THE CONDOR

AJOURNAL OF AVIAN BIOLOGY UWWN-Y OF IDAHOl_lBRARY

Volume 93 Number 2 AUG !2 &# ay 1991

The Condor93:209-235 0 The Cooper Ornithological Society I99 1

THE SEQUENCE OF MOLTS AND PLUMAGES IN PAINTED BUNTINGS AND IMPLICATIONS FOR THEORIES OF DELAYED PLUMAGE MATURATION’

CHRISTOPHER W. THOMPSON Department of Zoology, State University,Tempe, AZ 85287-1501

Abstract. Within a week of fledging,juvenile Painted Buntings, ciris, undergo a previously unrecognizedfirst prebasicmolt that includes most to all body plumage except greater primary and greater secondarycoverts. This molt occurs from early June to early October, usually on the breeding ground. The resulting first basic plumage is similar to juvenal plumage, but is more adult female-like in color in both sexes. First-year, but not adult, Painted Buntings subsequentlyundergo a presupplemental molt during which the outer four or five primaries, inner four to six secondaries,all remiges, and all body plumage except some to all greaterprimary coverts typically are replaced.The resultingsupplemental plumage is identical in both sexesand is more adult female-like in color than the previous first basicplumage. Presupplementalmolt usuallyoccurs between early Septemberand early November. Most (90%) Painted Buntingsalong the easterncoast ofthe United Statesundergo this molt on the breeding ground prior to fall migration; whereas, in the western United States, most (60%) molt in exclusively migratory areas of the desert southwesternUnited Statesand northwestern . These desert locationsexhibit a large increasein plant and life in response to “monsoon” rains that occur predictably in these areas in late summer and fall. Selection may have favored evolution of molt-migration strategiesin Painted Buntings in the western United States as a mechanism to allow them to molt in areas with greater food resourcesthan exist at the same time of year on their breeding or wintering ground. Painted Buntingsare one of only six passerinesin which molt-migration has been documented. All age and sex classesundergo a partial prealtemate molt that was describedpreviously (Fisk 1974) but has been overlooked by all subseauentinvestigators. This molt usuallv occurson the wintering ground, and is typically limited to the her& breast, and belly. Most plumagegrown by subadultmales during first prealtemate molt is adult female-like in color. This is the only passerineknown in which sexuallymature subadultmales grow adult female- like rather than adult male-like plumage during prealtemate molt. Juvenal plumage has poor structural integrity compared to subsequentplumages. Its structure facilitates heat transfer to the young by brooding females, but is poorly adapted to protectingjuveniles from adverseeffects of abiotic factors. Therefore, selectionmay have favored replacement of juvenal plumage with a structurally strongerfirst basic plumage as soon as possible after fledging. First-year males and females may undergo presupplementalmolt to prevent being iden- tified as first-year by adult females and being dominated by them (and possibly by adult males as well if adult males do not usually dominate adult females), or to reduce predation on themselves.That no subadultmales acquire a winter plumage intermediate in color between that of adult males and adult females in winter is the first empirical support for Rohwer et al.‘s (1980) model which indicates that selectionshould favor subadultmales with plumagesthat are completely adult female-like or adult male-like more than plumages that are intermediate between adult males and adult females. These resultsalso support the winter female mimicry hypothesis(Brown and Brown 1988), are consistentwith the winter

I Received 17 October 1990. Final acceptance3 January 1991.

12091 210 CHRISTOPHER W. THOMPSON

cryptic hypothesis (Rohwer and Butcher 1988), and are inconsistentwith the winter status signalinghypothesis (Rohwer 1975, 1982). Becausesubadult males grow plumage during first prealtemate molt that is adult female- like in color, this strongly suggeststhat the resulting plumage functions during their first potential breeding season to mimic females. This supports the summer female mimicry hypothesis(Rohwer et al. 1980), but is inconsistentwith the summer cryptic (Sclander 1965, 1972), summer statussignaling (Lyon and Montgomerie 1986, Montgomerie and Lyon 1986) andjuvenile mimicry hypotheses(Lawton and Lawton 1986, Foster 1987).This also supports Rohwer et al.3 (1980) model discussedabove. Adult males exhibit a greater rate of feather replacement during prealtemate molt than any other age or sex class. This suggeststhat the extent of this molt is less constrained in adult males than other age and sex classes,thus weakly supporting the molt constraints hypothesis (Rohwer and Butcher 1988). Becausesubadult males exhibit only a limited prealtemate molt, their summer plumage may not be functionally independent of their winter plumage. Thus, it is not possible to conclusively determine whether subadult male plumagesare adaptive during winter, sum- mer, or both. Becausepresupplemental body molt is essentially complete, this clearly in- dicates that failure of subadult males to obtain fully adult male-like supplementalplumage can not be due to energeticconstraints. Key words: Molt migration: delayedplumage maturation: Passerina;museum specimens; color; energetics;constraints.

INTRODUCTION ually monochromatic and dichromatic, that have In nearly all speciesof birds, females reach sexual a variety of mating systemsand migratory strat- and somatic maturity at the same time (but see egies, and that occur throughout the world in Ligon 197 1, Hussell 1983, Stutchbury and Rob- diverse habitats ranging from arctic tundra at ertson 1987, Mountjoy and Robertson 1988). high latitudes to tropical rain forests at low lat- However, males of many speciesdo not attain itudes. somatic maturity in plumage or bare part color All hypotheses for the evolution of delayed for months to years after reaching sexual matu- plumage maturation propose that the relatively rity. In sexually mature but somatically imma- cryptic plumage of subadult males is adaptive ture (subadult sensuRohwer et al. 1980, Rohwer either during their first non-breeding season and Butcher 1988) males, these characters are (winter hypotheses)or their first potential breed- more cryptic than those in adult males (Rohwer ing season (summer hypotheses). The function et al. 1980, Lawton and Lawton 1986, Butcher of plumage color during breeding season(usually and Rohwer 1989) and usually are very similar summer) may be considered independent of to those of adult females. In many of these spe- plumage color during nonbreeding season (usu- cies subadult males successfullybreed occasion- ally winter) only when a complete body molt ally to regularly (e.g., Diamond 1972, Forshaw occurs between these seasons (Rohwer 1986, and Cooper 198 1, Proctor-Gray and Holmes Rohwer and Butcher 1988, Butcher and Rohwer 1981,Steenhofetal. 1983,Goodwin 1986,Grant 1989). 1986, Lawton and Lawton 1986, Rohwer and Although most studies of delayed plumage Butcher 1988). Such delayed acquisition of adult maturation have argued that subadult plumages plumage by sexually mature birds is termed de- evolved as adaptations for breeding, Rohwer and layed plumage maturation. Delayed plumage Butcher (1988) elegantly and unambiguously maturation has been studied almost exclusively demonstrated that subadult plumages often are in North American (but see Foster adaptive only in winter. Males of 16 specieswear 1987, Jar-vi et al. 1987) but its occurrence is subadult plumages only during their first winter. geographically and taxonomically more wide- In all of these species,subadult males undergo a spread than is generally recognized. It occurs in complete body molt in spring which causesthem at least seven orders and 11 families of non- to lose their adult female-like winter plumage passerinesand at least 21 families of passerines and acquire a summer plumage that is very sim- (e.g., Mayr 1933, 1934; Lack 1968; Diamond ilar or identical to that of adult males. This in- 1972; Forshaw and Cooper 198 1; Schodde 1982; dicates that subadult plumage in these species Grant 1986; Lawton and Lawton 1986; Rohwer must be an adaptation to winter. and Butcher 1988) including speciesthat are sex- Males of 21 species wear subadult plumage DELAYED PLUMAGE MATURATION 211

during their first winter and subsequentsummer. tings are an ideal choice of speciesin which to In all of these species,subadult males undergo a look for a spring molt. In addition, three winter limited springbody molt thereby acquiring a more and five summer hypothesesimply testable pre- adult male-like plumage. Because this molt is dictions regarding subadult male molts and incomplete, summer plumage may not be func- plumages. These predictions were tested by de- tionally independent of winter plumage. There- termining the sequenceof molts and plumages, fore, this finding does not unequivocally support and resulting changesin plumage color, exhibited either winter or summer hypotheses.All plum- by male Painted Buntings from hatching to so- age replaced during spring molt in these 37 spe- matic maturity. The relevant winter and summer cies is always adult male-like and never adult hypothesesare summarized below. The predic- female-like or cryptic (Rohwer and Niles 1979, tions of these hypothesesare summarized in Ta- Rohwer et al. 1983, Rohwer 1986, Rohwer and ble 1 and explained in the discussion. Butcher 1988). All winter and summer hypothesesassume that Fourteen species of North American passer- competition is intense among age classesof males ines have equally cryptic winter and summer during winter and summer, respectively. Lessex- subadult male plumages (Table 1, category 3 in perienced subadult males are presumably com- Rohwer and Butcher 1988). Becausethere is no petitively disadvantaged relative to older males. evidence of a spring molt in any of these species, All of these hypotheses, except the breeding Rohwer and Butcher (1988) explicitly assume thresholdhypothesis (Studd and Robertson 1985, that these speciesdo not change plumage color Montgomerie and Lyon 1986) also proposethat from winter to summer becausethey lack a spring delayed plumage maturation has evolved in sub- molt. However, only one of these species, the adult males either to reduce predation or to re- Northern Oriole, Zcterusgalbula (Rohwer and duce aggressiondirected toward them by adult Manning 1990) has been specifically investigat- males, and differ only in the proposedproximate ed for the occurrenceof a spring molt. Evidence mechanisms by which subadult males reduce ag- that subadult males of any of these speciesre- gression directed toward them. The breeding place a cryptic winter plumage with either an threshold hypothesis is an extension of theory equally or more cryptic or a more adult female- developed by Wiley (1974, 1981) and Witten- like summer plumage during a complete spring berger (1978, 1979), and is not addressedby this body molt would clearly demonstrate, in the spe- study because it makes no testable predictions ties in question, that it is more adaptive for them regarding seasonalchanges in plumage color and to wear subadult male than completely adult- does not invoke a specific adaptive proximate male-like plumage in summer. If the spring body function to subadult plumage. Instead, it argues molt is incomplete, this does not necessarilyin- that selection for reduced male-male competi- dicate that it is more adaptive for them to wear tion has favored the evolution of subadult plum- subadult male than completely adult-male-like age and increased body size in birds that expe- plumage in summer, but may indicate that it is rience intense male-male competition. more adaptive to wear a subadult plumage that is completely adult female-like rather than in- WINTER HYPOTHESES termediate in color between adult females and The winter status signaling hypothesis(Rohwer males (Rohwer et al. 1980, 1983; Rohwer and 1975, 1977, 1978a, 1982, 1985; Rohwer and Butcher 1988). Rohwer 1978; Ketterson 1979; Rohwer and I initiated this study to determine whether a Ewald 198 1; Rohwer and Butcher 1988) states spring molt occursin one of these 14 species,the that plumage color functions intraspecifically to Painted Bunting, Passerina ciris. Subadult male honestly signal dominance rank within social Painted Buntings are very cryptic in color and gmups, conspicuous color being dominant to exhibit the least change in plumage color from more cryptic color. Status signals may also func- winter to summer compared to subadult males tion to indicate inherent cost-benefit asymme- of the other 13 species.In addition, subadults of tries among group members (Parker 1974, May- congeneric Indigo Buntings, Passerina cyanea, nard Smith and Parker 1976, Ewald and Rohwer undergo a spring molt of some to all body plum- 1980). age, except greater primary coverts (Rohwer The winter cryptichypothesis (Ewald and Roh- 1986). This suggeststhat Painted Buntings may wer 1980, Rohwer et al. 1983, Rohwer 1986, exhibit a similar molt. Therefore, Painted Bun- Rohwer and Butcher 1988) proposesthat the rel- 212 CHRISTOPHER W. THOMPSON

TABLE 1. Seasonalchanges in plumagecolor in subadultmale PaintedBuntings and their implicationsfor winter and summerhypotheses for the evolutionof delayedplumage maturation (DPM).

