deprtment of iology, University of Ssktchewn, Ssktoon, Cnd S7N 5E2 portion of til fethers of ohemin wxwings (ombycill grrulus), we performed two seprte experiments where: () preen oils nd/or soil were removed, or (b) preen oils (from blck-billed mgpies Pic pic or egle owls ubo bubo) were dded. Preen oil ddition reduced brightness but incresed UV hue nd yellow chrom. UV chrom ws reduced by the ddition of mgpie (but not owl) preen oil. Soil ccumultion hd little effect on plumge reflectnce in the UV rnge but significntly reduced yellow chrom. ccording to models of vin vision, both of these effects re detectble by birds nd biologiclly meningful when compred with nturl vrition between the sexes nd ge clsses. We conclude tht preen oil nd soil ccumultion cn significntly ffect the UV visible reflectnce of crotenoid-bsed plumges. s such trits typiclly dvertise individul qulity, preening nd soiling hve the potentil to modify the informtion content of crotenoid-bsed plumge trits nd how these signls re perceived by receivers. Keywords Crotenoids. Colortion. Honest signls. Preen wxes. Uropygil glnd. Sexul selection Introduction Plumge colortion is mjor trget for sexul selection in birds (Hill nd McGrw 26). ecuse it often relibly indictes individul qulity, colortion cn ffect mting success or dominnce, thereby influencing individul fitness (Hill nd McGrw 26). Fether colortion is determined by differentil wvelength reflection cused by fether microstructure, or by the selective wvelength bsorption of pigments (e.g., melnins, crotenoids, porphyrins, etc.) (Hill nd McGrw 26b). lthough plumge color my chnge slightly with time (e.g., Örnborg et l.

2 ehv Ecol Sociobiol 22; McGrw nd Hill 24; Figuerol nd Senr 25), the color of fethers is much more stble thn tht of other ornments like beks, legs, wttles, eye rings, or skin (e.g., Fivre et l. 23; Mrtínez-Pdill et l. 27; Pérez- Rodríguez 28). However, there is incresing evidence tht birds cn modify fether reflectnce by the ppliction of some endogenous or exogenous gents such s oils, wxes, skin secretions, fether powders, nd soil (Montgomerie 26; Delhey et l. 27). The functions of these cosmetic modifictions of plumge colortion re still poorly known, lthough some uthors hve highlighted their signling potentil, prticulrly in sexul selection (Negro et l. 1999; Delhey et l. 27). Preen oils re produced by the uropygil glnd of most bird species nd consist of mixture of lipids tht birds ctively pply to fethers (Jcob nd Ziswiler 1982). The primry role of uropygil glnd secretions is to mintin the flexibility nd wterproofing of fethers (Jcob nd Ziswiler 1982), but recent studies hve proposed other possible functions: preen oils could protect fethers from degrding bcteri (urtt nd Ichid 1999; Shwkey et l. 23) or prevent fether color degrdtion due to exposure to sunlight (Reneerkens nd Korsten 24; Surmcki28). In ddition, preen oils my modify plumge colortion by cting s opticl filters (Piersm et l. 1999). Thus, birds could rpidly, nd perhps lso reversibly, djust plumge reflectnce by pplying vrible mounts or types of preen oils to fethers (Reneerkens nd Korsten 24). Interestingly, preen oils mostly bsorb light in the ultrviolet (UV) rnge (3 4 nm) of the spectrum (Reneerkens nd Korsten 24; Delhey et l. 28). Mny birds re sensitive to UV light (Cuthill 26; see below) nd respond to vrition in plumge UV reflectnce (Husmnn et l. 23), so preen oil ppliction could influence how UV-reflective plumges re perceived. Plumge colortion cn lso be ltered by the ccumultion of soil nd other exogenous substnces (Montgomerie 26). In some cses, birds delibertely pply these substnces for cmouflge or sttus signling (e.g., Montgomerie et l. 21; Negro et l. 1999). However, in most cses, the ccumultion of dirt in fethers is ccidentl nd voided by birds, which try to keep their plumge clen by the mens of preening nd bthing (Zmpig et l. 24; Wlther nd Clyton 25; Lenouvel et l. 29). Dirt ccumultion ffects the insulting nd wterproofing properties of the plumge, s well s their reflectnce, which is often used s n indictor of individul qulity (e.g., Zmpig et l. 24). lthough some recent studies hve tested the effect of preen oils on fether color, results hve been mixed (Reneerkens nd Korsten 24; Surmcki nd Nowkowski 27; Delhey et l. 28; López-Rull et l. 21). Furthermore, besides the importnce of including the UV rnge (3 4 nm) when studying vin colortion, this hs been done by only two studies (Reneerkens nd Korsten 24; Delhey et l. 28), none of which focused on crotenoid-bsed color. Crotenoid-pigmented plumges re prticulrly interesting becuse they re mong the commonest signls of individul qulity in birds (Hill nd McGrw 26) nd my relibly indicte forging bility, prsite levels, nd overll helth sttus (reviewed by McGrw 26). Moreover, crotenoid-pigmented fethers typiclly show bimodl pttern of reflectnce, with mrked pek in the UV nd plteu in the yellow red visible spectrum (>5 nm) (e.g., leiweiss 25; Shwkey nd Hill 25; ndersson nd Prger 26; see lso Fig. 1). Therefore, preen oils nd soil hve the potentil to influence reflectnce in the UV nd visible prt of the spectrum nd how these signls re perceived. Species differ in color perception, prticulrly in the UV prt of the visul spectrum. In fct, birds cn be split into two bsic groups ccording to their visul system: species tht re ultrviolet nd violet sensitive, herefter UVS nd VS (Cuthill 26). This is minly bsed on the specific sensitivity of the cones responding to shorter wvelengths (UVS species possess higher sensitivity t shorter wvelengths thn VS species), lthough there my be other differences in the rest of the visul spectrum. For instnce, retinl oil droplets my modify color perception by birds due to their filtering properties (Vorobyev 23; Hrt nd Hunt 27). Using two seprte experiments, we tested the effects of preen oils nd soiling on the UV visible reflectnce of the yellow bnd of til fethers of ohemin wxwings (ombycill grrulus). We selected this model species nd trit becuse the yellow til bnd is pigmented by crotenoids (primrily cnry xnthophylls), s demonstrted by the closely relted cedr wxwing ombycill cedrorum (Hudon nd rush 1989). This is conspicuous color trit lrgely influenced by diet (Hudon nd rush 1989) tht reflects both in the UV nd visible prts of the spectrum (Fig. 1). ohemin wxwings lso possess yellow tips on primry fethers nd red wxy tips on secondry fethers, lso pigmented by crotenoids. In this study, we first test the effect of removl of preen oil nd soil on the UV visible reflectnce of this trit. lso, in second experiment, we test the effect of preen oil ddition on the sme trit. lthough other closely relted oscine species possess UVS cones (Ödeen nd Håstd 23), the visul rnge of wxwings is unknown. Therefore, we nlyzed whether our experimentl mnipultions resulted in chnges detectble by both UVS nd VS species, in order to discuss the potentil influence of these substnces on the signl content of crotenoid-bsed plumge.

3 ehv Ecol Sociobiol reflectnce (%) fter experimentl tretment reflectnce difference with controls Methods b -1 soil removed oil nd soil removed Fether collection wvelength (nm) The birds used in this study were ll crcsses slvged fter hitting windows on the cmpus of the University of Ssktchewn, Ssktoon, Cnd. ohemin wxwings re regulr wintering birds in the city nd cn be commonly found in flocks of hundreds to bout 2, (Mountjoy 25 nd personl observtion). Multiple birds from single flock often died together. On three occsions, five or six birds were seen striking window t one time fter it ppered tht the flock hd been disturbed by predtor. Experiment 1: soil nd preen oil removl control soil removed oil nd soil removed Fig. 1 Effects of experimentl tretment in experiment 1 (see Methods) on fether reflectnce. Represents the reflectnce spectr of the yellow tip of til fethers of ohemin wxwings (see photo insert) fter soil removl (gry solid line), oil nd soil removl (thick blck solid line), nd control tretment (thin blck solid line). b Shows the difference in reflectnce between fethers fter soil (gry dshed line) or oil nd soil removl (blck dshed line) nd controls. Verticl brs denote SE t 5-nm intervls For this experiment, three djcent til fethers from ech bird (N=4) were collected. We excluded the centrl nd outer fethers to keep the smple from ech bird homogeneous in terms of size, preen oil, nd dirt content. Fethers from ech individul were ssigned to three experimentl tretments: control, soil-removed, or preen oil- nd soilremoved. ll fethers were first dried in n oven t 4 C for 1 h. Fethers were then weighed (to the nerest.1 g) using n electronic blnce (Gibertini E-5-S, Miln, Itly), nd the reflectnce spectrum of the yellow tip (Fig. 1) ws mesured (see below) before the experimentl mnipultion. Preen oil nd soil removl ws performed using similr protocol to tht described by Surmcki nd Nowkowski (27). Oil/soil-removed fethers were put in flsk contining 5 ml of chloroform/methnol mixture (2:1), shken, left for c. 2 min, nd shken gin before removing them from the flsk nd drying them in n oven t constnt temperture (4 C). Soil-removed fethers were treted in the sme wy, but in this cse, the flsk contined 5 ml of distilled wter. This tretment did not remove ll the soil ttched to the fether, s some of the dirt ws mixed with preen oil (wter is polr solvent nd ws unble to remove dirt mixed with preen oil). Control fethers were shken in n empty flsk nd plced in the sme oven fterwrds. Once they were completely dry fter 5 h, fethers were weighed gin nd their reflectnce spectr mesured. These two mesures of fether weight would llow us to confirm the effectiveness of our experimentl mnipultion, s fether weight chnges were expected to follow the following pttern: oil/soil-removed fethers > soil-removed fethers > control fethers. During these procedures, fethers were crefully mnipulted using gloves nd forceps, nd plced in hermetic plstic bgs in between mesurements. Experiment 2: preen oil ddition For this experiment, 23 pirs of djcent til fethers were collected (s in experiment 1, outer nd centrl fethers were excluded) nd their preen oils nd soil removed following the sme procedure (i.e., chloroform/methnol wshing) described bove. fter this, fethers were weighed nd their color mesured. We were not ble to collect preen oils from wxwings, so we smpled uropygil glnd secretions from two other species: blck-billed mgpies (Pic pic) nd egle owls (ubo bubo) (three individuls from ech species were smpled). Mgpies were obtined from nother study tht required their euthniztion nd were kept frozen for 6 months. Mgpie smples were obtined fter defrosting the birds nd gently pressing their uropygil glnds the sme dy of the experiment. Egle owl smples were collected from three live birds kept in locl rehbilittion center (Centro de Recuperción de Fun Silvestre El Chprrillo, Ciudd Rel, Spin). These birds hd stbilized bone frctures in the wings nd could not be relesed to the wild but were in good condition s indicted by regulr checks by veterinry stff of the center. Owl smples were collected in similr wy, nd smples were kept frozen for 1 week until the

4 ehv Ecol Sociobiol experiment ws performed. We ssumed tht this short period of freezing does not produce substntil chnge in the opticl properties of preen oils, s supported by recent evidence (Delhey et l. 28). Pirs of fethers from individul wxwings were ssigned to two different groups depending on the origin (mgpie or owl) of the preen oil employed. In the mgpie group, ech pir of fethers (N=13) ws split into two groups: mgpie preen oil-dded ( mgpie oil ) or control ( mgpie control ), i.e. no preen oil ddition. Similrly, pirs of fethers ssigned to the owl group (N=1) were split into owl preen oil-dded ( owl oil ) or control ( owl control ). Preen oils were defrosted t mbient temperture before ppliction. For preen oil dditions, smll drop of oil ws pplied t the bse of the yellow til bnd nd crefully spred over the whole yellow distl prt to be mesured using piece of plstic film. Control fethers were mnipulted in the sme wy, but no substnce ws dded. fter these procedures, fethers were weighed gin nd their reflectnce mesured. These mesures of fether weight before nd fter the ddition of preen oil would llow us to verify tht the dditions were within the nturl rnge (i.e., chnges in fether