Possiblechanges in plumagecolor in first-year Possiblechanges in plumagecolor in subadult malesin any specieswith DPM from summer malesin any specieswith DPM from winter to to winter’ that would suppmi(S), be consistent summerthat would support(S), be consistent with (C) or be inconsistentwith (I) the hy with (C) or be inconsistentwith (I) the hypo- pothesis.Observed change in first-yearmale thesis.Observed change in subadultmale Name of Hypothesis PaintedBuntings is italicized PaintedBuntings is italicized Winter StatusSignaling More male-like (S) More male-like(C) No change(C) No change(C) More female-like (I) More female-like(S) Winter Cryptic More conspicuous(I) More conspicuous(C) No change(C) No change(C) Lessconspicuous (S) Lessconspicuous (I) Winter FemaleMimicry More male-like (I) More male-like(C) No change(C) No change(C) More female-like (S) More female-like(I) SummerStatus Signaling No predictions Lessfemale-like (S) No change (C,= P) More female-like(I) SummerCryptic No predictions More conspicuous(I) No changed(C,5 P) Lessconspicuous (S) SummerFemale Mimicry No predictions Lessfemale-like (I) No chang&J (I,6 S) More female-like(S) JuvenileMimicry No predictions Lessjuvenile-like (I) No chang& (S,” Zl”) More juvenile-like (S) Molt Constraints No predictions No predictions

I Changein plumagecolor from ~ummerto winter may resultfrom first prebasicmolt, presupplementalmolt, featherwear, or a combination thereof. i No changein plumagecolor would be consistentwith this hypothesisonly if subadultmale winter plumagecolor is intermediatebetween that of adult malesand females,which is not the casein PaintedBuntings. ’ No changein plumagecolor would be inconsistentwith this hypothesisif subadultmale winter plumagecolor is completelyadult female-like, as occursin PaintedBuntings. 4 43% of subadultmale PaintedBuntings grow someadult male-like(blue) head plumage during first prealtemate molt (seeTable 2). 5No changein plumagecolor would be consistentwith this hypothesisonly if subadultmale winter plumagecolor is maximallycryptic, which is not the casem PaintedBuntings. 6No changein plumagecolor would be inconsistentwith thishypothesis only if subadultmale winter plumage color is not completelyadult female- like, which is not the casein PaintedBuntings. ’ No changein plumagecolor would supportthis hypothesis only if subadultmale winter plumagecolor is completelyadult female-lie, as occurs in PaintedBuntin s. 8 Becausesuba d:ult male and femalePainted Buntings are identicalin plumagecolor in winter, it is not possiblefor subadultmales to become more adult female-likein plumagecolor in summer. e No changein plumagecolor would supportthis hypothesisonly if subadultmale winter plumagecolor is completelyjuvenile-like, which is not the casein PaintedBuntmgs. ‘0 No changein plumagecolor would be inconsistentwith this hypothesisif subadultmale winter plumagecolor is not completelyjuvenile-like, as is the casein PaintedBuntings.

atively cryptic plumage of subadultsreduces their behavior of males to many or all adult females conspicuousnessto both predators and conspe- other than their mates and to subadult males in cific adult males during their first winter. adult female-like plumage, one must assume 1) The winterfemale mimicry hypothesis (Brown that reciprocal altruism (Trivers 197 1) exists and Brown 1988) suggeststhat, in speciesin which among males toward these females or 2) that by adult males preferentially allow females access doing so, males may increase their probability to resources,subadult males with adult female- of pairing and/or mating with these females in like plumage may mimic females to increasetheir the future. accessto resourcesin winter. Subadult males may exploit this relationship between adult males and SUMMER HYPOTHESES females by wearing an adult female-like plumage The summer status signaling hypothesis (Lyon that deceives adult males into identifying them and Montgomerie 1986, Montgomerie and Lyon as females and behaving subordinately toward 1986) proposesthat subadult male plumage color them. For this hypothesisto explain subordinate is a reliable signal of fighting ability rather than DELAYED PLUMAGE MATURATION 213

a deceitful indicator of sex. It requires that 1) having subordinately toward them, thereby in- male-male competition is intense due to a strong- creasing accessto food by subadult males. For ly male-biased operational sex ratio, 2) females this hypothesis to explain subordinate behavior choose mates based on intrinsic male attributes of males to many or all adult females other than such as plumage color, but not on extrinsic cues their mates and to subadult males in adult fe- such as territory quality, and 3) subadult males male-like plumage, one must assume 1) that re- have “lower intrinsic resourceholding potential ciprocal altruism (Trivers 1971) exists among than adult [males]” which precludes them from males toward these females or 2) that by doing winning territorial contests with adult males. It so, males may increase their probability of pair- argues that when these conditions obtain, sub- ing and/or mating with these females in the fu- adult males do not pose a threat to adult males ture. and, therefore, should be tolerated by them. The juvenile mimicry hypothesis(Lawton and The summer cryptichypothesis (Selander 1965, Lawton 1986, Foster 1987) proposesthat 1) ju- 1972; Lack 1968; Proctor-Gray and Holmes venile rather than female-like morphological and 198 1) assumesthat the primary function of con- behavioral characteristics are retained by sub- spicuouscolor is to increase successin male-male adult males during their first potential breeding competition for breeding resources,but that con- season,and 2) adult males display lessaggression spicuous color also increases predation and/or toward subadult than adult males because sub- intraspecific competition relative to more cryptic adult males exhibit juvenile-like rather than color. It also assumesthat, on average, subadult female-like or adult male-like characteristics. males that attempt to breed during their first Foster (1987) further assertsthat retention ofju- potential breeding season achieve lower repro- venile-like characteristics during the breeding ductive successthan breeding adult males. Last, season by sexually mature subadult males de- it argues that the relatively higher reproductive ceitfully signals nonreproductive status,whereas successthat may be gained by subadult males in Lawton and Lawton (1986 and pers. comm.) more adult male-like than adult female-like maintain that retention of such juvenile-like breeding plumage is outweighed by their pro- characterscould act either as deceitful signals of portionally higher mortality than subadult males age or as honest signals of subordinance. in more adult female-like breeding plumage. The molt constraintshypothesis (Rohwer 1986, The summerfemale mimicry hypothesis(Roh- Rohwer and Butcher 1988) proposes that sub- wer 1978b, 1983; Rohwer and Niles 1979; Roh- adult males are unable to attain adult breeding wer et al. 1980, 1983) is based on the premise plumage by their potential breeding season be- that adult males minimize aggressiontoward fe- causeofeither 1) an inability to acquire sufficient males during breeding season to facilitate pair- resourcesto supply the energy for a late winter bonding and extra-pair copulations. It assertsthat or spring molt or 2) genetic constraints that have adult female-like breeding plumage in subadult prevented the evolution of a late winter or spring males is adaptive. For a short time after arrival molt unique to subadult males (Parkes 1967). on the breeding ground, subadult males behave Cryptic plumagesworn by subadult males during cryptically. As a result, adult males mistake sub- their Hurstpotential breeding season are, there- adult males for females for a sufficient time for fore, viewed as maladaptive. subadult males to establish site dominance and Last, if subadult males do not replace all of gain breeding territories capable of attracting re- their body plumage during a prebreeding molt in productive females. This strategy should in- late winter or spring, then their winter and sum- crease both survivorship and reproductive suc- mer plumages may not be functionally indepen- cessof subadult males. dent. In such cases,no definite conclusions can An alternative explanation for female mimicry be made regarding the seasonto which subadult is that selection may favor “males to allow fe- male plumages are adapted. males accessto critical resources that may be channeled into the production of eggsand young, METHODS in which males have genetic interest” (Brown The data for this study were collected from live 1984). Subadult males may exploit this relation- specimens examined during a field study of this ship between adult males and females by wearing species at McKinney Falls State Park an adult female-like plumage that deceivesadult (30”11’28”N, 98”43’19”W), Travis County, Tex- males into identifying them as females and be- as from May through July, 1984, and over 2,600 214 CHRISTOPHER W. THOMPSON

museum skins collected during all months of the culated by summing total primary, secondary, year from throughout their breeding and win- and rectrix molt scores.Painted Buntings have tering ranges. All museum skins were examined 18 primaries, 18 secondaries, and 12 rectrices. under a 100-W incandescent light with a 3.5x Adults replace all flight feathersduring definitive jeweler’s optiviser. prebasic molt. Therefore, maximum total pri- mary, secondary, rectrix, and total flight feather SCORING MOLT molt scoresare 90,90,60, and 240, respectively, Terminology regarding names of molts and in adults. plumages follows Humphrey and Parkes (1959). First-year (born the current calendar year) birds The presenceof molting featherswas determined typically replace the outer four or five primaries, by lifting the featherswith forceps.The intensity inner four to six secondaries, and all rectrices ofbody molt was scoredseparately on the crown, during presupplemental molt. Primary molt is chin and throat, face (auricular region, ocular most frequently initiated at P6 (see data below). region, and lores), back, rump, breast, and belly. Therefore, first-year birds that had replaced P6 Definitions ofbody regionsfollow Rohwer (1986). or more distal primaries received a scoreof 5 for Following Rohwer (1986), the intensity of body each such primary. First-year birds that had re- molt in each region was scored as follows: 0 = placed P5 or more proximal primaries received no new feathers, 10 = l-20% new feathers, 30 a score of 0 for each such primary. Similarly, = 21110%, . . . 90 = 81-99%, 100 = body molt following replacement of the tertials (S7-S9), completed. To avoid mistaking adventitious molt subsequent secondary molt was most frequently for true molt, I assignedmolt scoresof 0 to spec- initiated with S5 (seedata below). Therefore, fust- imens showing less than five growing feathersin year birds that had replaced S5 or more proximal only a single body region. Specimenswith at least secondariesreceived a score of 5 for each such two growing feathersin each of two or more body secondary. Birds that had replaced S4 or more regions were considered to be in molt. The total distal secondariesreceived a score of 0 for each percentageof growing featherson each specimen such secondary. Therefore, maximum total pri- was estimated by calculating the average of the mary, secondary, rectrix, and total flight feather molt scoresfor the forehead, crown, nape, face, molt scoresare 40,50,60, and 150, respectively. and chin and throat regions, and is hereafter re- ferred to as total body molt score. In analyses of RATE AND DURATION OF prealtemate body molt, body molt scoreswere FLIGHT FEATHER MOLT used for areas on the head only because of the The rate and duration of flight feather molt in rarity of observing prealtemate body molt in oth- individuals, as opposed to populations, was es- er body regions on museum specimens (see dis- timated by regression of date on total primary cussion below). In addition, the location and ex- and total flight feather molt scores (see Pimm tent of blue and red plumage was noted for all 1976 for statistical methodology). females and subadult males. Flight feathers were scored on the following AGE AND SEX DETERMINATION scale modified from Ginn and Melville (1983): The plumages of males and females are identical, 0 = old feather, 1 = missing feather, 2 = new except for some blue head plumage in some sub- feather lessthan % grown, 3 = new feather greater adult males in first alternate plumage, until they than 13‘ grown and less than 2/3 grown, 4 = new begin definitive prebasic molt after their first po- feather greater than V3 grown and less than full tential breeding seasonat 12 to 17 months of age grown, 5 = new feather full grown and tm- (Table 2). Determination of sex of these speci- sheathed. Molting flight feathers were identified mens was based on information provided on by their symmetrical loss and replacement. Only specimen labels. Males and females in definitive feathers in molt, as opposed to feathers lost or basic or definitive alternate plumage were sexed replaced adventitiously, were counted. Total pri- by plumage. Subadults in supplemental or first mary, secondary, and rectrix molt scores were alternate plumage were distinguished in agefrom calculated by summing molt scoresfor individ- females in definitive basic or definitive alternate ual primaries (P) from both wings, secondaries plumage by the retention of some to all juvenal (S) from both wings, and all rectrices, respec- (brown) remiges and greater primary coverts tively. Total flight feather molt score was cal- (Storer 195 1, Fisk 1974). Similarly, first-year

216 CHRISTOPHER W. THOMPSON birds in juvenal or first basic plumage were dis- sex class(summarized in Table 2) and compare tinguished from second-year (born the previous these data to analogous data on Indigo Buntings calendar year) birds in first alternate plumage by (Rohwer 1986). I argue that similarities and dif- differences in plumage color (see Table 2). ferences in the sequenceof molts and plumages between thesetwo speciesindicate selection pres- DEFINITION OF BREEDING RANGE sureswhich have influenced the evolution of molt During the breeding season,Painted Buntings are strategiesin these two species. distributed as two disjunct eastern and western populations (Fig. 1). Unexpectedly, the breeding FIRST PREBASIC MOLT ranges of these two disjunct populations do not Sequenceand extent. This molt is exhibited only correspond to the breeding rangesof the two rec- by first-year males and females and has not been ognized subspecies,ciris and pallidior (American distinguished by previous investigators (Dwight Ornithologists’ Union 1957). Therefore, the dis- 1900, Storer 195 1, Fisk 1974, Oberholser 1974) junct eastern and western breeding populations as a separate molt from the subsequent presup- are hereafter referred to as eastern and western plemental molt (described below). Most or all Painted Buntings, respectively. Very few sight body plumage exceptgreater primary and greater records of Painted Buntings exist for 550 km secondary coverts is usually replaced. Incoming from eastern (8 lo5 1’W) to western Ala- first basic body feathers are distinguished from bama (87”42’W) (Ogden and Chapman 1967 and incoming juvenal feathersby their brighter color references cited therein). Similarly, of 1,625 (discussedbelow, see Table 2) and greater struc- specimenscollected within the breeding range of tural integrity (Dwight 1900). Replacement be- P. ciris, only 12 (0.7%) were collected within this gins on the upper back, followed by the head and 550 km gap, of which none were collected west nape, and then the back and rump. The breast, of 85”W. Subsequent discussion of breeding and flanks, and belly begin molting shortly after onset wintering ranges refers to exclusive ranges, and of molt on the dorsal side. First-year specimens does not include areas where breeding and win- were considered to be in first basic plumage if tering ranges overlap (Fig. 1). Geographic coor- most of their juvenal plumage (except greater dinates of collection localities were determined primary and greater secondary coverts) was re- from gazetteers,atlases, and maps. placed by first basic plumage and if they did not exhibit any molt as defined above. Of52 first- STATISTICAL ANALYSIS year birds in first basic plumage, 23 retained ju- Frequency data were analyzed in all casesusing venal body plumage other than greater primary the G-test with Williams’ correction (Sokal and and greater secondary coverts. These 23 speci- Rohlf 198 1:7 10). Total body molt score was mens all had retained juvenal breast, belly and treated as a ranked variable because it is the flank feathers, while only six had retained ju- average of body molt scoresof body areas that venal back or rump feathers; none had retained differ significantly in surfacearea and number of juvenal feathers on their head, chin or throat. feathers. In comparisons of rate and timing of Timing. Parmelee (1959) collected a series of onset of flight feather molt, variances of pairs of fledglingsof known agesin (UOKLA regression models were tested for homoscedas- 3407, 3409, 3410, 3412, 3414, and 3442; see ticity before comparing Y-intercepts and slopes. acknowledgments for explanation of museum The level of significancewas defined as P < 0.05 name abbreviations). These specimens indicate in all tests. that this molt begins within 15 days of hatching and within one week after fledging when some DESCRIPTION OF MOLTS juvenal body feathers and all remiges and rec- All descriptions of molts and plumages in Paint- trices are still sheathed and growing. This molt ed Buntings have been incomplete (Storer 195 1, is completed by about 35 days after hatching. Fisk 1974) or partially incorrect (e.g., Wilson and This was also the casefor nestlings and fledglings Bonaparte 183 1, Dwight 1900, Coues 1903, examined in Austin, (Thompson, unpub. Sprunt 1968, Oberholser 1974). Thus, I present data). BecausePainted Buntings are often multi- data concerning the timing, sequenceand extent brooded and fledge young as early as early June (proportion of plumage replaced) of molts, and and as late as early September (e.g., Parmelee color of the resulting plumages for each age and 1959) juveniles may be found in first prebasic DELAYED PLUMAGE MATURATION 217