weight in experiment 2 [oil ddition] were similr to those of experiment 1 [oil removl]). Color mesurements The reflectnce of the yellow distl prt of the til fethers ws mesured using n vspec US2 spectrophotometer connected to deuterium hlogen light source (DH2, vntes, Eerbek, Netherlnds) through bifurcted fiberoptic probe providing 45 to norml ngle of illumintion/ recording. Three consecutive mesurements were conducted per fether, removing the probe from the fether between mesurements. Mesurements were done in prtilly drk room to void possible interference from mbient light. Reflectnce vlues (in 3-nm steps, from 3 to 7 nm) were clculted reltive to Spectrlon 99% white stndrd reference (Lbsphere, Congleton) nd to the drk, nd computed using the progrm Spectrwin 5.. The white reference ws checked every six fethers to ensure the stbility of the light source. Following Montgomerie (26b) nd ndersson nd Prger (26), we summrized spectrl dt by clculting the following vribles: (1) men brightness (verge reflectnce in the intervl 3 7 nm), (2) UV hue (spectrl loction, in nnometers, of the reflectnce pek in the UV rnge), (3) UV chrom (difference between mximum nd minimum reflectnce in the intervl 3 4 nm, divided by brightness); (4) yellow hue (wvelength of mximum slope in the 4 7 intervl), nd (5) yellow chrom (difference between mximum nd minimum reflectnce in the intervl 4 7 nm, divided by brightness). Correltions between these vribles re shown in Online Resource 1. Significnt effects of our tretments on the color vribles do not necessrily imply tht these effects could be perceived by birds. Therefore, we lso quntified whether spectrl differences could be detectble by birds by using color discrimintion model (Vorobyev nd Osorio 1998; Vorobyev et l. 1998). We used visul prmeters of UVS (Hrt et l. 2) nd VS birds (Hrt 22) to compute the chromtic (ΔS) nd chromtic (ΔQ) contrsts in the vin visul spce s result of experimentl mnipultions. The unit for ΔS nd ΔQ is the just noticeble difference (JND); vlues of ΔS or ΔQ higher thn 1 indicte color differences noticeble by birds. ΔS nd ΔQ clcultions were done using the progrm vicol (Gomez 26). Further mthemticl detils of these models cn be found in Delhey et l. (28), Lenouvel et l. (29), nd vilés nd Soler (29). Sttistics We tested the effect of tretments on color vribles using generl liner mixed models. In experiment 1, smpling time ( before or fter experimentl mnipultion) nd tretment (control, soil-removed, or oil/soil-removed) were entered s fixed effects nd individul fether s rndom vrible to ccount for repeted mesures (before nd fter tretment) on the sme fether. In experiment 2, the sttisticl models included time, tretment (preen oil-dded or control), nd oil type (owl or mgpie) s fixed fctors. To ssess whether the effect of preen oils differed depending on preen oil type t different times, the three-wy interction of time tretment oil type nd ll two-wy interctions between these effects were lso included in the model. To ssess whether chnges in fether reflectnce in both experiments would be noticeble by birds, we computed ΔS nd ΔQ (for UV nd UVS eye types) between control nd experimentl fethers before nd fter mnipultions in both experiments. We tested whether ech contrst between control nd treted fethers significntly incresed s result of the tretments using pired t tests. Men brightness, ΔS, nd ΔQ vlues were log-trnsformed to chieve normlity (Kolmogorov Smirnov test), lthough untrnsformed vlues were used in tbles nd grphs. ll tests re two tiled, nd ll dt re given s men ± stndrd error (SE). Results Experiment 1: soil nd preen oil removl Experimentl tretments cused significnt differences in the percentge of weight lost by fethers (time tretment

5 ehv Ecol Sociobiol interction: F 2,117 =48.5, P<.1). Oil/soil-removed fethers lost more weight thn soil-removed fethers nd soilremoved fethers lost more weight thn control fethers, nd ll differences between groups were significnt (ll P<.1; Fig. 2). The effect of experimentl tretment on fether color is shown in Tbles 1 nd 2 nd Figs. 1 nd 2. Men brightness ws not ffected by tretment. In contrst, UV hue nd UV chrom showed significnt vrition depending on the tretment (Fig. 2c, d). lthough oil/soil-removed fethers did not show significnt chnge in UV hue (F 1,39 =2.42, P=.13), vritions exhibited were significntly different from those of both control fethers (time tretment interction: F 1,78 =5., P=.3) nd soil-removed fethers (time tretment interction: F 1,78 =11.1, P=.1). In soil removed fethers, UV hue incresed over time (F 1,39 =11.3, P=.2). In contrst, UV hue did not increse significntly over time in control fethers (F 1,39 =2.53, P=.12). Overll, the increse in UV hue observed in soil-removed fethers ws not sttisticlly different from tht observed in control fethers (smpling time tretment interction: F 1,78 =.7, P=.41). UV chrom did not chnge significntly in control nd soil-removed fethers (both P>.15), the chnges being similr mong tretment groups (time tretment interction: F 1,78 =1.6, P=.31). However, oil/soil-removed fethers showed significnt increse in UV chrom (F 1,39 =37.3, P<.1), different from tht of control (F 1,78 =1.6, P=.2) nd soil-removed fethers (F 1,78 =16.1, P<.1) (Fig. 2d). Yellow hue incresed similrly in ll groups (no significnt time tretment interction, Tble 1 nd Fig. 2e). Chnges in yellow chrom