BREEDING RANGE

m WINTERING RANGE

. DEFINITIVE PREBASIC FLIGHT FEATHER MOLT

0 PRESUPPLEMENTAL FLIGHT FEATHER MOLT

FIGURE 1. Map depicting the breeding and wintering rangesof Painted Buntings,and the locations at which first-year specimens in presupplemental flight feather molt and second-year or older specimens in definitive prebasicflight feather molt were collected.Numbers adjacentto symbolsindicate number of specimensexhibiting molt. Two numbers adjacent to a symbol indicate that both first-year and second-yearor older specimenswere found in flight feather molt at the same locality. The first number indicates the number of second-yearor older specimens, and the second number indicates the number of first-year specimens.The breeding and wintering distributions are based on 1) 776 collection localities of specimens examined in this study, 2) Banding Laboratory banding summary data and recovery data, 3) records of breeding, migratory, and wintering Painted Buntings in a) Birdlore, Audubon Field Notes, and American Birds, b) The Vermillion Flycatcher (newsletter of the Tucson Audubon Society), and c) the following distributional accounts:Dearborn (1907) Carriker (19 lo), Howell (1911, 1928, 1932) Barbour (1923, 1943) Nice (1931) Bailey (1928) Bennitt (1932) Griscom (1932) Dickey and Van Rossem (1938) Oberholser (1938, 1974), Burleigb (1944, 1958) Goodrich (1946) Sprunt (1954, 1968) Lowery (1955) Friedmann et al. (1957) Urban (1959) Ligon (1961) Smithe and Paynter (1963) Phillips et al. (1964), Slud (1964) Johnston (1965) Smithe (1966) Ogden and Chapman (1967) Sutton (1967) Monroe (1968), Taber (1968), Land (1970) Sprunt and Chamberlain (1970), Paterson (1972) Peterson and Chalif (1973) Imhof (1976) Bond (1980), Potter et al. (1980) Norris and Elder (1982), Wetmore et al. (1984) Wood and Schnell(l984), Jamesand Neal (1986) Toups and Jackson(1987) Root (1988) Ridgely and Gwynne (1989), Stiles et al. (1989) Taylor et al. (1989) and Robinson (1990). 218 CHRISTOPHER W. THOMPSON

sample size 1 14 21 19 37 21 35 30 31 25 25 20 25 19 100 g 90 lJUVENAL PLUMAGE g 60 =FIRST PREBASIC g 70 MOLT & 60 mFIRST BASIC cr, 50 PLUMAGE O5 3040 OPRESUPPLEMENTAL MOLT 2 20 IXISUPPLEMENTAL E 10 PLUMAGE 0 =PREALTERNATE JUN JUL AUG SEP OCT NOV DEC MOLT DATE FIGURE 2. Timing of molts and plumagesof first-yearPainted Buntings (males and femalescombined). molt from early June through mid-October (Fig. range of P, ciris.This indicates that most indi- 2, Table 3). viduals complete this molt within the breeding To compare the timing of this molt between range. However, ten of 20 first-year birds col- sexes, first I compared the temporal frequency lected in exclusively migratory areas were in first distributions of first-year male and female spec- prebasic molt, indicating that this molt often is imens collected during the molting period from completed during migration. No specimenswere June through October. These frequency distri- collected on the wintering ground in this molt. butions did not differ significantly (Gad,= 4.399, No specimen in first basic plumage (21 males, df = 8, P > 0.75; see Table 3) and, therefore, 25 females, six sex unknown) exhibited any blue allowed comparison of the frequency distribu- or red body plumage. First basic plumage in both tions of first-year male and female specimensin sexesis significantly brighter olive-green dorsally molt across months of the molting period. The and yellow-green ventrally than is juvenal plum- timing of molt did not differ significantly be- age. This plumage is similar to definitive female tween the sexes(G,, = 2.939, df = 8, P > 0.95; plumage, but is considerably more cryptic. Table 3). Location.Most specimens (88 of 98) in first PRESUPPLEMENTAL MOLT prebasic molt were collected on the breeding Sequenceand extent. This molt is exhibited only ground. In addition, 48 of 52 specimens in first by first-year males and females. All previous in- basic plumage were collected within the breeding vestigators(e.g., Dwight 1900, Storer 195 1, Fisk 1974, Oberholser 1974) have mistaken this molt TABLE 3. Temporal frequency distribution of first- year male and female Painted Buntingsin first prebasic for first prebasic molt. All body plumage, all rec- molt. u-ices, and typically the outer four or five pri- maries and inner four to six secondaries(Table Male Female 4a) are replaced. In about 18% of males and 25% speci- speci- mens Male mens Female of females, one to five outer greater primary co- exam- specimens exam- v$$y ined in molt ined verts also are replaced (Table 4b) at the same Date (fo % bo (a) % bo time as their corresponding primaries. Replace- June 1-15 1 100.0(l) 0 0.0 (0) ment of the four most proximal greater primary June 16-30 10 90.0 (9) 4 50.0 (2) coverts was never observed (Table 4b). There July 1-15 10 100.0 (10) 7 100.0 (7) was no difference between sexesin the frequency July 16-31 10 60.0 (6) 2; 77.8 (7) of replacing one or more greater primary coverts Aug. 1-15 15 86.7 (13) 65.0 (13) Aug. 16-31 10 50.0 (5) 1: 75.0 (6) compared to replacing none (Gad,= 3.12 1, df = Sept. 1-15 11 54.5 (6) 47.4 (9) 1, P > 0.05) or in the frequency distribution of Sept. 16-30 14 14.3 (2) 15 66.7 (1) their replacement (G,, = 7.14 1, df = 5, P > 0.1; Oct. l-15 19 5.3 (1) 10 0.0 (0) see Table 4b). The sequence of body molt ap- DELAYED PLUMAGE MATURATION 219

TABLE 4a. Site of initiation of primary and second- TABLE 5. Temporal frequency distribution of first- ary molt (after S7S9 molt) by first-year male and fe- year male and female Painted Buntings in presupple- male Painted Buntings during presupplementalmolt. mental body and/or flight feather molt.

Site of PP.ICfdl Percent initia- PWWlt PWtX?nt Site of of male of female tion of ofmale of female initiation speci- speci- second- speci- speci- of p~ok$ry Inens mens (n = 397) (n = 283) St (nYh) (nz:3)

P6 80.1 76.3 S6 39.1 44.0 Sept. l-15 11 63.6 (7) 19 47.4 (9) P5 13.3 17.7 s5 51.5 48.7 Sept. 16-30 14 78.6(11) 15 73.3 (11) P4 4.3 4.6 s4 9.4 7.3 Ott 1-15 19 36.8 (7) 10 70.0 (7) P3 2.3 1.4 - - - Oct. 16-31 12 33.3 (4) 12 41.7 (5) Nov. 1-15 12 33.3 (4) 12 91.7 (11) Nov. 16-30 9 22.2 (2) 11 9.1 (1) pears to follow that describedpreviously for first prebasic molt. Primary molt is initiated most green and less brown) to supplemental remiges frequently at P6 (Table 4a) and proceedsin as- and greater primary coverts as they wear. Sup- cending order. Secondary molt begins with S7- plemental remigesare not sufficiently longer than S9, which do not appear to be replaced in any the juvenal remiges they replace to allow dis- particular order, and is subsequentlyfollowed by crimination of subadult males and females from S4, S5, and/or S6 in ascending order. Rectrices adult females basedon remex length. In addition, were molted simultaneously in most cases(13 of retained remiges become more green due to wear 14), as indicated by their similar length during than do retained upper greater primary coverts. replacement, and rarely in ascending order (1 of Thus, retained greater primary coverts are the 14). First-year birds were considered to be in best criterion for distinguishing males and fe- presupplemental molt if they either exhibited males in supplemental or first alternate plumage symmetric flight feather molt or exhibited body from females in definitive alternate or definitive molt and had replaced their greater secondary basic plumage. coverts. Timing. Presupplemental body and flight No specimens (n = 159 males, 137 females) feather molt typically occursfrom early Septem- in supplemental plumage exhibited any blue ber through late October. However, a few spec- feathers;red featherswere found in 11 (7%) males imens exhibited a small amount of body molt and one (< 1%) female. Supplemental plumage until late November (Fig. 2, Table 5). is somewhat brighter olive-green dorsally and Regression of date on primary molt score in- considerably brighter yellow-green ventrally than dicates no significant relationship between these first basic plumage. Birds in this plumage can be variables for males (n = 20) (one-tailed t-test, t distinguished from adult females by their re- = 0.0908, df = 18, P > 0.43, females (n = 19, tained juvenal inner primaries, inner greater pri- df = 17, P > 0.2), or both sexescombined (t = mary coverts, and outer secondaries. Retained 0.4379, df = 35, P > 0.25). However, regression juvenal remiges and greater primary coverts be- ofdate on total flight feather molt score(TFFMS) come considerably more similar in color (more doesindicate a significantrelationship (t = 2.5017, df = 36, P < 0.01). Comparison of Y-intercepts TABLE 4b. Frequency of replacement of juvenal (at X = 86, the minimum TFFMS in the data (brown) with green outer greater primary coverts by first-year male and female Painted Buntingsduring pre- set; seeFig. 3) and slopesof regressionsfor males supplemental molt. versus females indicates no significant difference in the date of onset (t = 0.8702, df = 37, P > “-V&W 0.2) or rate (t = 0.3499, df = 36, P > 0.5) of Percentofmale Percentof female RPhced specimens (n = 428) specimens(n = 324) molt. The average duration of molt is 47 days. Location. The location of presupplemental 0 82.5 75.0 1 3.7 6.5 flight feather molt differs dramatically between 2 5.6 5.9 eastern and western Painted Buntings. No spec- 3 6.1 11.7 imens were collected on the wintering range in 4 1.9 0.6 first basic plumage. This indicates that most, if 5 0.2 0.3 not all, Painted Buntings must initiate this molt 220 CHRISTOPHER W. THOMPSON

were collected in areasof overlap between breed- 0 NO”1. ing and wintering ranges (discussed further in Y=8.74+o.o1ox 0 discussion), and only two (10%) were collected 2 Z r”=O 14 0 on the wintering ground, probably having initi- OCTI- OO ated molt in migratory areas as discussedabove. k 2 O 8 0 This indicates that most, if not all, western Paint- 2 ed Buntings that undergo this molt in migratory d SEPT1 areas do not continue their migration to the win- tering ground until their flight feather molt is AlJG1, 1 1 1 I 0 15 30 45 60 75 90 105 1ao135 150 complete. TOTALFLIGHT FEATHER MOLT SCORE PREALTERNATE MOLT FIGURE 3. Regressionof collection date againstto- tal flight feather-molt score during presupplemental Sequenceand extent. After completing this study, molt for subadults(males and females combined). In I found that this molt had been recognized pre- the regressionmodel, Y is expressedas a decimal form of month, e.g., for the date 15 October, Y = 10.50. viously by Fisk (1974) who stated that “both first-year [second-year]and adult birds had a par- tial spring molt, replacing the abdominal, at least before reaching the wintering ground. Seventeen some breast feathers, and at least some feathers (89%) eastern specimens in this molt were col- in the orbital region, lores, and chin. This molt lected on the breeding ground (Fig. l), indicating is rapid, the short facial and chin quills exploding that eastern Painted Buntings usually begin this in a day or two.” Fisk also recognized that sub- molt on the breeding ground. The remaining two adult males grow female-like feathersduring this (11 o/o)eastern specimens in molt were collected molt when she stated that “the fresh straw yellow on the wintering ground (Fig. l), but probably abdominal feathers of the immature [subadult initiated this molt on the breeding ground. In male or female] bird, or the adult female, are addition, seven eastern specimenswere collected partially grown one day and fully grown two days from within their breeding range in supplemental later. . . . An immature male may acquire a scat- plumage, indicating that eastern Painted Bun- tering of blue feathers in the head at this time, tings often complete this molt on the breeding but . . . the ventral parts remain yellow.” Others ground. (e.g., Storer 195 1) have noted that subadult males In contrast to eastern Painted Buntings, only may have some blue head plumage in spring but one (5%) western Painted Bunting was collected not in winter, but have attributed this change to on the breeding ground in this molt, and this wear or adventitious molt rather than a preal- specimen was collected very close to a wintering temate molt. In addition, it is quite surprising area along the Gulf of Mexico (Fig. 1). Further, that, subsequentto Fisk (1974) all investigators no western specimens were collected from their of delayed plumage maturation overlooked Fisk’s breeding range in supplemental plumage. These discovery of prealtemate molt in Painted Bun- results indicate that western Painted Buntings tings, and the growth of female-like feathers by rarely begin, much lesscomplete, presupplemen- subadult males during this molt. tal molt on their breeding ground. An additional My data completely corroborate Fisk’s results eight (40%) specimens in molt were collected in that this molt occurs in all age and sex classes. exclusively migratory areas of southeastern Ar- Except blue and, rarely, red plumage in some izona, Sonora, and northern Sinaloa, indicating subadults and adult females (Table 2, Fig. 4) all that nearly half of all western Painted Buntings alternate plumage in all age and sex classesis undergo at least part of this molt in exclusively identical in color to the plumage that preceded migratory areas. In addition, western specimens it. The alternate plumages of adult females and were collected in these migratory areas in first subadult males and females are essentially iden- basic plumage (n = 4) and presupplemental body tical (except for some or all greater primary co- molt (n = 6) prior to beginning flight feather verts) being bright olive-green dorsally and yel- molt. This shows that many western Painted low-green ventrally. However, a greater Buntings do not begin this flight feather molt proportion of subadult males exhibit limited blue until they reach these migratory areas. Of the head plumage than either subadult females (G,, remaining specimens in this molt, nine (45%) = 33.492, df = 1, P < 0.001) or adult females DELAYED PLUMAGE MATURATION 221

I SUBADULT MALES

&&I SUBADULT FEMALES

A-A SUBADULTS

_ JAN FEB h4AR APR MAY JUN JAN FEB MAR AFR’ MAY

MONTH MONTH OF YEAR FIGURE 4. Percentageof Painted Buntingswith any FIGURE 5. Percentageof head plumagein molt dur- blue plumage by age-sex class,excluding adult males, ing prealternate molt in Painted Buntings by age and from January through June. sex class.Values represent means f 1 SE.

(G,, = 30.405, df = 1, P < 0.001; Fig. 4). In- on areas of the body other than the head was coming alternate head plumage of subadult males observed in only two subadult males, one sub- was green in 14 specimens (AMNH 60063, adult female, three adult females, and one adult 785749,365308; DMNH 48703,48707; FMNH male. However, examination of museum skins 22930; LSU: 49448, 17629; MCZ 310689, underestimates the proportion of birds molting 261372; MLOCAL 49405, 26451; UMINN feathers on parts of the body other than the head 29875,29869), blue andgreen in three specimens becausemolting feathersare more difficult to see (FMNH 22932; LSU 99884; MLOCAL 25945) on the body than on the head of museum skins and only blue in one specimen (MCZ 26 137 1). as compared to live birds. This is illustrated by Total body molt score(TBM) was used to eval- four lines of evidence. uate molt intensity among age and sex classes First, museum skin labels often indicated pres- (Fig. 5). Molt intensity was significantly higher ence of prealtemate body molt that I could not in adult males than females (n = 697, Mann- find, e.g., the label ofan adult male (LSU 113340) Whitney U test, U = 35,677, P = 0.02 l), but did stated “moderate overall body molt,” but I could not differ between subadult males and females not find any. (n = 224, U = 6,350, P = 0.557). In addition, Second, many areas on the body of museum molt intensity was significantly higher in adult skins can not be examined becausethey are cov- males than subadult males and females (n = 786, ered by the wings and legs of the specimens. U = 59,216, P = 0.007) but did not differ be- Third, among specimens of Painted Buntings tween adult females and subadult males and fe- in prealtemate molt, Fisk (unpub. data) found males (n = 359, U = 15,111, P = 0.998). molting ventral and dorsal feathers in a much It is not possible to distinguish between com- higher proportion of live specimens (15 or 6 1) pletely grown (unsheathed) alternate body feath- than I did in museum skins (7 of 87) (G-test, G,, ers and feathers of the preceding plumage of the = 3.93 1, df = 1, P < 0.05). This is a valid com- same color. Therefore, on museum specimens, parison for the following reasons. First, there is it also is not possible to determine either the no significant difference among the frequency average extent of this molt among age and sex distributions of subadult and adult male and fe- classesor the proportion of molt completed in male specimens examined in this study for the individuals. However, it is probable that ex- period from December through May (G,, = amination of museum skins underestimates the 1.007, df = 15, P > 0.995; Tables 6a, b) or in proportion of birds molting feathers on parts of Fisk’s study for the period from January through the body other than the head. April (G,, = 1.470, df = 6, P > 0.95; Table 7). Collection of molt data from live versus mu- This indicates that no age or sex class is dispro- seum specimens. Based on my examination of portionately representedin the data in any month museum skins, molt appearsto be limited to the within either study. Second, the proportion of head in most cases(85 of 92 specimens). Molt specimensin molt doesnot differ among age and 222 CHRISTOPHER W. THOMPSON

TABLE 6a. Temporal frequencydistribution of first prealtematemolt observedin museumspecimens of subadultmale and femalePainted Buntings.

MaIfS FCXtUkS examined Males in head Males in body’ examined Femalesin head Femalesin body’ Month (n) molt % (n) molt % (n) vo molt % (n) molt % (n) Dec. 17 17.6 (3) 0.0 (0) 23 8.7 (2) 0.0 (0) Jan. 14 0.0 (0) 0.0 (0) 35 8.6 (3) 0.0 (0) Feb. 26 11.5 (3) 0.0 (0) 32 3.1 (1) 0.0 (0) Mar. 36 19.4 (7) 2.8 (1) 32 21.9 (7) 3.1 (1) Apr. 63 9.5 (6) 0.0 (0) 38 0.0 (0) 0.0 (0) May 129 0 (0) 0.8 (1) 67 0.0 (0) 0.0 (0) ’ ’ Thesespecimens may alsoexhibit head molt. sex classesin this study during the period from geststhat this molt is more extensive in all age December through May (G,, = 3.436, df = 3, P and sex classesthan is indicated by my analysis > 0.25; Tables 6a, b) or in Fisk’s study from of museum specimens. January through April (G,, = 1.805, df = 2, P Timing. First prealternate molt begins in mid- > 0.25; Table 7). This indicates that the fre- December, only shortly after completion of pre- quency of molt doesnot differ significantly among supplemental molt in late November, and is age and sex classesand, therefore, allows age and completed by mid-May (Fig. 5, Tables 6a and sex classesto be combined for further analysis. 7). Timing of this molt is also reflected by the Third, there is no significant difference between development of blue head plumage in subadults Fisk’s study (unpub. data) and this study in the (Fig. 4). Similarly, definitive prealtemate molt frequency distribution of specimens(age and sex occursfrom early January through late May (Fig. classescombined) from January through April 5, Tables 6b and 7). At the population level, (G,, = 6.596, df = 3, P > 0.05; Tables 6a and presupplemental and definitive prebasic molts 6b combined versus Table 7). nearly overlap in time and first and definitive Fourth, growing “pin” feathers are often lost prealtemate molts in subadults and adults, re- from specimens during specimen preparation. spectively. Fisk usually observed only a small number of That prealtemate molt in subadults and adults growing feathers in individual birds at any one is a functionally separateevent from the fall molt time during this molt, although she estimated that preceded it, and not a continuation of it, is individuals rarely had as much as 80% of their strongly suggestedby three points. First, in all body feathers in molt. In addition, Fisk’s data age and sex classes,the percentage of birds in indicate that molt occursin subadults and adults prealtemate molt increasesfrom late December in most, if not all, feather tracts of the head and through February and March, and decreases body including greater and lessertail coverts and through April and May (Tables 6a, b and 7). median and lesserwing coverts, and that molt is Second, specimens in presupplemental and de- extensive in at least some individuals. This sug- finitive prebasic molt complete head molt usu-

TABLE 6b. Temporalfrequency distribution of definitive prealtematemolt observedin museumspecimens of adult male and femalePainted Buntings.

M&S FelMleS examined Males in head Males in body’ examined Femalesin head Femalesin body> Month (4 molt % (n) molt % (n) (n) molt % (n) molt % (n) Dec. 50 0.0 (0) 0.0 (0) 22 0.0 (0) 0.0 (0) Jan. 84 8.3 (7) 0.0 (0) 25 4.0 (1) 4.0 (1) Feb. 88 18.2 (16) 0.0 (0) 29 20.7 (6) 3.4 (1) Mar. 103 13.6 (14) 0.0 (0) 21 19.0 (4) 0.0 (0) Apr. 287 0.0 (0) 0.3 (1) 60 5.0 (3) 3.3 (2) May 307 0.3 (1) 0.0 (0) 74 0.0 (0) 0.0 (0) I Thesespecimens may alsoexhibit head molt. DELAYED PLUMAGE MATURATION 223

TABLE 7. Temporal frequency distribution of prealtemate molt observed in adult male and female live specimensof Painted Buntings (Fisk, unpublisheddata).

SY SY SY &3+99 MfPP M+PP ASYdd ASYM ASYdd ASY 99 ASY99 ASYOn examined in headmolt in body’ molt examined in headmolt in body’ molt examined in headmolt in body’ molt MOlltb 00 % (4 % (n) PO % (4 % 00 @o % @I % Qo Jan. 20 0.0 (0) 0.0 (0) 12 0.0 (0) 0.0 (0) 5 0.0 (0) 0.0 (0) Feb. 31 19.3 (6) 0.0 (0) :: 44.4 (8) 0.0 (0) 15 46.7 (7) 0.0 (0) Mar. 42 14.3 (6) 14.3 (6) 19.2 (5) 3.8 (1) 27.8 (5) 0.0 (0) Apr. 54 3.7 (2) 7.4 (4) 36 11.1 (4) 0.0 (0) :i 9.7 (3) 12.9 (4) ’ Thesespecimens may alsoexhibit head molt. ally by late September and always by mid-Oc- versus adults in molt during each month of the tober. Third, adult females and subadult males molting period. No significant difference in the and females frequently grow some blue and red frequency of molt between age classeswas de- feathers during this molt (Fig. 4), but such feath- tected in any month (Gad,5 3.434, df = 1, P > ers are never grown during presupplemental or 0.05; seeTables 6a, b), indicating that the timing definitive prebasic molt. of this molt does not differ significantly between To compare the timing of this molt among age age classes. and sex classes,first I compared the frequency An identical analysis of Fisk’s data (unpub.) distributions of male and female specimenswith- yielded similar results. The frequency distribu- in ageclasses from December throughMay. These tions of male and female specimens within age distributions differed significantly between sub- classesfrom January through April did not differ adult males and subadult females (Gadj= 15.48 1, significantly between adult males and females df = 5, P < 0.01; see Table 6a), but not between (G,,, = 0.909, df = 3, P > 0.75; see Table 7). adult males and adult females (G,, = 4.602, df Fisk did not sex subadults in her study because = 5, P > 0.25; see Table 6b). Therefore, com- it is not possible to do so except by laparotomy parison of the timing of this molt in males and or observing sex-specific behaviors. Therefore, females was done by comparing frequencies of the frequencydistributions of subadult males and 1) subadult males and subadult females in molt females could not be compared. Adult males and during each month of the molting period, and 2) females did not differ significantly in the timing adult males and adult females in molt acrossall of this molt (G,, = 0.453, df = 3, P > 0.9; see months of the molting period. No significant dif- Table 7). In addition, the frequency distributions ferences in the frequency of molt were found of subadult and adult specimens did not differ between subadult males and subadult females in significantly (Gad,= 0.563, df = 3, P > 0.9) and any month (Gadj5 1.293, df = 1, P > 0.25), or no significant difference in the timing of molt between adult males and adult females from De- between age classeswas detected (G,, = 1.400, cember through May (Gadj= 3.618, df = 5, P > df = 3, P > 0.25; see Table 7). 0.5). This indicates that the timing of this molt Location. Ninety specimensin first and defin- does not differ significantly between sexeswithin itive prealternate molt were collected on the win- age classes.Therefore, frequencies of male and tering ground as compared to only two on the female specimens were combined within age breeding ground. This indicates that this molt is classesin each month for further analysis. To usually completed on the wintering ground, but compare the timing of this molt between age rarely is continued on the breeding ground. This classes,I compared the frequency distributions is further supported by the 1) absence of preal- from December through May of subadult and temate molt in live specimens(adults: 19 males, adult specimens.These distributions differed sig- 8 females; subadults: 13 males, 9 females) mist- nificantly between age classes(G, = 11.385, df netted on the breeding ground in Austin, Texas = 5, P < 0.05). Therefore, comparison of the during May, 1984, and 2) lack of changein plum- timing of this molt in subadults versus adults age color in subadult males captured in April or was done by comparing frequencies of subadults May and recaptured later in the breeding season 224 CHRISTOPHER W. THOMPSON