differed, lthough not significntly, between tretments (Tble 1, Fig. 2f). Control nd oil/soil-removed fethers showed no significnt chnge in yellow chrom (F 1,39 =.3, P=.87 nd F 1,39 =2.3, P=.16, respectively). In contrst, soil-removed fethers showed significnt increse in yellow chrom (F 1,39 =12.7, P<.1). Hence, Fig. 2 Chnges, in men ± SE, ccording to tretment within experiment 1, in fether weight (in percent), b men brightness (in percent), c UV hue (in nnometers), d UV chrom (in percent), e yellow hue (in nnometers), nd f yellow chrom (in percent). lck, gry, nd open squres correspond to control, soilremoved, nd soil- nd preen oil-removed fethers, respectively. Different letters bove symbols indicte significnt (P<.5) differences between tretment groups. Cpitl letters indicte chnge significntly different from zero; uncpitlized letters indicte chnge not significntly different from zero (non-significnt chnge) chnge in fether weight (%) b chnge in men brightness (%) C d chnge in UV chrom (%) e chnge in yellow hue (nm) c chnge in UV hue (nm) b f chnge in yellow chrom (%) b. control soil removed oil/soil removed control soil removed oil/soil removed tretment tretment

6 ehv Ecol Sociobiol Tble 1 Effect of tretments (experiment 1; see Methods) on fether brightness, UV hue, UV chrom, yellow hue, nd yellow chrom of the crotenoid-pigmented portion of wxwing til fethers rightness UV hue UV chrom Yellow hue Yellow chrom F df P F df P F df P F df P F df P Time (T) , , , 117 < , 117 < , Tretment (TR).77 2, , , , , 117 <.1 T TR.48 2, , , 117 < , , Time (before or fter) nd tretment (control, soil-removed, nd soil/oil-removed) were included s fixed effects. Individul fether ws entered s rndom vrible chnges in yellow chrom differed significntly between control nd soil-removed fethers (F 1,78 =5.62, P=.2), but not between control nd oil/soil-removed fethers (F 1,78 =.98, P=.32) or between soil-removed nd oil/soil-removed fethers (F 1,78 =1.28, P=.26). ll these effects on fether reflectnce prmeters resulted in chnges in chromtic distnces (ΔS) noticeble by UVS or UV visul systems (Tble 2), lthough chnges were closer to the discrimintion threshold in the lter. In contrst, no detectble chnges in the chromtic distnce (ΔQ) were found between tretments, regrdless of the visul system (UVS or VS; Tble 2). Experiment 2: preen oil ddition Preen oil ddition cused n increse in fether weight in oiled fethers (time tretment: F 1,42 =53., P<.1) which ws similr in mgpie-oil nd owl-oil fethers (time tretment type of preen oil interction: F 1, 42 =.7, P=.79) (Fig. 3). The mgnitude of this increse (2.±.35%) ws similr to the net decrese in fether weight of oil/soil removl (i.e., fter discounting the chnge of control nd soil-removed fethers (1.6±.9%), nd the differences between the two were non-significnt (F 1,61 =2.6, P=.15). This indicted tht the mount of preen oil dded ws within the nturl rnge found in wxwing til fethers. The effects of experimentl preen oil dditions on fether color vribles re shown in Tble 3 nd Figs. 3 nd 4. Men brightness decresed in control fethers, but this ws significnt only in owl group (mgpie control: F 1, 12=1.99, P=.18; owl control: F 1, 9 =7.62, P=.2), lthough this decrese ws much more pronounced s result of preen oil ddition (significnt time tretment interction), irrespective of the type of oil (Tble 3, Fig. 4b). However, these chnges in fether brightness my not be noticeble by birds, s shown by the effects on chromtic contrst using UVS (mgpie oil: t 12 =1.85, P=.9; owl oil: t 9 =.1, P=.99) or VS visul systems (mgpie oil: t 12 =1.89, P=.8; owl oil: t 9 =.16, P=.87) (Tble 2). UV hue incresed fter preen oil ddition (mgpie-oil fethers: F 1,12 =11.1, P=.6; owl-oil fethers: F 1, 9 =81., P<.1), the effect being similr for both types of preen oil (non-significnt time tretment oil type interction; Tble 3, Figs. 3 nd 4c). In contrst, control fethers showed no significnt chnges in UV hue during the experiment (mgpie control: F 1,12 =3.1, P=.1; owl control: F 1,9 =198, P=.19). Tble 2 bsolute chnge in chromtic (ΔS) nd chromtic (ΔQ) distnces (in JNDs) between control nd experimentl fethers during experiment 1 (soil/preen oil removl) nd experiment 2 (mgpie or egle owl preen oil ddition) UVS eyes VS eyes ΔS ΔQ ΔS ΔQ Experiment 1 Control vs. soil-removed 2.1±.94*.5± ±.83*.18±.54 Control vs. oil/soil-removed 2.4±1.1*.4± ±.94**.2±.53 Soil removed vs. oil/soil-removed 1.5±.44*.2± ±.4**.22±.4 Experiment 2 Control vs. mgpie oil ddition 3.4±1.1* 1.4±.76*** 4.1±1.24* 1.5±.8*** Control vs. owl oil ddition 3.1±1.7**.59± ±1.5**.48±.87 Vlues correspond to post-tretment contrst minus pre-tretment contrsts between the specified groups, clculted for UVS nd VS eyes. For experiment 1, the chnge in ΔS nd ΔQ between soil-removed nd oil/soil-removed fethers is lso shown. Levels of significnce of the chnges fter pired t tests (pre-tretment contrsts vs. post-tretment contrsts between the specified groups) re noted by sterisks (*P<.1; **P<.5; ***P<.1). See Methods for further detils