NO” 1 quence: all primary and secondarycoverts, head 0

Y-6.18+0.0196 X 0 (including chin, throat, and nape) and lower back, and underside. Flight feather molt typically be- gins with Pl, followed shortly thereafter by the secondariesand rectrices. All primaries and S l- S6 molt in ascending order. S7-S9 molt first, followed by SlS6. Rectrices appear to be re- placed synchronously in most cases (21 of 28 specimens), and in ascending order in the rest (n 1 O 1 1 I I 0 10 20 30 40 50 60 70 80 90 = 7). PRIMARY MOLT SCORE Contrary to many earlier sources(e.g., Dwight FIGURE 6. Regressionof collectiondate against pri- 1900, Sprunt 1968, Oberholser 1974), the pres- mary molt scoreduring definitive prebasicmolt in ence of female-like olive-green body or flight Painted Buntingsfor adult malesand females(com- feathersin adult males is not an indication of an bined). In the regressionmodel, Y is expressedas‘ a decimalform of month, e.g.,for the date 15 October, incomplete prebasic molt by subadult males, but Y = 10.50. rather indicates adventitious feather loss and re- placement, i.e., replacement of feathers lost and at the same locality in Austin, Texas (n = 17) replaced at a time other than the normal molting and in coastal (T. A. Beckett III, period (Storer 1951, Fisk 1974). S7S9 and Pl- pers. comm.). P3 (occasionally 4 and 5) of adult males are usu- ally partly or totally bright green, being similar DEFINITIVE PREBASICMOLT in color to their bright green dorsal body plum- Sequenceand extent. This is a complete and very age, and thus are easily distinguished from ad- intense molt exhibited by second-year and older ventitiously replaced olive-green flight feathers. males and females. The intensity of this molt Timing. Adult prebasic molt typically occurs renders molting birds nearly flightlessduring the from early August through early October (Fig. intermediate stagesof this molt, thereby causing 7) but some specimensexhibited body molt until them to become very secretive at this time mid-November. Linear regressionof date on pri- (Thompson, unpub. data; T. A. Beckett III, pers. mary molt scoreindicates that duration of flight comm.). This is also illustrated by the relative feather molt is approximately 54 days (Fig. 6). lack of adult specimensin intermediate stagesof Comparison of Y-intercepts and slopes of re- molt (Fig. 6). Body molt begins on the upper back gressionsof date on primary molt scorefor males and progressesin the following overlapping se- and females indicates that flight feather molt does

sample size 69 44 62 21 13 59 57 27 27 20 100 - g 90 m ALTERNATE 3 60 PLUMAGE E; 70 m PREBASIC MOLT ; 60 0 BASIC !% 60 PLUMAGE O 40 g 30 2 20

- JUL AUG SEP OCT NOV DATE

FIGURE 7. Timing of molts and plumagesof second-yearand older male and female Painted Buntingsfrom July through November. DELAYED PLUMAGE MATURATION 225

not differ between the sexesin its rate (t-test, t = lo), prior to beginning flight feather molt. This = 0.8998, df = 93, P > 0.2) or date of onset (t shows that many western Painted Buntings do = 1.621, df = 94, P > 0.1). not begin this flight feather molt until they reach Location.Like presupplemental flight feather these migratory areas. In addition, three speci- molt, the location of definitive prebasic flight mens were collected in migratory areas in fresh feather molt differs dramatically between eastern definitive basic plumage, indicating that at least and westernPainted Buntings. No specimenswere some individuals complete molt before contin- collected on the wintering range in first or defin- uing their migration to the wintering ground. Of itive alternate plumage. This indicates that most, the remaining specimens in flight feather molt, ifnot all, Painted Buntings must initiate this molt 16 (26.7%) were collected in areas of overlap before reachingthe wintering ground. Forty (93%) between breeding and wintering ranges (dis- eastern specimensin this molt were collected on cussed further in discussion), and only seven the breeding ground (Fig. l), indicating that east- (11.7%) were collected on the wintering ground, em Painted Buntings usually begin this molt on probably having initiated molt in migratory ar- the breeding ground. The remaining three (7%) eas as discussedabove. This further suggeststhat eastern specimens in molt were collected on the most, if not all, western Painted Buntings that wintering ground (Fig. l), but probably initiated undergo this molt in migratory areas do not con- this molt on the breeding ground. This conclu- tinue their migration to the wintering ground un- sion is corroborated by Fisk (1974, unpub. data, til their flight feather molt is complete. pers. comm.) who studied Painted Buntings on the wintering ground in southern Florida. She COMPARISON OF MOLTS AND never captured any adults in first or definitive PLUMAGES BETWEEN INDIGO AND PAINTED BUNTINGS alternate -plumage -- prior to definitive nrebasic molt, but did capture 11 adults “in [adult pre- This comparison indicates that Painted and In- basic] molt between 25 September and 27 Oc- digo Buntings exhibit very similar seasonaltim- tober.” In addition, 13 eastern specimens were ing and extent of all molts in all age and sex collected from within their breeding range in de- classes.However, most molts and plumages dif- finitive basic plumage, indicating that eastern fer dramatically between Painted and Indigo Painted Buntings often complete this molt on the Buntings in two ways: 1) whether molt occurson breeding ground. This conclusion is corroborat- the wintering, migratory, or breeding ground, and ed by Tipton and Tipton (1978) who banded two 2) the seasonal change in male plumage color second-year males in first alternate plumage in and conspicuousnessresulting from molt. South Carolina on 21 and 23 July 1977. Both birds were recaptured at the same locality on 2 1 FIRST PREBASIC MOLT September 1977, both having completed defin- Juvenile male and female Painted Buntings begin itive prebasic molt “except for waxy sheathson first prebasic molt of most to all body feathers primaries eight and nine.” (except primary and secondarycoverts) within a In contrast to eastern Painted Buntings, within week of fledging that results in a first basic plum- the breeding range of western Painted Buntings, age that is brighter green dorsally and yellow only one (1.7%) western Painted Bunting was ventrally. In both males and females, first basic collected in this molt (Fig. I), and only one spec- plumage is more similar to definitive female imen was collected in definitive basic plumage. plumage than is the previous juvenal plumage. These results indicate that western Painted Bun- In Indigo Buntings, the age of onset and seasonal tings rarely begin, much less complete, this molt timing of this molt is very similar to that of on their breeding ground. An additional 36 (60%) Painted Buntings, and first basic plumage in both specimens in molt were collected in exclusively males and females is also more similar to female migratory areasof southeasternArizona, Sonora, definitive basic plumage than is the previous ju- and northern Sinaloa, indicating that most west- venal plumage (Rohwer 1986). The existence of em Painted Buntings undergo at least part of this this molt raises an obvious question. Why do molt in exclusively migratory areas. In addition, fledgling Painted and Indigo Buntings undergo western specimens were collected in these mi- this nearly complete body molt that does not gratory areasin first and definitive alternate plum- result in a substantial change in plumage color age (n = 9) and definitive prebasic body molt (n and subsequentlyundergo a secondcomplete molt 226 CHRISTOPHER W. THOMPSON only one to three months later? Juvenal plumage adults which typically replace S7S9 (Rohwer in both speciesis more compressibleand has less 1986). Adult males and females of both species structural integrity (lessinterlocking barbules and grow adult male- and female-like plumage, re- lower density of barbs on feathers) than subse- spectively, during this molt. quent plumages(Thompson, unpub. data). Thus, Wingfield (pers. comm.) concluded from a re- juvenal plumage probably facilitates heat trans- view of the literature that physiological control fer during brooding, but is poorly adapted to of plumage color typically is independent of sex protecting juveniles from adverse abiotic factors steroids in speciessuch as Painted Buntings that such as wind, rain, incident solar radiation, and do not change plumage color seasonallyafter re- physical abrasion. Selection may have favored productive maturity, but is dependent on sex evolution ofjuvenal plumage that facilitates heat steroids in species such as Indigo Buntings transfer from the brooding female to young in that do change plumage color seasonally after the nest, and replacement of this plumage as soon reproductive maturity. Testis size of subadult as possible after fledging with a structurally male Painted Buntings does not differ signifi- strongerplumage that offersmore protection from cantly from that of adult males from March the physical environment (Webb et al. 1990). through June (Thompson, unpub. data), and go- nadal hypertrophy is significantly advanced in PRESUPPLEMENTAL MOLT both subadults and adults by April. This suggests Within three months after completing first pre- that circulating plasma sex steroid (testosterone) basic molt, and usually before arrival on the win- levels are elevated in subadult male Painted tering ground, first-year male and female Painted Buntings when they grow green (female-like) Buntings subsequentlyundergo a presupplemen- feathers during first prealtemate molt, and, tal molt of all body plumage, all rectrices, the therefore, supports Wingfield’s conclusion that inner four to six secondaries,and the outer four physiological control of plumage color typically or five primaries. The resulting supplemental is independent of sex steroids in speciesthat do plumage is significantly brighter green dorsally not change plumage color seasonallyafter repro- and yellow ventrally than first basic plumage, ductive maturity. and is essentially identical in color to definitive However, Lowery (1955:58) states that “the female plumage. The extent and seasonaltiming brilliant livery of the male Painted Bunting is of this molt is very similar in Indigo Buntings. brought on by secretionsof the mature male go- However, this molt differs between Painted and nads which do not become activated until the Indigo Buntings in two striking ways. First, about spring following birth. Before this the male is 90% of Painted Buntings complete this molt be- attired like the female. Experimentally, in the fore arrival on the wintering ground whereas at laboratory, adult males can be made to assume least 90% of Indigo Buntings do not begin this the color of the female by injection of female molt until they reach the wintering ground (Roh- hormones. Conversely, the female can be made wer 1986). Second, supplemental plumage re- to take on the male attire by injection of male sulting from this molt is essentially identical to hormones.” Unfortunately, Lowery (195 5) gives female definitive basic plumage in subadult male no additional details or citations regarding these Painted Buntings, but is essentially identical to experiments. A comparison of circulating plasma male definitive basic plumage in subadult male sex steroid levels between subadult and adult Indigo Buntings (Rohwer 1986). males in Painted and Indigo Buntings would be a valuable contribution to our knowledge of the PREALTERNATE MOLT hormonal control of plumage color. In Painted Buntings, this molt 1) occurs almost exclusively on the wintering ground, 2) results in ADULT PREBASIC MOLT growth of adult female-like plumage by subadult Most Painted and Indigo Buntings complete this males, and 3) is similar in extent among age and molt before arrival on the wintering ground.This sex classes.In contrast, in Indigo Buntings this molt does not cause any seasonal change in molt 1) is continued on the breeding ground in plumage color in adult females of either species. about 17% of specimens, 2) results in growth of In males, however, this molt replaces conspic- exclusively adult male-like plumage by subadult uous breeding plumage with an identical con- males, and 3) is less extensive in subadults than spicuous winter plumage in Painted Buntings, DELAYED PLUMAGE MATURATION 227

but with a much more cryptic and female-like or definitive prebasic] molt before continuing on winter plumage in Indigo Buntings. to the wintering ground. This study documents This comparison of the molts and plumages molt-migration by first-year and adult Painted of Painted and Indigo Buntings illustrates that Buntings, and illustrates that the pattern of molt- both speciesexhibit the same sequenceof molts migration in this speciesis very similar to that during similar months of the year, but that the exhibited by Northern Orioles and Lazuli Bunt- location (breeding, migratory, or wintering area) ings. of molt and the seasonal changes in plumage Northern Orioles are classified into two geo- color and conspicuousnessdiffer dramatically. graphically contiguous subspecies: Baltimore Lazuli Buntings, Passerina amoena, exhibit the Orioles (I. g. galbula) and Bullock’s Orioles (I. same set of similarities and differences in their g. bullockii) which breed extensively throughout sequenceof molts and plumages(Young, in press), southern Canada and most of the continental thus further corroborating this conclusion. Al- United States east and west of central Texas, though no strongconclusions can be inferred from respectively. Lazuli Buntings have an extensive these comparisons, the apparent pattern of sim- breeding range in the midwestem and western ilarities and differences among these three spe- United States. Drought conditions often prevail cies of Passerina suggeststhat the extent, timing in the midwestem and western United States in with respect to breeding and migration, and lo- late summer and fall. At this time, Lazuli Bun- cation of molt, aswell as the seasonalcolor change tings (Young, in press) and Bullock’s Orioles caused by molt, are less phylogenetically con- (Rohwer and Manning 1990) usually migrate to strained than are the sequenceand calendar tim- and undergo a complete molt in desert areas of ing of molt. the southwestern United States and northwest- em Mexico that are outside their breeding and IMPLICATIONS FOR MOLT-MIGRATION wintering ranges. Conversely, Baltimore Orioles IN PASSERINES usually complete this molt prior to fall migration. In most birds, flight feather molt typically occurs Rohwer and Manning (1990) and Young (in press) on the breeding ground prior to migration or, argue that these desert locations exhibit a large more rarely, on the wintering ground after mi- increase in plant and insect life in response to gration. Alternatively, flight feather molt may “monsoon” rains that occur predictably in these begin on the breeding ground, be suspendeddur- areas in late summer and fall. In turn, they argue ing migration, and be completed on the wintering that selection has favored evolution of molt-mi- ground. Molt of flight feathersduring migration, gration strategiesin Bullock’s Orioles and Lazuli called molt-migration, is rare. It has been doc- Buntings as a mechanism to allow these species umented previously in many species of water- to molt in an area with greater food resources fowl (Herter et al. 1989 and refs. cited therein), than exist at the same time of year on their breed- and raptors (Thiollay 1978) and in some shore- ing or wintering ground. In addition, most Bal- birds, grebes, flamingos, cranes, and auk (Jehl timore Orioles, but not Bullock’s Orioles, prob- 1990) but in no other nonpasserines and only ably migrate acrossthe Gulf of Mexico en route five species of passerines: Western Kingbirds, to their wintering ground in Central and South Tyrannus verticalis(Phillips et al. 1964) Swain- America. Rohwer and Manning (1990) arguethat, son’s Thrushes,Catharus ustulatus (Cherry 1985), as a consequence, selection has favored Balti- Northern Orioles (Rohwer and Manning 1990) more Orioles, but not Bullock’s Orioles, to max- Lazuli Buntings (Young, in press), and Yellow- imize their flight efficiency during fall migration breasted Buntings, Emberiza aureola (Strese- by completing flight feather molt prior to fall mann and Stresemann 1966, cited in Young, in migration. press). The occurrence of molt-migration in Except in coastal areas, western Painted Bun- western Painted Buntings was previously rec- tings inhabit areas that often experience drought ognized by Phillips et al. (1964) who stated that, and food shortagein late summer and fall. Con- prior to “prebasic” [presupplemental or defini- versely, eastern Painted Buntings inhabit coastal tive prebasic] molt, Painted Buntings migrate areas that rarely experience drought and food from more easterly breeding areas to southeast- shortage in late summer and fall. In addition, em Arizona and northern Sonora where they be- like Northern Orioles, most eastern and western gin and complete “prebasic” [presupplemental Painted Buntings migrate acrossand around the 228 CHRISTOPHER W. THOMPSON

Gulf of Mexico, respectively, to their wintering ing ability as well. Thus, the winter status sig- ground in (Storer 195 1, Bird naling hypothesispredicts that the extent of adult Banding Lab, unpub. data). Figure 1 illustrates male-like winter plumage acquired by subadult that presupplementaland definitive prebasicflight males as a result of presupplemental molt should feather molt occurs on the breeding ground in be correlated with fighting ability and, thus, the eastern population in about 90% of cases,but should vary considerably among subadult males. in the western population in lessthat 5% of cases. In addition, the more conspicuousthat subadult About 50% of the western population undergo male winter plumage is, on average, relative to flight feather molt in exclusively migratory areas the preceding fall plumage, the stronger is the of southeastern Arizona, Sonora, and northern support for the winter status signaling hypothe- Sinaloa. An additional 20-25% of western spec- sis. imens in flight feather molt were collected in the Within three months after completing first pre- Rio Grande Valley of Texas where the breeding basic molt, and usually before arrival on the win- and wintering ranges of Painted Buntings over- tering ground, first-year male and female Painted lap. This area, like the desert southwest, expe- Buntings undergo a presupplemental molt of all riences similar “monsoon” conditions in late body feathers and many flight feathers. The re- summer that cause large increases in seedsand sulting supplemental plumage in all subadult . Thus, this area probably serves as an males is essentially identical in color to definitive alternate staging ground for flight feather molt. female plumage, and is similar to but more con- Only about 10% of specimens from the eastern spicuous than the previous first basic plumage. and western populations were found in flight This is inconsistent with the winter status sig- feather molt on the wintering ground. The sim- naling hypothesis. ilarities in the patterns of molt-migration among The winter cryptic hypothesis(Ewald and Roh- Painted Buntings, Lazuli Buntings, and Northern wer 1980, Rohwer et al. 1983, Rohwer 1986, Orioles suggeststhat these patterns evolved in Rohwer and Butcher 1988) predicts that sub- response to similar selection pressuresin these adult male winter plumage should be equally or species. more cryptic than the preceding fall plumage. Supplemental plumage is significantly more con- IMPLICATIONS FOR THEORIES OF spicuousthan the first prebasicplumage that pre- DELAYED PLUMAGE MATURATION ceded it. Thus, the data are inconsistent with this The implications of seasonalchanges in plumage hypothesis. color in subadult male Painted Buntings for win- The winterfemale mimicry hypothesis(Brown ter and summer hypothesesare summarized in and Brown 1988) predicts that subadult male Table 1 and explained below. winter plumage should be equally or more adult female-like than the preceding fall plumage. The WINTER HYPOTHESES data strongly support this hypothesis because All three winter hypothesespredict that plumage subadult males replace first prebasic plumage of subadult males should be lessconspicuous than during presupplemental molt with a supplemen- that of adult males in winter and assumethat the tal plumage that is identical in color to definitive benefits of such plumage in winter outweigh the female plumage (except for retained greater pri- costs of such plumage in summer. mary coverts). The winterstatus.signalinghypothesis(Rohwer The occurrence of this presupplemental molt 1975, 1978a, 1982, 1985; Rohwer and Butcher raises an obvious question. Why do subadult 1988) predicts that winter plumage of subadult males incur the energeticand potential predation males should reliably indicate their fighting abil- costsof molt in order to replace an unworn first ity. Subadult male Painted Buntings must com- basic plumage with a supplemental plumage that pete with adult males and females in winter. is similar in structure and color? The most likely Painted Buntings are often multi-brood and fledge explanation is that by acquiring a supplemental young as early as early June (e.g., LSU 13989) plumage that is indistinguishable from definitive and as late as early September (e.g., Parmelee female basic plumage, first-year males may 1) 1959, Sutton 1967). As a result, during their first suffer less predation than adult males, and/or 2) fall and winter, subadult males vary greatly in be mistakenly identified as adult females and relative age and, therefore, presumably in fight- thereby incur less aggressionfrom adults. DELAYED PLUMAGE MATURATION 229