7 ehv Ecol Sociobiol Fig. 3 Effects of blck-billed mgpie nd b egle owl preen oil ddition on wxwing fether reflectnce. Thin lines represent intct (control) fethers, thick lines represent fethers with preen oil pplied. Dshed lines in the lower grphs represent the difference in reflectnce between control nd fethers with mgpie or owl preen oil. Verticl brs denote SE t 5-nm intervls reflectnce (%) reflectnce difference blck-billed mgpie preen oil wvelength (nm) control preen oil dded reflectnce (%) reflectnce difference b egle owl preen oil control preen oil dded wvelength (nm) UV chrom ws lso ffected by the experimentl tretment (Fig. 4d); however, in this cse, the effect depended on the type of preen oil pplied (significnt time tretment type of preen oil interction, Tble 3). Mgpie control nd mgpieoil fethers showed decrese in UV chrom (F 1, 12 =5.84, P=.3 nd F 1, 12 =2.8, P<.1, respectively), which ws significntly more pronounced in the mgpie control thn in mgpie-oil fethers (time tretment interction: F 1, 24 =8.36, P=.8) (Fig. 3). In contrst, owl control nd owloil fethers did not show ny chnge in UV chrom (time: F 1, 18 =.78, P=.39; time tretment interction: F 1, 18 =.39, P=.54) (Fig. 3b). Yellow hue decresed during the experiment, similrly so in the different tretment groups (Tble 3, Fig. 4e). Finlly, yellow chrom incresed in ll groups, lthough this increse ws significntly greter in fethers where preen oils were dded, irrespective of the type of oil employed (Tble 3, Fig. 4f). ll these chnges in fether reflectnce resulted in significnt increse in chromtic distnce between control nd experimentl fethers for species possessing UVS (mgpie oil: t 12 =3.68, P=.3; owl oil: t 9 =2.73, P=.2) or VS visul systems (mgpie oil: t 12 =3.9, P=.2; owl oil: t 9 =2.44, P=.3) (Tble 2). Discussion In this study, we tested the effect of preen oils nd soil on fether reflectnce. This is the first experimentl ttempt to tese prt the effects of these two fctors on fether reflectnce. lso, this is the first study nlyzing the effects of preen oils on the reflectnce of crotenoid-bsed plumge cross the entire visul rnge of most birds (3 7 nm) nd tking into ccount the color discriminbility of the vin visul system. Preen oil ddition incresed UV hue (fether reflectnce peked t longer wvelengths in the 3 4 nm intervl), nd in the cse of mgpie s uropygil secretions, it lso reduced UV chrom. This is consistent with results of Tble 3 Effect of tretments (experiment 2; see Methods) on fether brightness, UV hue, UV chrom, yellow hue, nd yellow chrom of the crotenoid-pigmented portion of wxwing til fethers rightness UV hue UV chrom Yellow hue Yellow chrom F df P F df P F df P F df P F df P Time (T) , 42 < , 42 < , 42 < , 42 < , 42 <.1 Oil type (OT) , , , , , Tretment (TR) , , , , , T OT , , , , , 42.9 T TR , 42 < , , , , 42.5 TR OT.39 1, , , , , T TR OT.5 1, , , , , Time (before or fter) nd tretment (control or preen oil-dded) nd the type of preen oil (mgpie or owl) were included s fixed effects. Individul fether ws entered s rndom vrible

8 ehv Ecol Sociobiol Fig. 4 Chnges, in men ± SE, ccording to tretment within experiment 2, in fether weight (in percent), b men brightness (in percent), c UV hue (in nnometers), d UV chrom (in percent), e yellow hue (in nnometers), nd f yellow chrom (in percent) ccording to tretment within experiment 2. Circles nd tringles correspond to the mgpie nd owl preen oil group, respectively, wheres open nd blck symbols correspond to control nd oiled fethers, respectively. Different letters bove symbols indicte significnt (P<.5) differences between tretment groups. Cpitl letters indicte chnge significntly different from zero; uncpitlized letters indicte chnge not significntly different from zero (nonsignificnt chnge) chnge in fether weight (%) b chnge in men brightness (%) d chnge in UV chrom (%) e chnge in yellow hue (nm) c 4 f 25 chnge in UV hue (nm) chnge in yellow chrom (%) control oil control oil control oil control oil mgpie owl mgpie owl tretment tretment Delhey et l. (28), showing decrese in UV chrom nd overll brightness when preen oils were pplied to white fethers of mllrds (ns pltyrynchos) or UV-reflective fethers of blue tits (Prus ceruleus). In contrst, Reneerkens nd Korsten (24) did not find ny effect of preen oil removl on the reflectnce of fethers of red knots (Clidris cnutus), possibly becuse these fethers were pigmented by melnins nd reflected little in the UV rnge. The reduction of UV reflectnce by preen oils is likely explined by the reltively higher bsorbnce of preen oils t short wvelengths (Reneerkens nd Korsten 24; Delhey et l. 28). Preen oil ddition lso incresed yellow chrom but did not ffect yellow hue. This is consistent with López- Rull et l. (21),whoreportednincreseinyellow/red chrom fter preen oil ddition to crotenoid-pigmented fethers of house finches (Crpodcus mexicnus). In contrst, Surmcki nd Nowkowski (27) reported somewht different result, s yellow hue, but not yellow chrom, ws ffected by preen oil removl. However, tht study used different method to mesure color, bsed on digitl photogrphs, which my explin, in prt, these discrepncies. In ddition, their experimentl tretment lso removed ll possible dirt ttched to the fether surfce, mking it difficult to distinguish the effect of preen oil vs. soiling. We lso found tht preen oil ddition reduced men brightness, consistent with the results of Surmcki nd Nowkowski (27) nd López-Rull et l. (21), lthough these effects were below the discrimintion threshold of birds. positive effect of preen oil ccumultion on fether glossiness might hve lso been expected (Delhey et l. 27). However, unfortuntely, glossiness could not be quntified in our study from the reflectnce spectr mesured.