As discussed above, subadult male Indigo summer status signaling hypothesispredicts that Buntings undergo a presupplemental molt, usu- the extent of adult male-like summer plumage ally on the wintering ground, resulting in a winter exhibited by subadult males during their first po- supplemental plumage that is adult male-like in tential breeding seasonshould be correlated with color. As a result, Rohwer (1986) speculatedthat male fighting ability and, therefore, should vary subadult male Indigo Buntings achieve sufficient considerably among subadult males. In addition, experience after a month or two on the wintering the greater that the extent is, on average, of adult ground to be able to dominate adult females and male-like plumage worn by subadult males dur- compete with adult males. As a result, selection ing their first potential breeding seasonrelative may have favored the evolution of this molt as to the winter plumage that precededit, the stron- a mechanism for subadult males to acquire a new ger is the support for this hypothesis. Replace- honest statussignal which reflectstheir enhanced ment of winter plumage with an equally or more competitive ability. In contrast, I suggestthat the cryptic summer plumage by subadult males would average competitive ability attained by subadult be inconsistent with this hypothesis. male Painted Buntings on the wintering ground Subadult males 1) grow plumage during first is inferior to that of adult males and females. prealtemate molt that is almost exclusively adult Thus, selection may have favored subadult males female-like in color, and 2) exhibit remarkably to deceitfully mimic adult females by growing little variation in plumage color among individ- supplemental plumage that is adult female-like uals. This is the first speciesdocumented in which in color rather than to honestly signal subordi- sexually mature males usually grow adult female- nance to adults by growing supplemental plum- like rather than male-like plumage during a age that is juvenile-like in color. In addition, if prealtemate molt. However, I recently have found the energetic cost of molting is lesson the breed- that House Finches, Carpodacus mexicanus, in ing ground or during migration than on the win- central Arizona also undergo a limited prealter- tering ground (e.g., due to greater food avail- nate molt in which both second-year and older ability), this may explain why presupplemental males replace adult male-like (red) feathers only molt in subadults occursbefore reaching the win- with female-like (brown, rarely yellow) feathers tering ground in Painted Buntings but after (Thompson, unpub. data). This finding in Paint- reaching the wintering ground in Indigo Bunt- ed Buntings is inconsistent with the summer sta- ings. tus signaling hypothesis. However, the strength If second-year males replace their first winter of this conclusion is weakened because spring plumage with a lessconspicuous summer plum- molt is limited in extent. This result does not age, this would support the winter status signal- necessarilyimply that second-yearmales achieve ing hypothesis and be inconsistent with the win- greater lifetime inclusive fitness by wearing an ter cryptic and female mimicry hypotheses. adult female-like plumage in summer than they Replacement of winter plumage with an equally would if they wore an adult male-like plumage or more conspicuoussummer plumage would be in summer. Instead, it may indicate that selection consistent with all winter hypotheses.Since sec- favors them to wear an adult female-like plum- ond-year males replace winter plumage with a age becausethey are unable to attain a complete- nearly identical and adult female-like summer ly adult male-like plumage. In summer, selection plumage during first prealtemate molt, this is may favor subadult males with plumages that consistent with all three winter hypotheses. are either relatively adult male-like or adult fe- male-like, but not intermediate between the two. SUMMER HYPOTHESES Thus, in speciesin which spring molt is limited, The summer status signaling hypothesis(Lyon selection may favor subadult males that remain and Montgomerie 1986, Montgomerie and Lyon in adult female-like plumage by growing adult 1986) predicts that subadult male plumage color female-like feathers during spring molt, rather should reliably signal fighting ability. As dis- than subadult males that attain intermediate cussedabove, Painted Buntings are often multi- plumage by growing adult male-like feathers brooded (Parmelee 1959). As a result, during their (Rohwer et al. 1980, 1983; Rohwer and Butcher first potential breeding season, subadult males 1988). vary greatly in their relative agesand, therefore, The summer cryptichypothesis (Selander 1965, presumably in fighting ability as well. Thus, the 1972; Proctor-Gray and Holmes 198 1) would be 230 CHRISTOPHER W. THOMPSON strongly supported if subadult males exhibit a winter or first potential breeding season. How- late winter or spring molt resulting in a summer ever, it would be supported if subadult males plumage that is more cryptic than their previous molt less extensively than adult males during a first winter plumage. However, as discussed late winter or spring molt. Conversely, a com- above, the relative strength of this support de- plete prebreeding body molt by subadult males pends on how extensive this molt is. The greater would be inconsistent with the hypothesis. An the increasein conspicuousnessof plumage color equally extensive but incomplete prebreeding exhibited by subadult males from winter to their body molt by both subadult and adult males first potential breeding season,the weaker is the would suggestthat such a molt is either physi- support for this hypothesis. Because subadult ologically constrained or selectively favored. males grow plumage during first prealtemate molt The frequency of prealtemate molt does not that is adult female-like and not lessconspicuous differ between subadult and adult males. These in color relative to their previous winter plum- results are consistent with but do not support the age, this is inconsistent with the summer cryptic molt constraints hypothesis. However, the in- hypothesis. tensity of prealtemate molt is greater in adult The summerfemalemimicry hypothesis (Roh- males than any other age or sex class.This result wer et al. 1980, 1983; Rohwer 1983) would be suggeststhat extent of first prealtemate molt may strongly supported if subadult males exhibit a be more energetically limited in subadult than late winter or spring molt resulting in a summer adult males and, therefore, supports the molt plumage that is as or more adult female-like in constraints hypothesis. color than their previous first winter plumage. As evidence in support of the molt constraints Becausesubadult males grow plumage during first hypothesis, Rohwer (1986) documented that prealtemate molt that is adult female-like in col- prealtemate molt by subadult male Indigo Bunt- or, this strongly supports the summer female ings often is continued on the breeding ground mimicry hypothesis. However, as discussed after spring migration. Rohwer proposesthat in above, the relative strength of this support de- spring there is less competition for food on the pends on how extensive this molt is. Becauseof breeding ground than the wintering ground be- the limitations of molt studies using museum cause food is more abundant on the breeding skins, and because this molt does not cause a than the wintering ground. As a result, first preal- discernable changein plumage color or structure, temate molt may be favored to continue on the except on the head in some subadult males, I breeding ground after spring migration. How- could not accurately estimate the extent of this ever, Painted Buntings undergo prealtemate molt molt. To do so, one would have to perform a almost exclusively on the wintering ground, thus mark-recapture study on a wintering population. not supporting the molt constraints hypothesis. The juvenilemimicry hypothesis (Lawton and If the extent of prealtemate molt is limited by Lawton 1986, Foster 1987) would be supported a constraint (e.g., food), then winter and summer only if, during their first potential breeding sea- plumages of Painted Buntings may not be func- son, second-year males wear a juvenile-like tionally independent of one another. Thus, if plumage and behave like juveniles, and would subadult plumages in males are a greater fitness be refuted if they wear an adult male-like or adult advantage in winter than a fitness disadvantage female-like plumage, or if they exhibit aggressive in summer, then supplemental plumage could be reproductive behavior. The data are inconsistent viewed as an adaptation to winter, and retention with the hypothesis becausethe plumage of sub- of adult female-like plumage in summer as mak- adult males during their first potential breeding ing the best of a bad situation (Gross 1984). For season resembles definitive female and not ju- example, when suitable breeding habitat is lim- venal or first basic plumage, and because sub- ited, and adult males outnumber reproductively adult male Painted Buntings often defend breed- mature females, subadult males are forced to ing territories and exhibit reproductive behavior compete with adult males for mates and terri- similar to adult males (Lanyon and Thompson tories. Under these or analogous circumstances 1986, Thompson, unpub. data). in winter or summer, selection may favor sub- The molt constraintshypothesis (Rohwer and adult males that are either female-like or adult Butcher 1988) makes no predictions regarding male-like in appearance, but not intermediate plumage color of subadult males during their first between the two (Rohwer et al. 1980,1983; Roh- DELAYED PLUMAGE MATURATION 231 wer and Butcher 1988). Rohwer et al. (1980) ing that season. In this way subadult male modeled the evolution of subadult male plum- plumagesmay have become adaptive during both ages and concluded that evolution of interme- winter and summer. diate plumages should be rare. Two lines of ev- The energetic cost of molt is poorly under- idence from this study support Rohwer et al’s stood. Walsberg (1983) concludes from a survey model. First, subadult males could wear a winter of the literature that molting “represents the ma- plumage that is intermediate in color between jor event of somatic production during the an- definitive male and female winter plumage and nual cycle” and requires an increase in energy wear an adult female-like plumage in summer expenditure that “significantly affect[s] the en- by replacing the adult male-like part of their ergy relationship of free-living birds.” Thus, molt winter plumage with adult female-like plumage is integrated in the annual cycle with reproduc- during first prealternate molt. Second, because tion and migration as one of the most socially subadult male Painted Buntings grow adult fe- and physiologically important and most ener- male-like rather than adult male-like plumage getically expensive events during the annual cy- during first prealtemate molt, this suggeststhat cle. Unfortunately, as Rohwer and Butcher(1988) they cannot undergo a sufficiently extensive and Rohwer and Manning (1990) recently com- prealtemate molt to achieve a completely adult mented, qualitative and quantitative knowledge male-like appearance in summer. In turn, this of the sequence,timing, location, and extent of suggeststhat selection may favor either a reduc- molts is sorely lacking and desperately needed tion in the extent of prealtemate molt and/or for the majority of birds worldwide. Such infor- growth of female-like rather than adult male-like mation about molt is essential for addressing first alternate plumage by subadult males. That many central questions in evolutionary biology subadult males could, but do not, wear a plumage and ecology regarding predation, intraspecific in winter, summer, or both that is intermediate mimicry, thermoregulation, physiological con- in color between definitive male and female straints, and paedomorphosisamong others.This plumage is the first empirical data that support study emphasizes that such information about Rohwer et al.‘s (1980) model. However, in spe- molt can be extracted from museum specimens, cies with complex social systems, such as co- and used to addressquestions regarding the evo- operative breeders, it is likely that young males lution of delayed somatic maturation, molt strat- of some species may molt into either a com- egies, and the integration of these life history pletely adult male-like plumage or a completely traits with other life history parameters. Hope- female-like plumage as suggestedby Rohwer et fully, this study will encourageothers to address al.‘s model. For example, this may occurin Lovely the evolution of delayed plumage maturation, as Fairy Wrens, Malurus amabilis, a cooperatively well as other central questions in biology, by do- breeding Australian ,in which Schodde ing similar studies on other relatively unstudied (1982) states that “some young birds seem to species. Most scientists have assumed that ab- change directly into [male] nuptial dress during sence of seasonal change in plumage color in- the post-juvenile moult (e.g., Harrison 1974) dicates absence of molt. As mentioned above, while others colour up like old females, partic- fourteen species of North American passerines ularly if they are subordinate members of a com- have subadult male plumages that differ very munal group (K. A. Muller, pers. comm.).” little between winter and summer (Table 1, cat- If subadult plumages in males are a greater egory 3 in Rohwer and Butcher 1988). Of these fitness advantage in summer than a fitness dis- species,only the Northern Oriole (Rohwer and advantage in winter, then adult female-like Manning (1990) Painted Bunting (this study), plumage could be viewed as adaptive in summer and House Finch (Thompson, unpub. data) have but potentially maladaptive in winter. In addi- been specifically investigated for the occurrence tion, selection may have favored the evolution of a spring molt. This study indicates that ab- of delayed plumage maturation during one sea- senceof changein plumage color between winter son (winter or summer) only. Retention of and summer does not always indicate an absence subadult plumage during a season when it was of molt but, rather, may indicate growth of sim- initially maladaptive may have acted as a prea- ilarly colored plumage. This suggeststhat similar daptation for the evolution of a secondary func- investigations of spring molt in the other 11 spe- tion (e.g., status signaling, female mimicry) dur- cies, especially Purple Finches (Carpodacuspur- 232 CHRISTOPHER W. THOMPSON pureus), Cassin’s Finches (C. cassinii), Great Ba- sity of Museum of SouthwesternBiology sin populations of House Finches, and Olive (2), Western New Mexico University (1) American Museum of Natural History (AMNH) (223), Buffalo Warblers (Peucedramus taeniatus) are strongly Museum of Science(2), Cornell University (33), North warranted, and likely to further advance our un- Carolina State Museum (1 l), Cincinnati Museum of derstanding of the evolution of delayed plumage Natural History (9) University of Oklahoma (UOK- maturation. LA) (87). National Museum of Natural Science (Ot- tawa: Ontario) (14) Royal Ontario Museum (45) ‘Or- ACKNOWLEDGMENTS egon State University Department of Zoology (2), Oregon State University Department of Fisheries and This paper is dedicated to the memory of Erma “Jon- Wildlife (2) Academy of Natural Sciencesof Phila- nie” Fisk who contributed extraordinarily generously delphia (54) CarnegieMuseum of Natural History (66), of her time, data, expertise, and inspiring enthusiasm Charleston Museum (25) Fort Worth Museum of Sci- to help further this project and my career, as she did ence and History (14). Midwestern State University for so many others throughout her life. Her assistance (5) Rob and Be& Welder Wildlife Foundation (S), improved this paper immeasurably. Initial stagesof Texas Tech University (3) Texas A and M University this project were conductedat the University of Texas (50) University of Dallas (17) Baylor University at Austin and were funded by University of Texas Uni- Strecker Museum (5) Dallas Museum of Natural His- versity ResearchInstitute ResearchGrants in 1984 and tory (20) University of Texas at Austin (16) Virginia 1985. Band recovery data were kindly supplied by the Polytechnic University (30) Thomas Burke Memorial Bird Banding Laboratory. T. A. Beckett III, D. Norris Washington State Museum at University of Washing- and D. F. Parmelee were very helpful and supportive ton (1 I), Milwaukee Public Museum (6) and British in many ways during the initial stagesof this project. Museum of Natural History (174). I am especiallygrateful to G. S. Butcher, G. Hill, and S. Rohwer for much helpful discussion and advice throughout this project, and for sharing unpublished LITERATURE CITED data with me regarding plumage color and delayed plumage maturation. G. S. Butcher, D. Norris, S. Roh- .~MERICAN ORNITHOLOGISTS' UNION. 1957. Check- wer, and J. C. Wingfield kindly shared unpublished list of North American birds. 5th ed. American manuscriptswith me that greatly benefitted this study. Ornithologists’ Union, Baltimore, MD. J. P. Collins. M. E. Douglas. M. C. Moore, and D. BAILEY, F. M. 1928. Birds of New Mexico. New Mex- Young provided much statistical help and advice. M. ice Dept. Fish and Game, Washington, DC E. Douglas, D. K. Hews, R. Knapp, C. A. Marler, M. BARBOUR.T. 1923. The birds ofCuba. Memoirs Nut- C. Moore, D. L. Pearson, S. Rohwer, R. L. Rutowski, tall dmithol. Club 6: 130. G. E. Walsberg,B. Young, and an anonymousreviewer BARBOUR, T. 1943. Cuban ornithology. Memoirs read the manuscript and provided much helpful crit- Nuttall Omithol. Club 9: 129. icism. I am greatly indebted and deeply grateful to my BENNITT,R. 1932. Checklist of the birds of Missouri. wife who was an endlesssource of advice, supportand Univ. MO. Stud. 7:60. inspiration throughout the project. BOND. J. 1980. Birds of the West Indies. 4th ed. I am also very grateful to the following institutions Houghton Mifflin, , NY. and their curators for their loans of specimensto me BROWN,M. B., AND C. R. BROWN. 1988. Access to (number of specimens in parentheses):Museum of winter food resourcesby bright- versus dull-col- Northern Arizona (2) University of ArkansasMuseum ored House Finches. Condor 90:729-73 1. (4), University of British Columbia (4) California BROWN,C. R. 1984. Light-breasted Purple Martins Academy of Science(40) San Diego Societyof Natural dominate dark-breasted birds in a roost: impli- History (lo), Natural History Museum of Los Angeles cations for female mimicry. Auk 101: 162-l 64. County (2 l), University of California at Los Angeles BURLEIGH.T. D. 1944. The bird life of the Gulf Coast (37), University of California at Berkeley(80) Western region of Mississippi. Occas. Pap. Mus. Zool. La. Foundation of Vertebrate Zoology (43) Occidental State Univ. 203329~490. _ College Moore Laboratory of Zoology (MLOCAL) BURLEIGH.T. D. 1958. Georeia birds. Univ. Okla- (223) Denver Museum of Natural History (18) Yale homa Press, Norman, OK: Peabody Museum of Natural History (25), Delaware BUTCHER,G. S., AND S. A. ROHWER. 1989. The evo- Museum of Natural History (DMNH) (70) U.S. Na- lution of conspicuousand distinctive coloration tional Museum of Natural History (182) University for communication in birds, p. 51-108. In R. F. ofMiami (24). Tall Timbers ResearchStation (7) Flor- Johnston [ed.], Current ornithology. Vol. 6. Ple- ida State University Museum (4), Florida State Mu- num Press, New York, NY. seum (18). Universitv of Central Florida (22). Uni- CARR~KER,M. A., JR. 1910. An annotated list of the versity%of (13X Chicago Academy of Science birds of Costa Rica, including Cocos Island. Ann. (5) Field Museum of Natural History (204) Coe Col- Carnegie Mus. 6:885. leae (2). Universitv of Kansas (4 1). SouthwesternCol- CHERRY,J. D. 1985. Early autumn movements and lege Museum of Natural History’(7) State prebasicmolt of Swainson’sThrushes. Wilson Bull. University (LSU) (142) Harvard Museum of Com- 97:358-370. parative Zoology (MCZ) (195) Allan R. Phillips (6) Corms, E. 1903. Key to North American birds. 5th University of Michigan (185), University of Minnesota ed. Dana Estes, Boston, MA. (UMINN) (52) University of Mississippi (7), Univer- DEARBORN,N. 1907. Catalogueof a collectionof birds DELAYED PLUMAGE MATURATION 233

from Guatemala. Field Mm. Nat. Hist. Pub. 125, female Tree Swallows. J. Field Omithol. 54:312- Omithol. Ser. 1:118-l 19. 318. DWOND, J. M. 1972. Avifauna of the easternhigh- IMHOF, T. A. 1976. Alabama birds. Univ. Alabama landsofNew Guinea. Nuttall Omithol. Club Mon- Press, University, AL. ogr. 12. Cambridge, MA. JAMES,D. A., ANDJ. C. NEAL. 1986. Arkansasbirds: DICKEY,D. R., AND A. J. VAN ROSSEM. 1938. The Their distribution and abundance.Univ. Arkan- birds of El Salvador. Field Mus. Nat. Hist. Publ. sas Press, Fayetteville, AR. Zool. Ser. 23583-584. JKRVI,T., E. RBSKAFI, M. BAKKEN,AND B. ZUMSTEG. DWIGHT,J., JR. 1900. The sequenceof plumagesand 1987. Evolution of variation in male secondary moults of the passerinebirds of New York. Ann. sexual characteristics:a test of eight hypotheses N.Y. Acad. Sci. 13:73-360. applied to Pied Flycatchers. Behav. Ecol. Socio- EWALD,P. W., ANDS. ROHWER. 1980. Age, coloration biol. 20:161-169. and dominance in nonbreeding hummingbirds: a JEHL,J. R., JR. 1990. Aspectsof the molt migration, test of the asymmetry hypothesis. Behav. Ecol. p. 102-l 13. In E. Gwinner [ed.], Bird migration: Sociobiol. 71273-279. physiology and ecophysiology. Springer-Verlag, FISK, E. J. 1974. Wintering populations of Painted New York. Buntings in southern Florida. Bird-banding 45: JOHNSTON,R. F. 1965. A directory to the birds of 353-359. Kansas. Univ. Kans. Mus. Nat. Hist. Misc. Publ. FORSHAW,J. M., AND W. T. COOPER. 198 1. Parrots 41. of the world. 2nd ed. David and Charles, London, KETTERSON, E. D. 1979. Status signaling in Dark- U.K. eyed Juncos.Auk 96:94-99. _ _ FOSTER,M. S. 1987. Delayed maturation, neoteny, LACK. D. 1968. Ecological adautations for breedinn and social system differencesin two manakins of in birds. Methuen, London,U.K. the genus Chiroxiphia. Evolution 4 1~547-558. LAND, H. C. 1970. Birds of Guatemala. Livington, FRIEDMANN, H., L. GRISCOM,AND R. T. MOORE. 1957. Wynnewood, PA. Distributional checklist of the birds of Mexico. LANYON,S. M., AND C. F. THOMPSON. 1986. Site Part. 2. Pacific Coast Avifauna No. 33. fidelity and habitat quality as determinants of set- G~NN,H. B., AND D. S. MELVILLE. 1983. Moult in tlement pattern in male Painted Buntings.Condor birds. British Trust for Ornithology. Hertford- 88:206-210. shire, U.K. LAWTON,M. F., AND R. 0. LAWTON. 1986. Hetero- GOODRICH,A. L., JR. 1946. Birds in Kansas. Report chrony, deferred breeding, and avian sociality, p. Kansas State Board Agriculture, June, 1945. To- 187-22 1. In R. F. Johnston [ed.], Current omi- peka, KS. thology. Vol. 3. Plenum Press, New York, NY. Goo~wr~, D. 1986. Crows of the world. 2nd ed. LIGON.J. D. 1961. New Mexico birds and where to Univ. of Washington Press, Seattle, WA. find them. Univ. New Mexico Press, Albuquer- GRANT, P. R. 1986. Ecology and evolution of Dar- que, NM. win’s finches. Princeton Univ. Press, Princeton, LIGON,J. D. 1971. Late summer-autumnal breeding NJ. of the Piiion Jay in New Mexico. Condor 73: 147- GRISCOM,L. 1932. The distribution of bird-life in 153. Guatemala: a contribution to the study of the or- LYON, B. E., AND R. D. MONTGOMERIE.1986. De- igin of Central American bird-life. Bull. Am. Mus. layed plumage maturation in passerinebirds: re- Nat. Hist. 64:364-365. liable signallingby subordinate males?Evolution 40:605-615. GROSS,M. R. 1984. Sunfish, salmon, and the evo- LOWERY,G. H., JR. 1955. Birds of Louisiana. 1st ed. lution of alternative reproductive tactics in fishes, Louisiana State Univ. Press, Baton Rouge, LA. p. 55-75. In G. Potts and R. Wooton [eds.], Fish reproduction: Strategies and tactics. Academic MAYNARDSMITH, J., AND G. A. PARKER. 1976. The Press, New York, NY. logic of asymmetric contests. Anim. Behav. 24: 159-175. HARRISON,C.J.O. 1974. Maluridae. Part 2, p. 199- 214. In B. P. Hall [ed.], Birds of the Harold Hall MAYR, E. 1933. Birds collected during the Whitney expeditions 1962-1970. British Mus. Nat. Hist., South Sea expedition. XXVII. Notes on the vari- London, U.K. ation of immature and adult plumages in birds and a physiological explanation of abnormal HERTER,D. R., S. M. JOHNSTON,AND A. P. WOODMAN. plumages.Am. Mus. Novit. 666. 1989. Molt migration of scotersat Cape Peirce, Alaska. Arctic 42:248-252. MAYR, E. 1934. Birds collected during the Whitney South Sea expedition. XXIX. Notes on the genus HOWELL,A. H. 1911. Birds of . U.S.D.A. Biological Survey Bull. 38:68. Petroica. Am. Mus. Novit. 714. HOWELL,A. H. 1928. Birds of Alabama. 2nd. ed. MONROE,B. L., JR. 1968. A distributional survey of Alabama Dem. Fish Game, Montaomerv, AL. the birds of Honduras. American Ornithologists’ HOWELL,A. H. _ 1932. Florida bird-life. Coward- Union Monogr. 7. McCann, New York, NY. MON~GOMERIE,R. D., AND B. E. LYON. 1986. Does HUMPHREY,P. S., AND K. C. PARKES. 1959. An ap- longevity influencethe evolution of delayed plum- proach to the study of molts and plumages.Auk age maturation in passerinebirds? Am. Nat. 128: 76:1-31. 930-936. HUSSELL,D. J. T. 1983. Age and plumage color in MOUNTJOY,D. J., ANDR. J. ROBERTSON.1988. Why 234 CHRISTOPHER W. THOMPSON

are waxwings “waxy”? Delayed plumage matura- Romvnn, S. 1983. Testing the female mimicry hy- tion in the Cedar Waxwing. Auk 10561-69. pothesis of delayed plumage maturation: a com- Nma, M. M. 1931. The birds of Oklahoma. Rev. ed. ment on Proctor-Gray and Holmes. Evolution 37: Pub. Okla. Biol. Surv. 3. 421423. NORRIS,D. J., ANDW. H. ELDER. 1982. Distribution Rorrwaa, S. 1985. Dyed birds achieve higher social and habitat characteristicsof the Painted Bunting status than controls in Harris’ Sparrows. Anim. in Missouri. Trans. MO. Acad. Sci. 16:77-83. Behav. 33:1325-1331. OB~RHOLS~R,H. C. 1938. The bird life of Louisiana. Rot-rwnn, S. 1986. A previously unknown plumage La. State Dep. Conserv. Bull. 28. of first-year Indigo Buntings and theories of de- OBWIOLSER.H. C. 1974. The bird life of Texas. Vol. layed plumage maturation. Auk 103:28l-292. 2. Edited, and with distribution maps and addi- ROHWER,S., AND G. S. BUTCHER. 1988. Winter ver- tional material by E. B. K&aid, Jr., Univ. of Tex- sussummer explanationsof delayedplumage mat- as Press, Austin, TX. uration in temperate passerine birds. Am. Nat. OGDEN,J. C., ANDF. L. CHAPMAN. 1967. Extralimital 131:556-572. breeding of Painted Buntings in Florida. Wilson ROIIWER,S., AND P. W. EWALD. 1981. The cost of Bull. 79~347. dominance and advantage of subordination in a PARKER,G. A. 1974. Assessmentstrategies and the badge signalling system. Evolution 35:441454. evolution of fighting behavior. J. Theor. Biol. 47: Roxwnn, S., AND J. MANNING. 1990. Differences in 223-243. timing and number of molts for Baltimore and PARKES,K. C. 1967. Prealtemate molt in the Summer Bullock’s Orioles: implications to hybrid fitness Tanager. Wilson Bull. 79:456458. and theoriesof delayed plumagematuration. Con- Pwrn~. D. F. 1959. The breedina behavior of the dor 92:125-140. Painted Bunting in southern Oklahoma. Bird- ROHWER,S., AND D. M. Nuns. 1979. The subadult banding 30:1-18. plumage of male Purple Martins: variability, fe- PARMELEE, D. F. 1964. Survival in the Painted Bun- male mimicry and recent evolution. Z. Tierpsy- ting. Living Bird 3~5-7. chol. 5 1:282-300. PATERSON, A. 1972. Birds of . Durrell ROHWER,S., AND F. C. Rorrwnn. 1978. Status sig- Pubs., Brattleboro, VT. naling in Harris Sparrows: experimental decep- PETERSON,R. T., AND E. L. (=HALIF. 1973. A field tions achieved. Anim. Behav. 26:1012-1022. guide.to Mexican birds: field marks of all species ROHWER,S., S. D. Fnnrwn~~, ANDD. M. Nuna. 1980. found in Mexico. Guatemala. Belize (British Hon- Delayed plumage maturation in passerineplum- duras), El Salvador. Houghton Mifflin, Boston. ages and the deceptive acquisition of resources. PHILLIPS,A., J. -HALL, .&D G. MONSON. 1964. Am. Nat. 115:40&437. The birds of Arizona. Univ. Arizona Press. Tuc- Romvnn, S., W. P. KLQN, ANDS. HEARD. 1983. De- son, AZ. layed plumage maturation and the presumed PIMM, S. 1976. Estimation of the duration of bird prealtemate molt in American Redstarts. Wilson molt. Condor 78:550. Bull. 95:199-208. POITER, E. F., J. F. PARNELL,AND R. P. TEULINGS. ROOT,T. 1988. Atlas of wintering North American 1980. Birds of the Carolinas. Univ. North Car- birds: an analysis of Christmas bird count data. olina Press, Chapel Hill, NC. Univ. Chicago Press, Chicago, IL. PROCTOR-GRAY,E., ANDR. T. HOLMES. 1981. Adap- SCHODDE,R. 1982. The Fairy-wrens: a monograph tive significanceof delayed attainment of plumage of the Mahuidae. Lansdowne, New York, NY. in male American Redstarts:tests of two hypoth- SELANDER,R. K. 1965. On mating systemsand sexual eses.Evolution 351742-751. selection. Am. Nat. 99: 129-l 4 1. Rn>oELY, R. S., ANDJ. A. G\KYNNE,JR. 1989. Aguide SELANDER,R. K. 1972. Sexual selection and dimor- to the birds of Panama with Costa Rica, Nicara- phism in birds, p. 180-230. In B. Campbell [ed.], gua,and Honduras. Princeton Univ. Press,Prince- Sexual selection and the descent of man 187l- ton, NJ. 1971. Aldine, Chicago, IL. ROBINSON,J. C. 1990. An annotated checklist of the SLUD,P. 1964. The birds of Costa Rica: distribution birds of Tennessee.Univ. TennesseePress, Knox- and ecology. Bull. Am. Mus. Nat. Hist. 128:373- ville, TN. 374. Rorrwnn, S. 1975. The social significanceof avian SM~THE,F. B. 1966. The birds of Tikal. Natural His- winter plumagevariability. Evolution 29:593-610. tory Press, Garden City, NY. SM~TH~,F. B., AND R. A. PAYNTER,JR. 1963. Birds Rorrwna, S. 1977. Status signalling in Harris Spar- ofTikal. Guatemala.Bull. Mus. Comp. Zool. Harv. rows: some experiments in deception. Behaviour Univ. 128:307. 61:107-129. SOKAL,R. R., ANDF. J. ROHLF. 1981. Biometry. 2nd. Romvnn, S. 1978a. Reply to to Shields on avian win- ed. W. H. Freeman. San Francisco. CA. ter plumage variability. Evolution 32:670-673. SPRUNT,A., JR. 1954: Florida bird’life. Coward- Rortwna, S. 1978b. Passerinesubadult plumages and McCann, New York, NY. the deceptive acquisition of resources:test of a SPRUNT,A., JR. 1968. Eastern Painted Bunting, p. critical assumption. Condor 80: 173-179. 137-l 54. In A. C. Bent [ed.], Life historiesof North Rorrwnn, S. 1982. The evolution of reliable and un- American cardinals,grosbeaks, buntings, towhees, reliable badgesof fighting ability. Am. Zool. 22: finches, sparrows, and allies. Smithsonian Inst. 531-546. Press, Washington, DC. DELAYED PLUMAGE MATURATION 235

SPRUNT,A., JR.,AND E. B. Chamberlain. 1970. South Tours, J. A., AND J. A. JACKSON. 1987. Birds and Carolina bird life. Rev. ed.. with a supnlement bv birding on the Mississippicoast. Univ. Mississippi E. M. Burton. Univ. South Carolina P&s, Colum- Press,Jackson, MS. bia, SC. Tmvnas, R. L. 1971. The evolution of reciprocal al- S~EENHOF,K., M. N. KOCHERT,AND J. H. DOREMUS. truism. Q. Rev. Biol. 46:34-57. 1983. Nesting of subadult Golden Eagles in URBAN, E. K. 1959. Birds from Coahuila, Mexico. southwesternIdaho. Auk 100:743-747. _ Univ. Kans. Publ. Mus. Nat. Hist. 11:443-516. STILES.F. G.. A. F. SKUTCH.AND D. GARDNER. 1989. WALSBERG,G. E. 1983. Avian ecologicalenergetics, A’guide to the birds of’Costa Rica. Cornell Univ. p. 161-220. In D. S. Famer, J. R. King, and K. Press, Ithaca, NY. C. Parkes [eds.], Avian biology. Vol. 7. Academic STORER,R. W. 1951. Variation in the Painted Bun- Press,New York, NY. ting. Occas. Pap. Mus. Zool. Univ. Mich. 532: l- WEBB,D. R., W. P. PORTR~,AND P. A. MCCLURE. 11. 1990. Development of insulation in juvenile ro- STRESEMANN,E., AND V. SIXFSEMANN. 1966. Die dents:functional compromisein insulation.Funct. mauser der Vogel. J. Omithol. 107:1-445. Ecol. 4:25 l-256. STUDD,M. V., ANDR. J. ROBERTSON.1985. Life span, WETMORE,A., R. F. PASQUIER,AND S. L. OWN. 1984. competition, and delayed plumage maturation in The birds of the Republic of Panama. Part 4. Pas- male passerines:the breeding threshold hypothe-. . seriformes: Hirundinidae (Swallows) to Fringilli- sis. Am. Nat. 126:101-l 15. _ dae (Finches). Smithsonian Inst. Press, Washing- Sru-rc~~u~~, B. J., ANDR. J. ROBERTSON.1987. Two ton, DC. methods of sexingadult Tree Swallowsbefore they WIW, R. H. 1974. Evolution of social organization begin breeding. J. Field Omithol. 58:236-242. and life-history patterns among grouse. Q. Rev. Surro~. G. M. 1967. Oklahoma birds: their ecoloav Biol. 49:201-207. and distribution, with comments on the avifauna WILEY, R. H. 1981. Social structureand individual of the southern Great Plains. Univ. Oklahoma ontogenies:problems of description, mechanism, Press, Norman, OK. and evolution, p. 105-133. In P. H. Klopfer and TABER,W. 1968. Western Painted Bunting, p. 154- P.P.G. Bateson [eds.], Perspectives in ethology. 155. In A. C. Bent [ed.], Life histories of North Vol. 4. Plenum Press, New York, NY. American cardinals,grosbeaks, buntings, towhees, WILSON, A., AND C. L. BONAPARTE. 1831. American finches, sparrows, and allies. Smithsonian Inst. ornithology or the natural history of the birds of Press, Washington, DC. the United States. Edited by R. Jameson. Vol. 2. TAYU)R. W. K.. B. H. ANDERSON.AND H. M. STE- Constable, Edinburgh, Scotland. VE&ON. 1989. Breeding range expansionsof the W~I-~EN~ERGER,J. F. 1978. The evolution of mating , Painted Bunting, and Blue Gros- systemsin grouse.Condor 80: 126-l 37. beak in Florida with new records for Seminole W~I-TEN~ERGER,J. F. 1979. A model for delayed re- County. Fl. Field Nat. 17:l-10. production in iteroparousanimals. Am. Nat. 114: THIOLLAY.J. M. 1978. Les miarations de rapacesen 439-446. Afrique occidentale: adapt&ions ecologicpiesaux WOOD, D. S., AND G. D. SCHNELL. 1984. Distribu- fluctuations saisonnitres de production des Cco- tions of Oklahoma birds. Univ. Oklahoma Press, systemes.Terre Vie 32:89-133. Norman, OK. TI~ON, S. R., AND I. H. TIPTON. 1978. Some notes YOUNG, B. 1991. Annual molts and interruption of on Painted Buntings.North American Bird Band- the fall migration for molting in Lazuli Buntings. er 3:26. Condor 93:236-250.