9 ehv Ecol Sociobiol Given the importnce of plumge colortion for socil signling (Hill nd McGrw 26), the effects of preen oils nd dirt on fether colortion could influence how crotenoid-bsed signls of individul qulity re perceived. In fct, our experimentl tretments influenced color trits tht differed between the sexes nd ge clsses of ohemin wxwings (see Online Resource 2), nd these effects were noticeble either by the UVS or VS species models, s reveled by the significnt chnges in chromtic contrsts between control nd experimentl fethers. Thus, it seems resonble to expect tht soil nd preen oil ccumultion would ffect the signl content of crotenoid-bsed colortion in t lest this species. The yellow chrom of crotenoid-pigmented fethers reflects their crotenoid content (Sks et l. 23; Isksson et l. 28), which my be informtive of severl spects of individul qulity (reviewed by McGrw 26). Therefore, the positive effect of preen oils on yellow chrom found in this study suggests tht they could serve cosmetic function (Delhey et l. 27), cting s enhncers of the signl. The honesty of this signl-enhncing mechnism would require the existence of some costs for preen oil ppliction (Zhvi nd Zhvi 1997). These costs my rise directly from preen oil production: the ctivity of the uropygil glnd is stimulted by incresed testosterone (Ghosh nd httchryy 1996) nd mintining high testosterone levels my be costly due to the collterl immunosuppressive effects of the hormone (Folstd nd Krter 1992). lso, the preen oil production my be constrined by individul nutritionl sttus (Ok nd Okuym 2). Preen oil ppliction to fethers my lso be influenced by indirect costs, such s the mount of time nd energy required for preening. This is not negligible cost, s preening ctivities consume significnt portion of bird s dytime budget (Wlther nd Clyton 25), nd the energetic cost of preening hs been estimted s twice the bsl metbolic rte (Goldstein 1988). We lso found tht preen oil ddition incresed UV hue nd, in the cse of mgpie preen oil, UV chrom. Unlike yellow chrom, the UV reflectnce of crotenoid-pigmented plumges is cused by fether microstructures (Shwkey nd Hill 25). lthough UV nd yellow reflectnce peks of crotenoid-pigmented fethers hve been positively relted (Mys et l. 24; Shwkey et l. 26), UV chrom does not seem to be relted to the fether s crotenoid content (Shwkey et l. 26). Therefore, more reserch is needed on the biologicl mening of UV reflectnce of crotenoid-bsed plumges (e.g., Shwkey nd Hill 25; Shwkey et l. 26; Glván et l. 28) to interpret the effects of preen oils on UV reflectnce. Nevertheless, the effect of preen oils on UV reflectnce could be importnt in those species with structurl plumge ornments where high UV-blue reflectnce plys role in socil signling (e.g., Sheldon et l. 1999; Sieffermn nd Hill 25). In those cses where high UV chrom/low UV hue re cues for sexul selection, individuls my experience trdeoff between signl mximiztion nd receiving the benefits of properly oiled plumge (Jcob nd Ziswiler 1982; urtt nd Ichid 1999), thus dding n extr cost to signl mximiztion tht my increse signl honesty. nother interesting finding is tht preen oil from blckbilled mgpies (Psseriformes) ffected UV chrom, wheres preen oils from egle owls (Strigiformes) did not. recent study (Delhey et l. 28) tested the effect of preen oils from 51 species belonging to 12 vin orders on UV/visible reflectnce of white UV-reflective surfce (Teflon tpe) nd found tht uropygil secretions reduced overll brightness nd especilly UV chrom, nd prticulrly for non-psserine secretions. Unfortuntely, tht study did not include preen oils from ny species of owl. The contrsting effect of mgpie nd owl preen oils on fether UV chrom is difficult to interpret under signling scenrio becuse mgpies nd owls hve different visul systems. Corvids, where blck-billed mgpies re included, re VS species (Ödeen nd Håstd 23) nd cn perceive vritions in the ner-uv rnge (Cuthill 26). In contrst, owls nd other nocturnl rptors pprently lck UVS or VS cones (owmker nd Mrtin 1978; Cuthill 26), lthough recent experiment (Prejo et l. 21) indictes tht t lest some species cn detect vritions in the UV rnge. We were not ble to use preen oils from wxwings; however, the removl of wxwing preen oil nd mgpie preen oil ddition exerted, s expected, similr but opposite effects on fether reflectnce prmeters. We believe tht our results re biologiclly meningful given tht the opticl properties of uropygil secretions of ll psseriforms re roughly similr (Delhey et l. 28). Our tretment with wter (soil-removed fethers) did not totlly remove soil, nd certin proportion of dirt likely remined ttched to the fethers, mixed with preen oils (personl observtion). Soil nd preen oil removl reduced UV hue nd incresed UV chrom (which is bsiclly opposite to the effect of preen oil ddition), indicting tht soil ppers to hve little effect on plumge reflectnce in the UV rnge s compred to preen oil. However, soil nd preen oil hd opposite effects on yellow chrom: preen oil ddition incresed yellow chrom, nd the sme effect ws obtined fter prtil soil removl. In greement with this result, we found tht concurrent removl of preen oil nd soil prtilly reverted the effect of soil removl lone (Fig. 2f). s yellow chrom is often indictive of individul qulity (see bove), our results re consistent with the hypothesis tht dirt ccumultion on fethers could reduce plumge ttrctiveness, nd so preening my enhnce the ttrctiveness of

10 ehv Ecol Sociobiol individuls (Zmpig et l. 24; Montgomerie 26; Lenouvel et l. 29; López-Rull et l. 21). We should mention tht, unexpectedly, control fethers showed some chnges in reflectnce during the experiment (prticulrly in Experiment 2). This is likely to be ttributble to fether hndling (e.g., disordered fether brbs) or some minor degrdtion of pigments during the experiment. lthough this does not ffect our results s ll fethers were mnipulted in the sme wy, extreme cre should be tken when hndling fethers in future studies involving color mesurements. In conclusion, lthough plumge colortion hs been trditionlly considered reltively sttic trit, severl studies hve shown tht it cn chnge due to fether wer or pigment degrdtion (Örnborg et l. 22; McGrw nd Hill 24; Figuerol nd Senr 25). Our results suggest tht fether colortion my lso be modified by preen oil content nd soiling of fethers. s plumge colortion usully cts s cue for sexul selection, it is likely tht individuls re ble to modify their ttrctiveness by investing in preening ctivities, thereby enhncing signl expression. cknowledgments We re grteful to Elenis Crespo nd Rsp de Prd (Centro de Recuperción de Fun Silvestre El Chprrillo ) for their help during preen oil collection from egle owls. Doris Gomez nd Jesús M. vilés kindly provided dvice on some computtionl detils of visul models. We lso thnk Jesús T. Grcí for genetic sexing of the birds nd Ismel Glván nd two nonymous referees for their comments on the mnuscript. L. P.-R. ws supported by postdoctorl contrct (7/28-) from Junt de Comuniddes de Cstill-L Mnch (JCCM) nd Jun de l Cierv contrct from the Spnish Ministerio de Cienci e Innovción (MICINN). Funding ws lso provided by Nturl Sciences nd Engineering Reserch Council of Cnd nd the Sturt nd Mry Houston Professorship in Ornithology (to G. R. ). References ndersson S, Prger M (26) Quntifying colors. In: Hill EG, McGrw KJ (eds) ird colortion, vol 1: mechnism nd mesurements. Hrvrd University Press, Cmbridge, pp vilés JM, Soler JJ (29) Nestling colortion is djusted to prent visul performnce in ltricil birds. J Evol iol 22: leiweiss R (25) Vrition in ultrviolet reflectnce by crotenoid bering fethers of tngers (Thrupini: Emberizine: Psseriformes). iol J Linn Soc 84: owmker JK, Mrtin GR (1978) Visul pigments nd colour vision in nocturnl bird, Strix luco (twny owl). Vis Res 18: urtt EH, Ichid JM (1999) Occurrence of fether-degrding bcilli in the plumge of birds. uk 116: Cuthill IC (26) Color perception. In: Hill EG, McGrw KJ (eds) ird colortion, vol 1: mechnism nd mesurements. Hrvrd University Press, Cmbridge, pp 3 4 Delhey K, Peters, Kempeners (27) Cosmetic colortion in birds: occurrence, function, nd evolution. m Nt 169:S145 S158 Delhey K, Peters, iedermnn PHW, Kempeners (28) Opticl properties of the uropygil glnd secretion: no evidence for UV cosmetics in birds. Nturwissenschften 95: Fivre, Grégoire, Préult M, Cézilly F, Sorci G (23) Immune ctivtion rpidly mirrored in secondry sexul trit. Science 3:13 Figuerol J, Senr JC (25) Sesonl chnges in crotenoid- nd melnin-bsed plumge colortion in gret tit Prus mjor. Ibis 147: Folstd I, Krter J (1992) Prsites, bright mles nd the immunocompetence hndicp. m Nt 139: Glván I, mo L, Snz JJ (28) Ultrviolet-blue reflectnce of some nestling plumge ptches medites prentl fvouritism in gret tits Prus mjor. J vin iol 39: Ghosh, httchryy SP (1996) Endocrine fctors in the regultion of the ctivity of the uropygil glnd of vin skin. Proc Indin Ntl Sci cd iol Sci 66: Goldstein DL (1988) Estimtes of dily energy expenditure in birds: the time-energy budget s n integrtor of lbortory nd field studies. m Zool 28: Gomez D (26) VICOL, progrm to nlyse spectrometric dt. Free progrm vilble t or from the uthor t dodogomez@yhoo.fr. ccessed 14 June 21 Hrt NS (22) Vision in the pefowl (ves: Pvo cristtus). J Exp iol 25: Hrt NS, Hunt DM (27) vin visul pigments: chrcteristics, spectrl tuning, nd evolution. m Nt 169:S7 S26 Hrt NS, Prtridge JC, Cuthill IC, ennett TD (2) Visul pigments, oil droplets, oculr medi nd cone photoreceptor distribution in two species of psserine bird: the blue tit (Prus ceruleus L.) nd the blckbird (Turdus merul L.). J Comp Physiol 186: Husmnn F, rnold KE, Mrshll NJ, Owens IPF (23) Ultrviolet signls in birds re specil. Proc R Soc Lond 27:61 67 Hill EG, McGrw KJ (eds) (26) ird colortion, vol 2: function nd evolution. Hrvrd University Press, Cmbridge Hill EG, McGrw KJ (eds) (26b) ird colortion, vol 2: mechnisms nd mesurements. Hrvrd University Press, Cmbridge Hudon J, rush H (1989) Probble dietry bsis of color vrint of the cedr wxwing. J Field Ornithol 6: Isksson C, Örnborg J, Prger M, ndersson S (28) Sex nd ge differences in reflectnce nd biochemistry of crotenoid-bsed colour vrition in the gret tit Prus mjor. iol J Linn Soc 95: Jcob J, Ziswiler V (1982) The uropygil glnd. In: Frner DS, King JR, Prkes KC (eds) vin biology, vol VI. cdemic, New York, pp Lenouvel P, Gomez D, Théry M, Kreutzer M (29) Do grooming behviours ffect properties of fethers in mle domestic cnries, Serinus cnri? nim ehv 77: López-Rull I, Pgán I, Mcís Grcí C (21) Cosmetic enhncement of signl colortion: experimentl evidence in the house finch. ehv Ecol 21: Mrtínez-Pdill J, Mougeot F, Pérez-Rodríguez L, ortolotti GR (27) Nemtode prsites reduce crotenoid-bsed signlling in mle red grouse. iol Lett 3: Mys HL, McGrw KJ, Ritchison G, Cooper S, Rush V, Prker RS (24) Sexul dichromtism in the yellow-brested cht Icteri virens: spectrophotometric nlysis nd biochemicl bsis. J vin iol 35: McGrw KJ (26) Mechnics of crotenoid-bsed colortion. In: Hill EG, McGrw KJ (eds) ird colortion. vol 1: mechnism nd mesurements. Hrvrd University Press, Cmbridge, pp

11 ehv Ecol Sociobiol McGrw KJ, Hill GE (24) Plumge color s dynmic trit: crotenoid pigmenttion of mle house finches Crpodcus mexicnus fdes during the breeding seson. Cn J Zool 82: Montgomerie R (26) Cosmetic nd dventitious colors. In: Hill EG, McGrw KJ (eds) ird colortion, vol 1: mechnism nd mesurements. Hrvrd University Press, Cmbridge, pp Montgomerie R (26b) nlyzing colors. In: Hill EG, McGrw KJ (eds) ird colortion, vol 1: mechnism nd mesurements. Hrvrd University Press, Cmbridge, pp Montgomerie R, Lyon, Holder K (21) Dirty ptrmign: behviorl modifiction of conspicuous mle plumge. ehv Ecol 12: Mountjoy DJ (25) Fmily ombycillide (Wxwings). In: del Hoyo J, Elliott, Christie D (eds) Hndbook of the birds of the world, vol 1: Cuckoo-shrikes to thrushes. Lynx Edicions, rcelon, pp Negro JJ, Mrglid, Hirldo F, Heredi R (1999) The function of the cosmetic colortion of berded vultures: when rt imittes life. nim ehv 58:F14 F17 Ödeen, Håstd O (23) Complex distribution of vin color vision systems reveled by sequencing the SWS1 opsin from totl DN. Mol iol Evol 2: Ok N, Okuym M (2) Nutritionl sttus of ded oiled rhinoceros uklets (Cerorhinc monocert) in the southern Jpn se. Mr Pollut ull 4: Örnborg J, ndersson S, Griffith SC, Sheldon C (22) Sesonl chnges in ultrviolet structurl colour signl in blue tits, Prus ceruleus. iol J Linn Soc Lond 76: Prejo D, vilés JM, Rodríguez J (21) Visul cues nd prentl fvouritism in nocturnl bird. iol Lett 6: Pérez-Rodríguez L (28) Crotenoid-bsed ornmenttion s dynmic but consistent individul trit. ehv Ecol Sociobiol 62: Piersm T, Dekker M, Dmste JSS (1999) n vin equivlent of mke-up? Ecol Lett 2:21 23 Reneerkens J, Korsten P (24) Plumge reflectnce is not ffected by preen wx composition in red knots Clidris cnutus. J vin iol 35:45 49 Sks L, McGrw KJ, Hõrk P (23) How fether colour reflects its crotenoid content. Funct Ecol 17: Shwkey MD, Hill GE (25) Crotenoids need structurl colours to shine. iol Lett 1: Shwkey MD, Pilli SR, Hill GE (23) Chemicl wrfre? Effects of uropygil oil on fether-degrding bcteri. J vin iol 34: Shwkey MD, Hill GE, McGrw KJ, Hood WR, Huggins K (26) n experimentl test of the contributions nd condition-dependence of microstructure nd crotenoids in yellow plumge colourtion. Proc R Soc Lond 273: Sheldon C, ndersson S, Griffith SC, Örnborg J, Sendeck J (1999) Ultrviolet colour vrition influences blue tit sex rtios. Nture 42: Sieffermn L, Hill GE (25) UV-blue structurl colortion nd competition for nest boxes in mle estern bluebirds. nim ehv 69:67 72 Surmcki (28) Preen wxes do not protect crotenoid plumge from bleching by sunlight. Ibis 15: Surmcki, Nowkowski JK (27) Soil nd preen wxes influence the expression of crotenoid-bsed plumge colourtion. Nturwissenschften 94: Vorobyev M (23) Coloured oil droplets enhnce colour discrimintion. Proc R Soc Lond 27: Vorobyev M, Osorio D (1998) Receptor noise s determinnt of colour thresholds. Proc R Soc Lond 265: Vorobyev M, Osorio D, ennett TD, Mrshll NJ, Cuthill IC (1998) Tetrchromcy, oil droplets nd bird plumge colours. J Comp Physiol 183: Wlther, Clyton DH (25) Elborte ornments re costly to mintin: evidence for high mintennce hndicps. ehv Ecol 16:89 95 Zhvi, Zhvi (1997) The hndicp principle. missing piece of Drwin s puzzle. Oxford University Press, Oxford Zmpig E, Hoi H, Pilsto (24) Preening, plumge reflectnce nd femle choice in budgerigrs. Ethol Ecol Evol 16: