sign in sign up
<<
Home , Anseranatidae, Plumage

The First Basic Problem: A Review of Molt and Homologies Author(s): Steve N. G. Howell, Chris Corben, Peter Pyle, Danny I. Rogers Reviewed work(s): Source: The Condor, Vol. 105, No. 4 (Nov., 2003), pp. 635-653 Published by: University of California Press on behalf of the Cooper Ornithological Society Stable URL: http://www.jstor.org/stable/1370569 . Accessed: 06/01/2012 16:38

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

University of California Press and Cooper Ornithological Society are collaborating with JSTOR to digitize, preserve and extend access to The Condor.

http://www.jstor.org FEATURE ARTICLES

The Condor 105:635-653 ? The Cooper Ornithological Society 2003

THE FIRST BASIC PROBLEM: A REVIEW OF MOLT AND PLUMAGE HOMOLOGIES

STEVE N. G. HOWELL1, CHRIS CORBENI, PETER PYLE1,3 AND DANNY I. ROGERS2'4 'Point Reyes Observatory, 4990 Shoreline Highway, Stinson Beach, CA 94970 2Johnstone Centre, Charles Sturt University, P.O. Box 789, Albury, NSW 2640,

Abstract. All have fundamentallysimilar patternsof plumage succession. Thus Humphreyand Parkes (1959) proposeda system of nomenclature(the H-P system), based on homologies, that has become standardfor molt studies in North America. However, presumablyanalogous similarities in patternbetween first basic and definitivebasic plum- ages have obscuredhomologies. Many plumagesconventionally known as "firstbasic" are better consideredas novel first-cycleplumages that lack homologous counterpartsin sub- sequent cycles. Consequently,current nomenclature does not consistentlyreflect between- homologies. Howell and Corben(2000b) proposedthat traditionaljuvenal plumage can be consideredan unambiguousstarting point for a terminologythat betterreflects pre- sumed homologies in basic ;alternate and other nonbasicplumages may not nec- essarilybe homologousbetween species. Fourunderlying strategies of increasingcomplexity incorporateall known patternsof plumagesuccession: the Simple Basic Strategy,the Com- plex Basic Strategy,the Simple AlternateStrategy, and the ComplexAlternate Strategy. We review inconsistencyin the H-P system; explain the four underlyingstrategies; and discuss how one can identify homologies (if any) between plumagesin first and subsequentcycles and among taxa. Many species have novel plumages added into their first plumage cycle; we argue that existing terminology for these plumages is unsuitableand we term them formative plumages attained by preformative molts. Finally, we provide examples of how this modified H-P system can be applied to diverse taxa of birds while reflectingthe ho- mology underlyingall basic plumagecycles. Ourrevision validatesthe flexibilityand utility of the H-P system. Key words: Complex Alternate Strategy, Complex Basic Strategy, formative plumages, molt terminology, plumage succession, Simple Alternate Strategy, Simple Basic Strategy.

El Problema del Primer Plumaje Bilsico: Una Revisi6n de las Homologifas de la Muda y del Plumaje Resumen. Todas las aves tienen patronesde sucesi6n del plumaje fundamentalmente similares.De este modo, Humphreyy Parkes(1959) propusieronun sistemade nomenclatura (el sistema H-P), basadoen homologias,el cual ha sido de uso comtinen estudiosde muda de plumajeen Norte America.Sin embargo,supuestas similitudes anailogas entre el primer plumaje baisicoy el plumaje definitivo basico han confundido las homologifas.Muchos plumajes convencionalmenteconocidos como "primerbasico" son consideradosmejor como plumajesoriginales del primerciclo que carecen de contraparteshom6logas en los ciclos siguientes.Consecuentemente, la nomenclaturaactual no reflejalas homologiasentre especies. Howell y Corben (2000b) propusieronque el tradicionalplumaje juvenil puede ser consideradocomo un punto de partidainequivoco para una terminologifaque refleje mejor las homologifaspresuntas en los plumajesbaisicos; los plumajesalternos y otros plu-

Manuscriptreceived 25 October2002; accepted9 May 2003. 3 Correspondingauthor. E-mail: [email protected] 4 Presentaddress: 340 Nink's Road, St. Andrews,Victoria 3761, Australia.

[635] 636 STEVE N. G. HOWELL ET AL.

majes no basicos pudieranno ser hom6logosentre especies. Cuatroestrategias de creciente complejidadincorporan todos los patronesconocidos de sucesi6nde plumajes:La Estrategia BaisicaSimple, La EstrategiaBaisica Compleja, La EstrategiaAlterna Simple, y La Estrategia Alterna Compleja.Examinamos ciertas inconsistenciasen el sistema H-P; explicamos las cuatroestrategias subyacentes, y discutimosc6mo se puedenidentificar homologias (cuando existen) entrelos plumajesdel primerciclo y de los ciclos siguientes,y entretaxa diferentes. Muchas especies tienen plumajes originales adicionales en su primer ciclo de plumaje; sostenemosque la terminologiaactual para estos plumajeses inadecuaday los denominamos como plumajes formativos, logrados por mudas preformativas. Finalmente, damos ejemplos de como este sistema H-P modificadopuede ser aplicado a diversos tipos de aves y al mismo tiempo reflejarla homologia subyacentea todos los ciclos de plumajes baisicos. Nuestrarevisi6n valida la flexibilidady utilidaddel sistema H-P.

INTRODUCTION The H-P system is founded upon a few tenets, foremost which are (1) that molts Molting (the normal and regular growth of among only that is, a bird can have no , by which plumages are attained) is one produce plumages; more and no fewer than it has molts; of the most important processes in the life cycle plumages that the molt of of every bird. Despite this, molt remains a rel- (2) complete (or near-complete) adults can be considered and thus atively little-studied subject in homologous across that is a (Thompson and Leu 1994, Rohwer 1999), and comparable species; (3) molting but not the physiological and environmental processes cyclic phenomenon (most, all, species' to an annual governing molt in wild birds are poorly under- life-history cycles correspond cy- but se is not a crite- stood (Voitkevich 1966, Payne 1972). It has long cle, chronological age per rion for been recognized, however, that all bird species making comparisons among species); that molts and should be named have fundamentally similar patterns of plumage (4) plumages of other in a bird's succession, and descriptive accounts of plumage independently phenomena life such as seasons or cycles exist for many species worldwide. cycle, breeding status; and that molts should be named on the basis The molt strategies of birds are inextricably (5) of linked with other aspects of their life history, the plumage they produce, not on the basis such as breeding cycles, food supply, and, in of the plumage they replace. The H-P system is some cases, migration cycles. Traditional ter- standard for molt and plumage studies in North minologies have used terms such as "breeding" America, although some have questioned its or "nuptial" plumages, or "summer" and "win- practicality (e.g., Willoughby 1992). Any system ter" plumages. Humphrey and Parkes (1959, that attempts to draw lines on natural processes 1963) pointed out, however, that for meaningful will be imperfect-Nature does not sleep in a comparisons to be made among species, a sys- Procrustean bed-but it is widely accepted that tem of nomenclature for molts and plumages the H-P system is the most practical for studying should be free from preconceptions related to molt and plumage succession (Rohwer et al. other life-history phenomena. Consequently, 1992, Thompson and Leu 1994). their 1959 paper introduced a system of nomen- In view of accumulating data, Howell and clature whereby variations in the patterns of Corben (2000a, 2000b) identified and addressed plumage succession may be described, com- an inconsistency in the H-P system: the nonco- pared, and contrasted among different groups of rrespondence between nomenclature and pre- birds by applying the concept of homology to sumed homology in so-called first basic plum- the study of molts. The defining criterion for ho- ages. Here we review current misuse of, and in- mology is common ancestry, but determining consistency in, the conventional H-P system; this can be problematic, if not impossible (Min- pursue the implications of Howell and Corben's dell and Meyer 2001). Nonetheless, observa- (2000b) reinterpretation by explaining the four tions and syntheses of broad patterns can pro- underlying molt strategies they identified; and duce hypotheses regarding presumed homology, summarize how these strategies apply to all which can then be tested. The so-called Hum- North American and Australian bird families phrey-Parkes (H-P) system provides a terminol- (representing a diverse spectrum of species with ogy for such hypotheses and has proved useful which we are familiar). We also define and name in comparative studies of molt. plumages and molts inserted into the first cycle; REVIEWOF MOLTAND PLUMAGEHOMOLOGIES 637 discuss how one may identify presumed homol- A crux of the H-P system is how one recog- ogies (if any) between plumages in first and sub- nizes homologies. Rohwer et al. (1992:298) ad- sequent cycles and across taxa; and provide case vocated that molt homologies be determined by studies to illustrate how this modified H-P sys- comparing "the timing, extent, and color change tem reflects presumed homology in all basic in each molt with that of closely related species plumages. This revised framework of first-cycle that have already been described." This ap- molts and plumages defines a more useful start- proach may reveal homologies among closely ing point for addressing questions about the evo- related species, but if the patterns of plumage lution of molt strategies, and about potential succession already described have been founded molt and plumage homologies. on an arbitrary point of reference (see below), then broader patterns of homology could be ob- MISUSE OF THE HUMPHREY-PARKESSYSTEM scured. Consequently, conventional methods of under-represent Readers are referred to Humphrey and Parkes determining homologies might the potential of the H-P system. To recognize (1959) for a full explanation of their system, but potential molt homologies it seems more helpful the gist is that the plumage attained by adult to identify and compare patterns of plumage birds in their complete (or near-complete) molt succession across a broad range of taxa. (which usually repeats on an annual cycle) may be considered homologous across species; other WHY THE CONFUSION? plumages have evolved as additions in response Misuse of the H-P system results largely from a to various types of selection. This presumed ho- combination of three factors: (1) an inherent in- mologous plumage shared by all birds is termed consistency relating to the starting point of no- basic plumage. Any second plumage added into menclature for plumage succession; (2) a failure the basic plumage cycle is termed an alternate to divorce plumage color and pattern from an plumage, which usually involves molt of only evaluation of homologies; and (3) widespread head and body feathers (i.e., not including all misconception that the H-P plumage terms basic remiges and rectrices). and alternate are simply synonyms for tradition- Conceptually brilliant, yet theoretically sim- al nonbreeding (or winter) and breeding (or the H-P system has been frequently mis- ple, summer) plumages, respectively. used, valiant to clarify its usage despite attempts The inherent inconsistency stems from the H- Rohwer et al. (1992) and and Leu by Thompson P system taking its starting point of cyclic plum- of misuse abound. For Trum- (1994). Examples age succession as the highly variable molt that Mitchell peter (Cygnus buccinator), replaces juvenal plumage with so-called first ba- treated the as (1994) plumage sequence juvenal sic plumage. A consequence of this is an arbi- to first alternate to definitive while for the basic, trary first cycle which, depending on species or Tundra Swan and (C. columbianus) Limpert even individuals, can start at any time over the Easton considered the as (1994) sequence juve- course of a bird's first year or two of life (Fig. nal to first basic to definitive basic. For the Rock lA). Humphrey and Parkes (1959) provided no Holder and Mont- Ptarmigan (Lagopus mutus), explanation why the molt immediately following gomerie (1993) described a plumage sequence juvenal plumage should be a prebasic molt. Pre- running from supplemental to basic to alternate, sumably they considered that any development while the White-tailed Ptarmigan (L. leucurus) of plumages should start from a basic plumage, plumage sequence is reported to run from basic a premise with which we agree in principle. It to supplemental to alternate (Braun et al. 1993). appears, however, that the perceived equivalence Both and both ptarmigans undoubtedly (based largely on phenotypic similarity of color have homologous plumage sequences, so why and pattern) of traditional adult non-nuptial and the confusion in terminology when using an in- first non-nuptial plumage was extended, by anal- herently simple system? Indeed, even Rohwer et ogy rather than any evident homology, to the al. (1992) and Thompson and Leu (1994) dis- equivalence of definitive basic and first basic agreed over how one should name first-year plumage in the H-P system (Humphrey and plumages of certain Passerina buntings with a Parkes 1959). view to recognizing so-called homologous Similarities in the color and pattern of plum- plumages. ages could, understandably, cloud an apprecia- 638 STEVE N. G. HOWELL ET AL. tion of homologies. The physiological processes ventional first basic plumage is absent, then in that govern pigmentation and molt tim- its stead some other plumage will be named first ing are independent, however, although usually basic. they are coincident (Voitkevich 1966). Thus DEFININGTHE FIRSTPLUMAGE CYCLE feather pigmentation is a superficial-and not infallible-criterion for identifying molt homol- There is something of a catch-22 here: the in- ogies. Examples of the independence of feather consistency in conventional H-P nomenclature is coloring and molt timing range from wading clearly illustrated by redefining the first plumage birds (Beebe 1914) and (Oring 1968), to cycle, but a logical pathway to redefining the ptarmigan (Dwight 1900), large (Howell first cycle can only be appreciated when one rec- 2001a), and (Herremans 1999). The ognizes that there is an inconsistency in the pre- last example is particularly sobering for those sent system. Howell and Corben (2000b) ad- who advocate plumage color or pattern as a cri- dressed this issue by two logical steps: (1) as- terion by which to identify homologous molts. suming the ancestral molt strategy is the sim- The Black-chested Prinia (Prinia flavicans) has plest possible; and (2) defining juvenal plumage two complete molts per year, and males typically as the first basic plumage. molt from a bright alternate plumage to a dull We agree with Humphrey and Parkes (1959: basic plumage. However, if environmental con- 24) that "plumages were not originally sexually, ditions become suitable for breeding during or seasonally, or developmentally dimorphic" and just before the prebasic molt, males can molt that such "primitive or ancestral plumages were from bright alternate plumage into an apparently most likely renewed periodically and completely identical bright basic plumage, or into a variably by protracted molts." We also agree with Ho- intermediate basic plumage (Herremans 1999). well and Corben (2000b:51) that it is most par- Certain first-cycle plumages (especially the simonious "to consider the ancestral molt strat- conventional first basic plumage of most Pas- egy to be the simplest possible, similar to that seriformes and ) resemble defin- manifested today by birds such as Procellarii- itive basic plumages in appearance. We contend formes ... and most ... which that these first-cycle plumages are for the most have only a single (i.e., basic) plumage per cy- part analogous, not homologous, with basic cle." We term this hypothesized ancestral pat- plumages of subsequent cycles, and that the re- of plumage succession the Simple Basic semblances in color simply reflect similar dy- Strategy (SBS); Howell and Corben (2000b) namics between hormones and pigmentation. named it the Primitive Basic Strategy. Molts cor- The great variability in the extent and timing of responding to those present in the Simple Basic many conventional first prebasic molts among Strategy occur in all birds and, it seems reason- and within species (discussed under Complex able to conclude, are the only molts that should Basic Strategy, below) lends support to the pre- be considered homologous across all species mise that these first-cycle molts are not homol- (Fig. 2). Thus, SBS molts comprise simply a sin- ogous with definitive prebasic molts (Howell gle complete or near-complete prebasic molt and Corben 2000b). once per cycle, and each cycle starts with a pre- By following the evolutionary scenario of an- basic molt (e.g., [Fulmarus cestral plumages outlined below, it becomes ap- glacialis] and Red-tailed [Buteo jamai- parent that a fundamental difficulty in applying censis] in Fig. 1). Any other molt strategy can H-P nomenclature to the conventional first cycle be viewed as building upon the Simple Basic is the absence of a prebasic molt that is homol- Strategy by the addition of extra molts into the ogous across or even within species (Fig. lA). basic plumage cycle. That is, the conventional first prebasic molt can A plumage cycle (often shortened simply to be complete, fairly extensive, very limited, or, cycle) "runs from a given plumage or molt to as argued by Howell and Corben (2000a) for the the next occurrence of the same plumage or Western (Larus occidentalis) and other molt" (Humphrey and Parkes 1959:3). Thus, large gulls, not present at all. This clearly causes plumage cycles can be basic cycles or alternate problems for conventional H-P nomenclature cycles, depending on which plumage is treated because if the first cycle has to be fixed from as the first plumage in the cycle. We use plum- the start of the "first prebasic molt" and a con- age cycle, or cycle, to mean a basic plumage REVIEWOF MOLT AND PLUMAGEHOMOLOGIES 639

(A)

Year 1 Year 2 Year 3 JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND NorthernFulmar PJ PB2 Red-tailed Hawk PJ PB2 Merlin(European) PJ PB2 PB3 Merlin(New World) PJ PB2 Glaucous-winged Gull PJ . PB2 PA2 PB3 Glaucous-winged Gull PJ PB2

(B)

Year 1 Year 2 Year 3 JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND NorthernFulmar PB2 PB3 Red-tailed Hawk PB2 PB3 Merlin(European__an_ PB2 PB3 Merlin(New World) PB2 PB3 Glaucous-winged Gull PB2 PA2 PB3 Glaucous-winged Gull PB2 PA2 PB3 FIGURE 1. Diagrammaticcomparison of molt cycles of NorthernFulmar, Red-tailed Hawk, Europeanand New WorldMerlins, and two strategiesof Glaucous-wingedGull. (A) illustratesconventional inconsistency of nomenclaturein traditionalfirst prebasic(PB 1) and second prebasic(PB2) molts both between and within taxa. Shadedzone representsconventional first plumage cycle. Note how presumedhomologies are not reflectedwhen numberingprebasic molts and plumage cycles. PJ = prejuvenalmolt; PAl = first prealternatemolt, etc. (B) illustratesrevised H-P nomenclaturethat reflects between-species homology in prebasicmolts andcycles. Shaded zone representsrevised first plumagecycle. PB1 equatesto PJ. PF1 = preformativemolt. cycle. The H-P cycle was defined in terms of an prove false; these may simply be analogous individual adult bird, but it can also be consid- plumages based on a widely shared function. A ered equivalent to the cycle of a species or pop- critical review of prejuvenal downy plumages ulation whose members share a common breed- could be helpful in this regard. Most bird species ing season. In most cases, even for species of have one such plumage while some reportedly tropical environments, the plumage cycle is an have two (e.g., ; Marchant and Higgins annual cycle (Prys-Jones 1982, Stutchbury and 1990). Some species apparently lack prejuvenal Morton 2001). Obviously, the starting point in downs (e.g., most hatch naked and life is not repeated in a cyclic manner for an molt directly into juvenal plumage; Winkler and individual bird, yet this event is a fundamental Christie 2002), whereas kiwis hatch wearing ju- part of the life cycle of all species. Consequent- venal plumage (Marchant and Higgins 1990). ly, following Howell and Corben (2000b), we These examples suggest that environmental and define the first plumage cycle as the period be- life-history traits affect the timing of acquisition tween the attainment of juvenal plumage and the of juvenal plumage, but the problem of analogy acquisition of the next basic plumage via a com- versus homology remains. For example, birds plete, or nearly complete, molt that corresponds with long nestling periods are those reported to to a molt in the Simple Basic Strategy. Thus, as have two downs. Has juvenal plumage in these a rule, the first cycle has a duration similar to species become modified into a second down, or subsequent basic cycles (Fig. 2, 3). has the prolonged nestling period caused a novel We acknowledge that treating juvenal plum- second down to evolve? Pending satisfactory an- ages as homologous is an assumption that may swers to such questions, treating juvenal plum- 640 STEVE N. G. HOWELLET AL.

Cycle 1 Cycle 2 Cycle 3

Simple Basic Juv B2 B3 B4 Strategy---

Basic "F B2 Compex ' B3

Simple Alternate Juv B Strategy B A2 B3 -1 A3 --B) ComplexAlternate Ju-v F Al B2 A2 B3 A3 B4 S trateg y - - - -- _-_- . . . . FIGURE2. Diagrammaticrepresentation of underlyingmolt strategiesthat build upon the Simple Basic Strat- egy (SBS; e.g., ):the Complex Basic Strategy(CBS; e.g., ),Simple AlternateStrategy (SAS; e.g., some large gulls), and Complex AlternateStrategy (CAS; e.g., Calidrissandpipers and small gulls). Following Humphreyand Parkes (1959), molts are indicated symbolically as breaks between plumages and cycles, and the area below the dashed line representsflight feathers(which can be replacedin some Fl and Al plumages). Note that a cycle extends from the initiationof one prebasicmolt to the initiationof the next, and that plumages (and molts) are consistentlynumbered in referenceto the cycle in which they occur;this consis- tency is absent from traditionalnomenclature (cf. Fig. lA). In the first cycle, a few CBS and CAS species may have a second preformativemolt; and CAS species can have an additional(presupplemental) molt insertedinto all cycles (cf. Fig. 4). Juv = JuvenalPlumage (= First Basic); Fl = FormativePlumage; Al = First Alternate Plumage,etc; B2 = Second Basic Plumage,etc.

ages as homologous is a useful approach to rec- er individual large gulls lack this partial molt ognizing broad patterns of plumage develop- (Temple 1972, Cramp and Simmons 1980, ment in birds (e.g., Table 1). Palmer 1988b, Forsman 1999, Howell 2001a). Thus, under conventional H-P terminology, INCONSISTENCYIN THE HUMPHREY-PARKES some Black-shouldered Kites, some Pacific Ba- SYSTEM zas, some Merlins, all American Kestrels, and Once an unambiguous first plumage cycle is de- some large gulls attain a second basic plumage fined, examples of inconsistency in the conven- via a complete prebasic molt that corresponds to tional first cycle (and consequent nomenclature) an SBS molt, while the surely homologous com- are easily identified. For example, Marchant and plete prebasic molt of some kites, bazas, Mer- Higgins (1993) reported that the "first prebasic lins, and gulls produces a first basic plumage molt" in Pacific Bazas (Aviceda subcristata) and (Fig. 1A). Black-shouldered Kites (Elanus axillaris) varies Essentially, when using conventional H-P no- from complete to partial, yet these complete and menclature, some species (or even individuals) partial molts are quite different: all individuals gain an extra "basic" plumage relative to the undergo the complete molt at the end of their Simple Basic Strategy and, as a consequence, first cycle, while some individuals also have an the facility of comparison is lost (Fig. lA). Roh- earlier partial molt inserted into the first cycle. wer et al. (1992:299) identified a similar prob- Merlins (Falco columbarius) in the Palearctic, lem of noncorresponding first basic and second American Kestrels (Falco sparverius), and in- basic plumages with Willoughby's (1986) inter- dividuals of some large gull species (Larus) pretation of first-cycle plumages in Aimophila have a partial molt in their first fall and winter sparrows, and noted that "between-species ho- prior to the complete prebasic molt at the end of mologies are lost in the definitive cycle"; how- their first cycle; Merlins in the Nearctic and oth- ever, they did not comment that the same dis-

FIGURE 3. Diagrammaticexamples of revised molt nomenclaturefor selected species (see Case Studies). Anna's (Calypte anna), LoggerheadShrike (Lanius ludovicianus),and Steller's Jay (Cyanocitta stelleri), used by Humphreyand Parkes (1959) to illustratetheir three major patternsof postjuvenalfeather REVIEWOF MOLT AND PLUMAGEHOMOLOGIES 641

YEAR 1 YEAR 2 YEAR 3 JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND

Cattle Egret (SAS) Juv Al B2 A2 B3

nada - (CBS) B2 2 JuV .... .------ AF

Mallard (CAS) Juv Al B2 A2 B3 A3

Surf Scoter (SAS) Juv Al B2 A2 3 A3

Wild (CAS) Juv I Al B2 \\A2 B3

(SBS) Juv 6B2 B3

(SAS) BA2 B3 ThayesGulGull JuvAl B2

(CAS)-[J-- Bonaparte's (CAS) Ju F1 Al B2 A2 B3 GullF - ...... -

Barn (SBS) Juv B263

Great Horned (CBS) F1 B2 B3 Owl .. . .

Anna's (CBS) Juv B2 B3 Hummingbird(CBS) ------

A2 Loggerhead Juv FB2 A B3 Shrike ((AS)------

Steller's Jay (CBS) B2 B3 FJuv ------replacement,involve only two strategies.Molts are representedby diagonallines betweenplumages and indicate averagetimings for species in NorthAmerica. The areaabove the dashedline representshead and body feathers, the area below the dashed line flight feathers.Divisions within the vertical scale indicate approximateextents of a molt when it does not involve all feathers(e.g., the molt producingFl in Mallarddoes not involve all head and body feathers). Plumage abbreviationsfollow Figure 2; numbersrefer to plumage cycles. SBS = Simple Basic Strategy;CBS = Complex Basic Strategy;SAS = Simple AlternateStrategy; CAS = ComplexAlternate Strategy. 642 STEVE N. G. HOWELL ET AL. crepancy exists in a wider context (e.g., between would require a change in terminology only for a hawk and a shorebird, or even within some the relatively few species following the SBS. species; Fig. lA). When one considers juvenal Clearly, fixing the start of the first plumage plumage or definitive basic plumage there is no cycle is a critical point for any nomenclature. ambiguity. Yet if one follows conventional H-P Because the basic plumage cycle is a fundamen- nomenclature, the term "first basic plumage" is tal characteristic of all species, Humphrey and ambiguous and arbitrary. This is contrary to one Parkes (1959) considered that any development of the H-P system's four criteria for a semanti- of plumage nomenclature should start from a ba- cally clean terminology: "the nomenclature sic plumage. Like workers before them, Hum- must be consistent" (Humphrey and Parkes phrey and Parkes (1959) focused on adult birds 1959:14). We suggest that it is preferable, and and worked backward from the relatively sim- possible, to name plumages such that (1) the ple, stereotyped adult molts to juvenal plumage; second basic plumage of one species corre- thus they equated traditional first winter plum- sponds to the second basic plumage of all other ages with first basic plumages and generated the species; and (2) plumages are numbered in ref- inconsistency detailed earlier. We believe it is erence to the cycle in which they occur (with more useful for a nomenclature to start from a cycles based on the SBS). fixed, nonarbitrary landmark rather than from a moving target. PROPOSEDREVISION TO HUMPHREY-PARKES For several reasons, NOMENCLATURE taking juvenal plumage as such a landmark is more logical than using If one accepts that molts corresponding to those the highly variable molt that replaces it. First, in the Simple Basic Strategy are the only molts juvenal plumage is shared by all birds and is that should be considered prebasic, that color more likely to be homologous across species and pattern can be misleading, and that first cy- than the molt by which some juvenal plumage cle molts need not be homologous with molts in is replaced. Second, juvenal plumage can intui- subsequent cycles, an that was missing re- tively be considered a bird's first basic plumage: veals itself (Table 1). Our reinterpretation ad- it is defined as "the first covering of 'true' heres to the tenets of the original H-P system ('vaned' or 'pennaceous') feathers" (Palmer except for the classification of many conven- 1972:72) and includes a bird's first remiges. tional first basic plumages. This may seem a rel- Third, it is always attained by a complete molt, atively minor, almost semantic, difference, but it completeness (or near-completeness) being a is significant in terms of identifying potentially unifying characteristic of definitive prebasic homologous molts and understanding the evo- molts recognized by the H-P system. For these lution of molt strategies. reasons, Howell and Corben (2000b) recom- mended that juvenal plumage can be considered ALIGNINGPLUMAGE CYCLES a bird's first basic plumage. Thus, a fixed point, Once inconsistency in conventional H-P nomen- presumed homologous in all birds, marks the clature is accepted, one is faced with the ques- start of the first plumage cycle. Traditional ju- tion of how to align cycles among species. Two venal plumage can still be called juvenal (and options present themselves, as discussed by Ho- birds in this plumage juveniles), but it should be well and Corben (2000b). The first is to consider recognized as synonymous with first basic plum- the complete prebasic molt at the end of a bird's age. first cycle to always be the first prebasic molt, When juvenal and first basic plumages are as is done conventionally with species that fol- considered equivalent, then conventional second low the Simple Basic Strategy (e.g., Procellarii- basic plumage for all species with alternate formes and most Falconiformes). This would re- plumages is still termed second basic. In SBS sult in a major upheaval of conventional termi- species, however, the conventional first basic nology (e.g., for all Passeriformes). The second plumage needs to be renamed second basic option is to consider the complete prebasic molt plumage, the second renamed third, and so on. at the end of a bird's first cycle the second pre- This is a relatively minor change, yet it allows basic molt, as done conventionally with species for a system of consistent nomenclature for all that have plumages added into the SBS. This prebasic molts and basic plumages of all species, REVIEWOF MOLT AND PLUMAGEHOMOLOGIES 643 something that conventional H-P nomenclature that nest on islands traditionally free from pred- fails to achieve (Fig. 1A versus iB). ators, and in larger Falconiformes that have rel- atively few predators. These species have rela- RECOGNIZINGUNDERLYING MOLT tively long chick or nestling stages during which STRATEGIES the young grow a strong juvenal (= first basic) Palmer (1972:74) stated "The great variation in plumage. Most or all individuals of these species modes of living that have evolved in birds in- do not breed in their first year, and the second cludes such great variation in the nature, timing, prebasic molt typically commences earlier than etc., of molts and resultant feather generations that of breeding adults. ... as to appear to mask any underlying pat- The Complex Basic Strategy (CBS). This strat- tern." Similarly, Jenni and Winkler (1994:8) egy applies to species in which a single molt concluded that "in non-passerines the variation (traditionally termed first prebasic) or (rarely) in moult and plumage cycles is so diverse that two molts are added into the first cycle but no a convincing common nomenclature is quite im- homologous molt(s) occur in subsequent cycles possible." Our cumulative experience in observ- (Fig. 2; it was termed the Modified Basic Strat- ing, investigating, describing, and documenting egy by Howell and Corben 2000b). The novel molt strategies for over 2000 diverse species, first-cycle molts replace whatever feathers pre- from temperate to tropical zones, and in both sumably need to be replaced for a bird to sur- Northern and Southern Hemispheres, leads us to vive its first year and enter into the definitive conclude, however, that an underlying order is cycle. A plumage similar in appearance to defin- present (which, indeed, it should be, assuming itive basic plumage usually results, but we con- that modern birds share a common ancestor). sider this similarity analogous, not homologous Once a consistent, fixed point from which to (see above). start a nomenclature of molts and plumages is The Complex Basic Strategy is found in a established (i.e., the synonymy of juvenal and fairly diverse group of birds (Table 1). One rea- first basic plumage), one can recognize consis- son for its frequent occurrence may be that ju- tent patterns of plumage succession, regardless venal plumage grows during the breeding sea- of how the molts and plumages are named. We son, while definitive prebasic molts generally propose that four underlying strategies of post- follow the breeding season. Consequently, the juvenal feather replacement can be identified, first cycle is usually slightly longer than defini- starting with, and building upon, the Simple Ba- tive cycles, and juvenal plumage would have to sic Strategy. These four strategies reflect life- be at least as durable as definitive basic plumage history traits to varying degrees, show patterns if it were to survive the first cycle. But juvenal of phylogenetic concordance (Table 1), and en- head and body plumage, at least, is generally compass all known patterns of plumage succes- poorer than subsequent basic plumages (Jenni sion. In order of increasing complexity, we term and Winkler 1994). One possible reason for this, these strategies the Simple Basic Strategy (SBS), for example in passerines, is that nestbound the Complex Basic Strategy (CBS), the Simple young are often susceptible to and so Alternate Strategy (SAS), and the Complex Al- they grow a functional, but not necessarily du- ternate Strategy (CAS; Fig. 2). In the first two rable, juvenal (= first basic) plumage with strategies, adults lack an alternate plumage, which they can leave the vulnerability of the while in the last two strategies adults have an nest. Subsequently they undergo a variably ex- alternate plumage. After the first cycle, molts tensive, novel molt (not repeated in subsequent follow a stereotyped pattern that we term the cycles) by which they attain stronger feathers definitive molt cycle (even though plumage may that protect them until the start of the second not yet have attained a definitive appearance). prebasic molt. The Simple Basic Strategy (SBS). This is the In most CBS species the inserted first-cycle simplest possible molt strategy and consists of a molt involves only head and body feathers, but single basic plumage per cycle, as defined above in some species it can be complete or nearly so, (Fig. 2). This pattern of plumage succession is thus appearing similar to a definitive prebasic relatively uncommon (Table 1), being found pri- molt. Close inspection usually reveals some dif- marily in ratites, in some large, long-lived sea- ferences, however: in the added first-cycle molt, birds (e.g., Sphenisciformes, ) birds are more likely to retain a few remiges, 644 STEVE N. G. HOWELL ET AL.

TABLE 1. Molt strategiesof NorthAmerican and Australianbirds by family or subfamily. follows AOU (1998) for overall sequence and North Americanfamilies, Christidisand Boles (1994) for Australian nonpasserines,Schodde and Mason (1999) for Australianpasserines, with families integratedby referenceto Gill (1994). Numbersin parenthesesindicate total numberof species considered,followed by numberof species for which adequatedata exist to assign them to a strategy.X = predominantmolt strategy(or strategies)within a family (>35% of species); + = less-frequentmolt strategy(<35% of species). ? = this molt strategymay occur,as reportedin the literatureor suspectedby us.

Simple Complex Simple Complex Basic Basic Alternate Alternate Strategy Strategy Strategy Strategy Casuariidae(2,1) x Gaviidae(5,3) x Podicipedidae(11,7) X Spheniscidae(12,12) x Diomedeidae(12,8) X Procellariidae(50,13) X ? Hydrobatidae(11,2) X Phaethontidae(3,0) ? (8,5) X Pelecanidae(3,0) ? ? Phalacrocoracidae(13,5) ? X Anhingidae(2,0) ? ? Fregatidae(4,0) ? ? Ardeidae(26,7) + X ? Threskiornithidae(7,2) x x ? Ciconiidae(2,0) ? Cathartidae(3,3) x Phoenicopteridae(1,0) ? ? (1,1) x Dendrocygninae(4,2) x (11,10) x ? (50,22) ? x x (38,30) x + (11,10) x + (1,1) x Megapodiidae(3,3) x (20,8) x x Odontophoridae(6,6) x + Rallidae (25,11) X ? x Aramidae(1,0) ? ? Gruidae(4,0) ? ? Otididae(1,1) X Turnicidae(6,5) X X Pedionomidae(1,1) X Scolopacidae(60,51) ? + X Rostratulidae(1,1) X Jacanidae(3,1) X ? Chionididae(1,1) x Burhinidae(2,1) X ? Haematopodidae(5,1) ? ? X Recurvirostridae(5,2) ? ? x Charadriidae(24,13) + ? x Glareolidae(2,2) x Laridae(58,44) + X X Alcidae (21,6) ? X X "Near-Passerines" (37,33) X ? Cacatuidae(14,6) X Psittacidae(39,25) X Cuculidae(19,19) X Centropidae(1,1) X Tytonidae(5,5) X Strigidae(24,22) X ? Podargidae(3,3) x Aegothelidae(1,1) X REVIEWOF MOLTAND PLUMAGEHOMOLOGIES 645

TABLE 1. Continued.

Simple Complex Simple Complex Basic Basic Alternate Alternate Strategy Strategy Strategy Strategy Caprimulgidae(12,11) X + Apodidae (8,3) ? x Trochilidae(16,16) x Trogonidae(2,2) X Alcedinidae(13,9) + x Meropidae(1,1) X Coraciidae(1,1) X Picidae (22,22) x Passerines Pittidae(4,3) x Tyrannidae(35,29) x x Menuridae(2,1) x Atrichornithidae(2,1) X Climacteridae(6,6) x Maluridae(22,21) X x Pardalotidae(4,4) x Acanthizidae(45,44) x Meliphagidae(72,66) X + Petroicidae(20,20) X Orthonychidae(2,2) x Pomatostomidae(4,4) x Laniidae(2,1) ? X Vireonidae(14,10) X X Eupetidae(8,8) X Neosittidae (1,1) X Pachycephalidae(14,14) X Dicruridae(19,16) x Paradisaeidae(4,4) X (14,7) X Campephagidae(8,7) X + Oriolidae(3,3) x Corvidae (23,23) x Corcoracidae(2,1) x Ptilonorhynchidae(10,1) X Alaudidae(3,2) X Dicaeidae (1,1) X Hirundinidae(13,9) x + Paridae(11,11) X Remizidae (1,1) X Aegithalidae(1,1) X Sittidae (4,4) x x Certhiidae(1,1) X Troglodytidae(9,9) X + Cinclidae (1,1) X Regulidae(2,2) X ? Zosteropidae(7,1) X Sylviidae (14,8) ? X Turdidae(20,16) x + Timaliidae(1,1) X Mimidae (10,9) X ? Sturnidae(4,1) x Motacillidae(9,9) X Bombycillidae(2,2) X Ptilogonatidae(1,1) X Peucedramidae(1,1) X ? Parulidae(53,45) X x Thraupidae(5,5) X x Emberizidae(49,47) X x Cardinalidae(10,10) x x Icteridae(22,13) X X Fringillidae (19,18) X X Passeridae(22,5) X 646 STEVE N. G. HOWELL ET AL. rectrices, or primary coverts (e.g., , documented for several species, for example, Trochilidae, Timaliidae). Also, comparisons White Ibis (Eudocimus albus; Beebe 1914, within species or among related species often Kushlan and Bildstein 1992) and scoters reveal differences. For example, in the ge- (Dwight 1914; see Case Studies, below), but it nus Neophema and the grosbeak Cardi- has been difficult to accept, or to identify as nalis, the conventional first prebasic molt in- such, because H-P convention considers the cludes none, some, or all flight feathers (Pyle postjuvenal molt as a first prebasic molt. 1997b, Higgins 1999). In such cases it is appar- The Simple Alternate Strategy has been found ent that complete postjuvenal molts represent mainly in relatively large, aquatic nonpasserines variation within an inserted molt rather than be- that, as a rule, do not breed in their first year, ing prebasic molts. We regard the usual occur- but critical study is needed to determine how rence of phenotypic differences between added widespread it is. Recently, Snyder and Russell first-cycle molts and definitive prebasic molts as (2002) described the Carolina Parakeet (Conu- evidence that these molts have different origins, ropsis carolinensis) as an SAS species. As with and are likely not homologous. But in cases the Complex Basic Strategy, the added first-cy- where no differences are readily apparent and cle molt presumably helps a bird survive to the there are no close relatives from which patterns definitive cycle, and it can be quite variable in can be inferred, how can one be sure that the extent, even within a species (e.g., Howell et al. complete postjuvenal molt is not actually a pre- 1999, Howell 2001a). In some groups (e.g., cer- basic molt? Following our definition of the first tain ducks and large gulls), we suspect the SAS cycle, complete postjuvenal molts are not pre- came about through loss of a second first-cycle basic if they do not correspond to molts in the molt (i.e., derived from the Complex Alternate Simple Basic Strategy. That is, these birds Strategy, following). Preliminary analysis sug- would gain an extra "basic" plumage relative to gests that birds exhibiting SAS include Gavii- the SBS (also see Howell and Corben 2000b). formes, Phalacrocoracidae (in which there may Species with complete or near-complete in- be no added first-cycle molt), some ducks (Ana- serted first-cycle molts (Pyle 1997a, Flannery tidae), and some large gulls. and Gardali 2000) usually appear to have good The Complex Alternate Strategy (CAS). This reason to replace flight feathers, given their life- strategy applies to species in which two (rarely history traits, as these feathers could become too three) plumages are added into the first cycle, worn to function adequately by the end of their while one (rarely two) added plumages occur in first breeding season, when the next molt occurs. definitive cycles (Fig. 2): for example, many Some are highly aerial and long-distance mi- Anatinae and Charadriiformes, and apparently grants (e.g., Trochilidae, Hirundinidae), some all Passeriformes that have alternate plumages as live in highly abrasive environments (e.g., adults (Table 1). One (rarely two) first-cycle [Chamaea fasciata]), and some are cav- molts (including conventional first prebasic) lack ity nesters that abrade their remiges when enter- a counterpart in the definitive cycle and function ing and leaving the nest (e.g., Picidae, European like the novel first-cycle molt of CBS species. [Sturnus vulgaris]). Rarely, the novel inserted molts can be complete The Simple Alternate Strategy (SAS). This and appear similar to a definitive prebasic molt, strategy applies to species in which at most only (e.g., Cassin's Sparrow [Aimophila cassinii]; a single plumage (traditionally termed first basic Willoughby 1986); see under Complex Basic and sometimes interpreted as comprising two Strategy for discussion of this situation. plumages) is added into the first cycle, and one Identifying the homology of inserted first-cy- plumage is added into definitive cycles (Fig. 2). cle molts in CAS species can be problematic. The inserted first-cycle molt usually appears ho- Conventionally the second inserted plumage has mologous with the prealternate molt of definitive been termed first alternate, but this may be in cycles. This strategy is relatively uncommon large part because the first was, inappropriately, (Table 1), and was not identified until Howell termed first basic. In some cases a review of and Corben (2000a) recognized it in the Western molt strategies within related families or within Gull, although over one hundred years ago an order may suggest a logical pathway for the Dwight (1901) noted this possibility for large of added plumages. In theory, a Com- gulls. The Simple Alternate Strategy has been plex Alternate Strategy could derive from either REVIEW OF MOLT AND PLUMAGE HOMOLOGIES 647 the CBS or the SAS. In the former case, defin- nology, the unique plumage of CAS species itive alternate plumage and the second inserted would have to be renamed supplemental. This first-cycle plumage appear homologous, as with would lead to contradictions and inconsistencies. passerines (see Case Studies, below). In the lat- For example, treating the unique first-cycle ter case, however, the second inserted first-cycle plumage of CBS species as alternate would plumage could equally be the novel plumage, wrongly imply homology with definitive alter- although we have found no case where the evo- nate plumages of closely related CAS species. lution of this sequence is suggested. To the con- Alternatively, treating the unique first-cycle trary, SAS may have derived from CAS in some plumage of CAS species as supplemental could cases (see Simple Alternate Strategy, above). wrongly imply homology with definitive supple- mental plumages (Fig. 4); it would also imply IDENTIFYINGAND NAMING FIRST-CYCLE that first-cycle molts homologous to those of MOLTSAND PLUMAGES CAS species do not occur in CBS species (the From the preceding synopsis of underlying molt unique first-cycle molt of CBS species cannot strategies, it is apparent that the first cycle can be called supplemental because there are only include one, two, or, rarely, three added plum- two plumages in the first cycle). These problems ages. We suggest it is most parsimonious to con- cannot be dismissed simply as the sort of con- sider molts and plumages present in definitive tradictions that will arise if homologies are cycles to have homologous counterparts in the sought between birds that are not closely related. first cycle, unless good evidence exists to the In groups such as the vireos (Vireonidae: Vireo) contrary. Thus in species following the Simple and wood-warblers (Parulidae: Dendroica) of Alternate and Complex Alternate Strategies, North America, or the honeyeaters (Meliphagi- first-cycle plumages can be identified that appear dae: Certhionyx) and grasswrens (Maluridae: homologous with definitive alternate plumages Amytornis) of Australia, some species have de- (see Case Studies). Still, because the majority of finitive alternate plumages while congeners do species exhibit Complex Basic or Complex Al- not (Pyle 1997b, Higgins et al. 2001). ternate Strategies (Table 1) they have at least We believe a logical solution to these prob- one first-cycle plumage that is not homologous lems is to introduce a new term to describe with plumages in the definitive cycle. plumages unique to the first cycle. Humphrey What should these unrepeated plumages be and Parkes (1959:17) themselves acknowledged called? The H-P system introduced the terms al- that some "cases may well show that parts of ternate and supplemental for plumages that can our fundamental thesis need to be altered or be added into a basic cycle. An alternate plum- broadened," and we believe this is such a case. age, attained by a prealternate molt, is any plum- In addition to solving nomenclatural problems, age inserted into the basic cycle. Any second it makes inherent sense that birds could have plumage (i.e., additional to alternate) added into plumages specifically evolved for the first cycle a basic cycle is termed a supplemental plumage, (as noted earlier). Unique postjuvenal molts and attained by a presupplemental molt, and it may plumages (i.e., ones that lack counterparts in precede or follow the alternate plumage. These subsequent cycles) have been identified recently definitions have in general proved to be unam- for a variety of species (Rohwer 1986, Thomp- biguous, practical, and helpful in studies of de- son and Leu 1994), and we are simply extending finitive cycles. Naming first-cycle molts and this concept to another predefinitive plumage. plumages, however, has often proved problem- We propose that the novel, first-cycle plumage atic (as noted above), perhaps because the defi- conventionally known as first basic plumage in nitions of alternate and supplemental were pro- CBS and CAS species can be termed formative posed with definitive cycles in mind. plumage attained by a preformative molt. Be- Use of the term alternate could, in theory, be cause it is a unique plumage, like juvenal plum- appropriate for the unique first-cycle plumage of age, it does not need to be numbered for the CBS species with a single inserted molt in the cycle in which it occurs (by definition a forma- first cycle. However, alternate is not available tive plumage occurs only in the first cycle) al- for use in the same context with CAS species though in plumage formulae the use of a suffix because they already have an alternate plumage "1" can be helpful (as in formative 1, or Fl). in the first cycle; following existing H-P termi- In addition to the conventional first basic plum- 648 STEVE N. G. HOWELL ET AL. Cycle1 Cycle2

Passerina bunting A B2 A2 B3 Ju-

Cycle 1 Cycle 2

Passerinabunting B Juy Al B2 A2

Cycle1 Cycle2

Ptarmigan Juv F 1 Al B2 A2 B3

Passerinabunting C Ju Fla AB2 A2 B3

FIGURE 4. Contrastof "first supplemental(Si)" plumage in Passerina buntingsA (Rohwer 1986), and B (Thompsonand Leu 1994) and ptarmigans;and our proposedrevision of first-cycleplumage terminologyfor the same buntings(C). Conventionand abbreviationsfollow Figure2. By termingnovel first-cycleplumages as formative plumages, ambiguity is removed concerningthe use of S1: in ptarmigan,Si is homologous with supplementalplumage in subsequentcycles; in buntingsit has been used for a novel, first-cycleplumage. age of CBS and CAS species, we suggest that plumages. Whereas juvenal and basic plumages the so-called supplemental plumages identified can reasonably be presumed homologous across by Rohwer (1986) and others can be viewed as species (Fig. IB and Fig. 2), nonbasic plumages auxiliary formative plumages (see Case Stud- (i.e., alternate, supplemental, and formative) ies). We also hope that recognition of prefor- need not be homologous among species, al- mative molts and their inherent variability ad- though usually they are so among closely related dresses problems raised by Willoughby (1986, taxa. This is an inevitable consequence of dif- 1992). We thus define a formative plumage as ferent evolutionary histories: for example, alter- any nonbasic plumage present in the first cycle nate plumages have undoubtedly evolved inde- but not in subsequent cycles. pendently among different groups of birds, as In the event that a species has more than one acknowledged by Humphrey and Parkes (1959). formative plumage, Fl can be used for the plum- In studies of molt and plumage succession age presumed homologous with the single for- among different taxa, one should therefore be mative plumage (Fl) of related species, while aware of a potential distinction between homol- Fla, Flb, etc., can be used for auxiliary for- ogies of nonbasic plumages within a species mative plumages, which may precede or follow compared to homologies between species. Fl. For example, if the second described for- CASE STUDIES mative plumage in Indigo Buntings (Passerina cyanea) appeared homologous to the single for- Applying our revision of the H-P system to all mative plumage of other Cardinalidae, then the birds is beyond the scope of this paper. Indeed, sequentially first, and novel, formative plumage a literature review indicates that critical data are would be termed Fla (Fig. 4). lacking for some species of nonpasserines (es- We close this section with a practical caution pecially in their first cycle) and that data are for workers looking to identify between-species sometimes contradictory between sources. In- homologies for formative (or any nonbasic) herent problems are that one can seldom follow REVIEWOF MOLTAND PLUMAGEHOMOLOGIES 649

individual wild birds over time, and that molt bird. Thus, the existence of Complex Alternate strategies (especially SAS) can be difficult to es- Strategy in requires substantiation (Table tablish from specimens. Consequently our re- 1). view should be considered provisional, and it is no more accurate than the data we summarize. ANATIDAE We review North American and Australian spe- The downy plumage phase of waterfowl is fol- cies because: reasonable data exist for most (1) lowed by juvenal plumage, and subsequent both Northern and species; (2) they encompass molts vary considerably in extent and timing Southern include a Hemispheres; (3) they great both among and within species (Palmer 1976a, of taxa, both resident and diversity migratory; 1976b). Members of the subfamily Anserinae and (4) include both and they tropical temperate (swans and geese) are usually considered to Thus should a wide species. they incorporate have only a single plumage per definitive cycle of possible molt range strategies. (e.g., Palmer 1976a). Juvenal plumage of An- The primary sources for our syntheses of molt serinae is partially replaced in fall and winter via strategies in Table I were data in regional hand- a variably extensive and often protracted prefor- books for North America (Palmer 1962, 1976a, mative molt; a complete prebasic molt then oc- 1976b, 1988a, 1988b, Pyle 1997b, Poole and curs in summer, after which molts are aligned Gill 1992-2002), the Western Palearctic (Cramp with the definitive cycle (i.e., CBS; Fig. 3). For and Simmons 1977, 1980, 1983, Cramp 1985), some geese, Cramp and Simmons (1977) re- and the Australian-Antarctic region (Marchant ported a prealternate molt of head and neck and Higgins 1990, 1993, Higgins and Davies feathers in adults, but we suspect they based 1996, Higgins 1999, Higgins et al. 2001, unpubl. their interpretation on seasonal criteria for nam- data). Although these sources used differing sys- ing so-called different molts when only a single, tems for naming molts and plumages, enough almost year-round prebasic molt presumably oc- data were usually provided to allow molt strat- curs (Gates et al. 1993). egies to be determined, at least provisionally. In Anatini and Aythyini, juvenal plumage is The examples given here align plumage cycles succeeded by formative plumage which, in among diverse groups of birds (Fig. 3), although many species, is very limited in extent and many examples underscore the need for more part of a more-or-less continuous transition into the data. Table 1 discusses taxonomic sequence for next molt, which the families; for orders we follow AOU (1998). produces bright plumage usually termed alternate (Palmer 1972, 1976a), ARDEIDAE e.g., (Anas platyrhynchos; Fig. 3). In most to retain most to Information for several species, such as. Black- , species appear all well into winter crowned Night- (Nycticorax nycticorax; juvenal plumage (Palmer Such a and Davis 1993, Howell and Pyle, unpubl. data) and 1976b; PP, unpubl. data). strategy, the of the and extent of two (Bubulcus ibis; Cramp and Sim- description timing molts in the first mons 1977, Marchant and Higgins 1990) indi- reported cycle (Palmer 1976b), recalls the cates that only a single added first-cycle molt confusion regarding two putative occurs over a bird's first winter, followed by a molts in large gulls and herons, when in fact complete or nearly complete prebasic molt in only a single molt may occur. Dwight (1914) summer and fall (Fig. 3). Humphrey and Parkes described molt in scoters (Melanitta) as SAS but (1963) discussed confusion surrounding the was unable to recognize it as such because of molts of Ardea herodias and A. cinerea, and we confusion derived from traditional terminology, suggest that the data they summarized more rea- for example, trying to distinguish arbitrary sonably accord with a Simple Alternate Strategy "postjuvenal" and "first prenuptial" molts or Complex Basic Strategy than with two puta- (within a single, protracted molt) by reference to tive molts (i.e., conventional first prebasic and season or color. Combined field and museum first prealternate). Late-molting birds could at- studies of (M. perspicillata; Howell tain more advanced-looking plumage, as occurs and Corben, unpubl. data) support this interpre- in large gulls (Howell 2001a), but only a single, tation (contra Palmer 1976b); thus, at least some potentially protracted, and variably extensive Mergini exhibit the Simple Alternate Strategy molt occurs in the first cycle of any individual (Fig. 3). 650 STEVE N. G. HOWELL ET AL.

GALLIFORMES examining the evolution of molt strategies in The molts of North American (Tetraoni- resident and migratory species. nae) and quail (Odontophoridae) were detailed "NEAR-PASSERINES" by Dwight (1900), and those of Wild Turkey (Meleagris gallopavo) by Leopold (1943): two Almost all North American and Australian near- inserted first-cycle molts reportedly occur in all passerines (sensu Pyle 1995, 1997b; see Table species, although subsequent data do not support 1) exhibit the Complex Basic Strategy, while a this for some quail (Raitt 1961). The prealternate handful exhibit the Simple Basic Strategy (no- molt of adult gamebirds is limited in most spe- tably Tytonidae; Fig. 3) and Complex Alternate cies (ptarmigan being a notable exception) and Strategy. Cramp (1985) considered the juvenal may not always occur, at least in females. In plumage of Tytonidae to be highly reduced and most species only the preformative molt is ex- modified to a downy plumage. This is contrary tensive, the second inserted first-cycle molt be- to the generally accepted definition of juvenal ing limited in extent and apparently homologous plumage (i.e., the first pennaceous plumage in with the definitive prealternate molt. In turkeys which birds fledge) and we agree with Higgins and ptarmigan, however, both the preformative (1999) that Tytonidae have two downs followed and first prealternate molts are extensive. Two by a durable juvenal plumage. When present, the extensive molts in ptarmigan parallel seasonal preformative molt of near-passerines varies from color change in the adult, while for turkeys Le- partial (e.g., Strigidae; Pyle 1997c; Fig. 3) to opold (1943) proposed that these molts were oc- complete (most Columbidae and Trochilidae; casioned by the great increase in size during a Pyle 1995). A second added molt occurs in the bird's first year of growth, and thus enable first- first cycle of some , with limited year birds to catch up to the definitive cycle. replacement of gorget feathers (Pyle et al. 1997); we consider this an auxiliary preformative molt CHARADRIIFORMES (Fig. 3). Considerable data are available for many species PASSERINES (Cramp and Simmons 1983, Cramp 1985, Mar- chant and Higgins 1993, Higgins and Davies Relatively good data exist for the timing and ex- 1996), and this cosmopolitan order includes spe- tent of molts in most North American passerines cies that exhibit all four molt strategies. The (Pyle 1997b and references therein; Table 1). only species in this order for which the Simple Passerina buntings in particular have been the Basic Strategy has been reported are the Chio- subject of several recent studies that used the nidae (Marchant and Higgins 1993) and Ivory term supplemental for a unique, first-cycle Gull (Pagophila eburnea; Howell 2001b; Fig. plumage (Rohwer 1986, Thompson 1991, Young 3). The Complex Basic Strategy occurs in the 1991). They exemplify the difficulties of iden- Rostratulidae and, apparently, some Jacanidae, tifying homologies by using the conventional H- among others (Marchant and Higgins 1993). P system: Thompson and Leu (1994) considered Data on Charadriidae are equivocal but the fam- the first postjuvenal plumage of Passerina bun- ily may include species with the Simple Alter- tings as a first supplemental plumage, followed nate Strategy (Pluvialis; Howell and Pyle 2002), by conventional first basic, the opposite of Row- Complex Alternate Strategy (most Charadrius), her et al. (1992). This involves a novel concept and CBS (Marchant and Higgins 1993). Most in molt terminology since, by H-P convention, Scolopacidae and small gulls (e.g., Bonaparte's Thompson and Leu's (1994) supplemental plum- Gull [L. philadelphia]) exhibit CAS, while most age occurs before the start of the first plumage large gulls exhibit SAS (Higgins and Davies cycle (Fig. 4). In identifying these plumages, 1996, Howell 2001a; Fig. 3). As in ducks and Rohwer et al. (1992) and Thompson and Leu passerines, the second added body plumage of (1994) relied heavily on comparisons with the the first cycle in and small gulls ap- definitive prebasic molts of Passerina. While pears homologous with definitive alternate this approach will produce an answer (one of the plumage. The cosmopolitan Charadriiformes first two molts will be more like definitive pre- pose many interesting questions (e.g., where basic than the other) it was based on the pre- does homology lie in first-cycle molts of remi- mise, which we reject, that one of these first two ges?) and could provide worthwhile subjects for molts had to be first prebasic. We consider in- REVIEW OF MOLT AND PLUMAGE HOMOLOGIES 651 stead that the first cycle of some Passerina bun- The ability to place any species in one of four tings simply involves two formative plumages underlying categories is a powerful tool for (Fig. 4). One still has the problem of determin- comparative studies, not least for the evolution ing which of these formative plumages is ho- of molt strategies across taxa. For example, the mologous with the single formative plumage of simplest basic and alternate strategies (SBS and related species. We suggest to solve this ques- SAS, respectively) occur only in nonpasserines tion one should compare the timing and extent (Table 1). When working with species whose of Passerina molts with the single preformative molt strategies are unknown one now has four molts of related species, and not with definitive options to consider as starting hypotheses, or prebasic molts. two options if it is known whether the species The only strategies known in North American in question has an alternate plumage. The ability and Australian Passeriformes are the Complex to categorize all species by these four strategies Basic Strategy and Complex Alternate Strategy can also highlight where potential exceptions oc- (Table 1). The more primitive families mostly cur, thus focusing attention on subjects for crit- exhibit CBS, although partial prealternate molts ical study. We accept, however, that the revised occur in some species (Table 1, Fig. 3). This nomenclature we have outlined, while more suggests that the ancestor had this streamlined than that previously available, is molt strategy, and it seems reasonable to pre- subject to modification with the acquisition of sume that the second added first-cycle plumage more data. This, like molt itself, is a function of of passerines with CAS is homologous with de- evolution. finitive alternate plumage, as both appear to have been inserted into the CBS. ACKNOWLEDGMENTS

CONCLUSIONS Foremost,we acknowledgeour debt to the far-sighted work of P. S. Humphreyand K. C. Parkes, and en- Other workers have described the goal of the H- courage studentsof molt and plumage succession to P system as being "to address questions regard- approachtheir subjects with open minds in the spirit the evolution of molt and that the H-P system was proposed40 years ago. We ing plumage sequenc- thank G. Elliott, C. M. L. and es in birds" and have advocated that "the H-P P. Battley, Hickey, Hug, S. Webb for discussions and inspiration,and P. S. system is the only existing method suitable for Humphreyfor thoughtfulcorrespondence on our pro- studying the evolution of molt and plumage suc- posed revision. W. Steele, A. Dunn, J. Matthew,and allowed to molt cession" and Leu 1994:778-779). E Coote kindly access prepublication (Thompson accounts for the Handbookof New Zea- We but have found some current use of Australian, agree land, and Antarctic Birds. Drafts of the manuscript the conventional H-P system to be arbitrary and benefitedfrom commentsby G. Ballard,T Gardali,R. thus inconsistent. In particular, automatically as- A. Hamilton,C. M. Hickey, D. Humple, D. Portelli, signing the first postjuvenal molt as the first pre- S. A. Rohwer, D. A. Sibley, D. Stralberg, C. W. N. Warnock,and S. Webb;P P basic molt has hindered the recognition of po- Thompson, Bricefio and H. G6mez de Silva G. helped with the Spanish tential homologies in patterns of molting. abstract.This is contributionnumber 977 of the Point Humphrey and Parkes (1963:503) concluded Reyes Bird Observatory. by asking "that critics provide for us and for other students of plumage succession an im- LITERATURE CITED and frame- proved conceptual terminological AMERICANORNITHOLOGISTS' UNION. 1998. Check-list work within which we may all continue produc- of North American Birds, 7th ed. The American tive research in this promising and important Ornithologists' Union, Washington, DC. field of ornithology." We suggest that the H-P BEEBE,C. W. 1914. Notes on the ontogeny of the White Guara alba. 1:241-248. system can be improved by (1) establishing an Ibis, Zoologica BRAUN, C. E., K. MARTIN,AND L. A. ROBB. 1993. unambiguous starting point (juvenal plumage) White-tailed Ptarmigan (Lagopus leucurus). In A. for molt and plumage nomenclature; (2) defining Poole and E Gill [EDS.],The birds of North Amer- a consistent first plumage cycle; and (3) recog- ica, No. 68. The Academy of Natural Sciences, nizing the existence of formative plumages. As Philadelphia, PA, and The American Ornitholo- DC. a consequence, four underlying molt strategies gists' Union, Washington, CHRISTIDIS,L., AND W. E. BOLES. 1994. The reveal themselves and serve as a frame- taxonomy working and species of birds of Australia and its territories. work from which to approach the subject of molt Royal Australian Ornithologists' Union Mono- and plumage sequence. graph 2, Melbourne. 652 STEVE N. G. HOWELL ETAL.

CRAMP, S. [ED.]. 1985. Handbookof the birds of Eu- HOWELL,S. N. G., ANDC. CORBEN.2000a. Molt cycles rope, the Middle East, and North Africa, volume and sequencesin the WesternGull. WesternBirds 4. OxfordUniversity Press, Oxford,UK. 31:38-49. CRAMP, S., AND K. E. L. SIMMONS[EDS.]. 1977. Hand- HOWELL,S. N. G., AND C. CORBEN.2000b. A com- book of the birdsof Europe,the Middle East, and mentary on molt and plumage terminology:im- NorthAfrica. Vol. 1. OxfordUniversity Press, Ox- plications from the WesternGull. WesternBirds ford, UK. 31:50-56. CRAMP, S., AND K. E. L. SIMMONS[EDS.]. 1980. Hand- HOWELL, S. N. G., J. R. KING, AND C. CORBEN. 1999. book of the birdsof Europe,the Middle East, and Firstprebasic molt in Herring,Thayer's, and Glau- NorthAfrica. Vol. 2. OxfordUniversity Press, Ox- cous-winged Gulls. Journalof Field Ornithology ford, UK. 70:543-554. CRAMP, S., AND K. E. L. SIMMONS[EDS.]. 1983. Hand- HOWELL,S. N. G., ANDP. PYLE.2002. Ageing and molt book of the birdsof Europe,the Middle East, and in nonbreeding Black-bellied Plovers. Western NorthAfrica. Vol. 3. OxfordUniversity Press, Ox- Birds 33:268-270. ford, UK. HUMPHREY, P. S., AND K. C. PARKES. 1959. An ap- DAVIS, W. E., JR. 1993. Black-crownedNight Heron proach to the study of molts and plumages.Auk (Nycticorax nycticorax). In A. Poole and E Gill 76:1-31. [EDS.],The birds of North America, No. 74. The HUMPHREY,P. S., ANDK. C. PARKES.1963. Comments Academy of Natural Sciences, Philadelphia,PA, on the study of plumagesuccession. Auk 80:496- and The AmericanOrnithologists' Union, Wash- 503. ington, DC. JENNI,L., AND R. WINKLER.1994. Moult and ageing DWIGHT,J. 1900. The molt of the North American Te- of Europeanpasserines. Academic Press, London. traonidae(quails, , and grouse).Auk 17: KUSHLAN, J. A., AND K. L. BILDSTEIN. 1992. White Ibis 34-51, 143-166. (Eudocimus albus). In A. Poole, P. Stettenheim, and Gill The birds of North DWIGHT, J. 1901. The sequence of moults and plum- E [EDS.], America, ages of the Laridae(gulls and ).Auk 18:49- No. 9. The Academy of NaturalSciences, Phila- 63. delphia, PA, and The American Ornithologists' DC. DWIGHT,J. 1914. The moults and plumages of the sco- Union, Washington, Oidemia.Auk 31:293-308. LEOPOLD,A. S. 1943. The molts of young wild and ters-genus domestic Condor45:133-145. FLANNERY, M. E., AND T. GARDALI. 2000. Incomplete turkeys. R. ANDS. L. EASTON.1994. Tundra Swan first prebasic molt in the Wrentit.Western Birds LIMPERT, J., 31:249-251. (Cygnus columbianus). In A. Poole and E Gill D. 1999. The of and the Mid- [EDS.], The birds of North America, No. 89. The FORSMAN, raptors Europe of Natural dle East. Poyser,London. Academy Sciences, Philadelphia,PA, R. D. C. AND C. and The AmericanOrnithologists' Union, Wash- GATES, J., E CAITHAMER,T. TACHA, DC. R. PAINE.1993. The annual molt cycle of Branta ington, canadensis interiorin relationto nutrientreserve MARCHANT, S., AND P. J. HIGGINS [EDS.]. 1990. Hand- book of New and Antarctic dynamics.Condor 95:680-693. Australian, Zealand, B. 1994. 2nd ed. Freemanand birds. Vol. 1 (partsA and B). Oxford University GILL,E Ornithology, Melbourne. Co., New York. Press, MARCHANT, S., AND P. J. HIGGINS [EDS.]. 1993. Hand- HERREMANS, M. 1999. Biannual moult in the complete book of Australian,New Zealand, and Antarctic Black-chestedPrinia Prinia flavescens. Ibis 141: 115-124. birds.Vol. 2. OxfordUniversity Press, Melbourne. MINDELL,D. P., AND A. MEYER. 2001. Homology HIGGINS, P. J. [ED.]. 1999. Handbook of Australian, in Vol. 4. Oxford evolving. Trends Ecology & Evolution16:434- New Zealand,and Antarcticbirds. 440. UniversityPress, Melbourne. MITCHELL,C. D. 1994. Trumpeter Swan (Cygnus buc- HIGGINS, P. J., AND S. J. J. DAVIES [EDS.]. 1996. F cinator). In A. Poole and F Gill [EDS.],The birds Handbookof Australian,New Zealand,and Ant- of North No. 105. The of Nat- arcticbirds. Vol. 3. Oxford Mel- America, Academy UniversityPress, ural Sciences, PA, and The Ameri- bourne. Philadelphia, can Ornithologists'Union, Washington,DC. J. M. AND W. M. STEELE HIGGINS, P. J., PETER, [EDS.]. ORING,L. W. 1968. Growth, molts, and plumages of 2001. Handbookof Australian,New Zealand,and the Gadwall.Auk 85:355-380. Antarcticbirds. Vol. 5. Oxford UniversityPress, PALMER,R. S. [ED.]. 1962. Handbook of North Amer- Melbourne. ican birds.Vol. 1. Yale UniversityPress, New Ha- HOLDER,K., ANDR. MONTGOMERIE.1993. Rock Ptar- ven, CT. migan (Lagopus mutus). In A. Poole and E Gill PALMER,R. S. 1972. Patterns of molting, p. 65-102. [eds.], The birds of North America,No. 51. The In D. S. Farner and J. R. King [EDS.], Avian bi- Academy of Natural Sciences, Philadelphia,PA, ology. Vol. 2. AcademicPress, New York. and The AmericanOrnithologists' Union, Wash- PALMER,R. S. [ED.]. 1976a. Handbook of North Amer- ington, DC. ican birds.Vol. 2. Yale UniversityPress, New Ha- HOWELL,S. N. G. 2001a. A new look at moult in gulls. ven, CT. 7:2-11. PALMER,R. S. [ED.]. 1976b. Handbook of North Amer- HOWELL,S. N. G. 2001b. Molt of the Ivory Gull. Wa- ican birds.Vol. 3. Yale UniversityPress, New Ha- terbirds24:438-442. ven, CT. REVIEWOF MOLT AND PLUMAGEHOMOLOGIES 653

PALMER,R. S. [ED.]. 1988a. Handbookof NorthAmer- ROHWER,S., C. W. THOMPSON,AND B. E. YOUNG.1992. ican birds.Vol. 4. Yale UniversityPress, New Ha- Clarifyingthe Humphrey-Parkesmolt and plum- ven, CT. age terminology.Condor 94:297-300. PALMER,R. S. [ED.]. 1988b. Handbookof NorthAmer- SCHODDE,R., ANDI. J. MASON.1999. The directoryof ican birds.Vol. 5. Yale UniversityPress, New Ha- Australianbirds. Passerines.Commonwealth Sci- ven, CT. entific and IndustrialResearch Organization, Col- Australia. PAYNE, R. B. 1972. Mechanismsand control of molt, lingwood, Victoria, N. E ANDK. RUSSELL.2002. Carolina Par- 103-155. In D. S. Farnerand J. R. [EDS.], SNYDER, R., p. King akeet In A. Poole and Avian biology. Vol. 2. Academic Press, New (Conuropsis carolinensis). York. E Gill [EDS.],The birds of North America, No. 667. The Birds of North America,Inc., Philadel- POOLE,A., ANDE GILL [EDS.]. 1992-2002. The birds The Birds of phia, PA. of North America(numbers 1-680). B. J. ANDE. S. MORTON.2001. Be- North PA. STUTCHBURY, M., America,Inc., Philadelphia, havioral ecology of tropical birds. Academic PRYS-JONES,R. P. 1982. Molt and weight of some land- Press, London. West Indies. Journalof Field birds on Dominica, TEMPLE,S. A. 1972. Sex and age characteristicsof Ornithology53:352-362. North AmericanMerlins. Bird Banding 43:191- PYLE,P. 1995. Incompleteflight-feather molts and age 196. in North Americannear-passerines. North Amer- THOMPSON,C. W. 1991. The sequence of molts and ican Bird Bander20:15-26. plumagesin PaintedBuntings and implicationsfor PYLE, P. 1997a. Molt limits in North Americanpasser- theories of delayed plumage maturation.Condor ines. North AmericanBird Bander22:49-90. 93:209-235. PYLE, P. 1997b. Identificationguide to NorthAmerican THOMPSON,C. W., ANDM. LEU. 1994. Determining ho- birds, part 1. Slate Creek Press, Bolinas, CA. mology of molts and plumagesto addressevolu- PYLE, P. 1997c. Flight-feathermolt patternsand age in tionaryquestions: a rejoinderregarding emberizid North American . Monographsin Field Or- . Condor96:769-782. nithology, number2. AmericanBirding Associa- VOITKEVICH,A. A. 1966. The feathers and plumage of tion, ColoradoSprings, CO. birds. OctoberHouse, New York. PYLE, P, S. N. G. HOWELL,AND G. M. YANEGA.1997. WILLOUGHBY,E. J. 1986. An unusual sequence of Molt, retained flight feathers, and age in North molts and plumages in Cassin's and Bachman's Condor88:461-472. American hummingbirds,p. 155-166. In R. W. Sparrows. Dickerman The era of Allan WILLOUGHBY,E. J. 1992. Incorrect use of the Hum- [COMPILER], Phillips: molt and for a festschrift.Horizon Communications, Albuquer- phrey-Parkes plumage terminology NM. buntingsof the genus Passerina. Condor94:295- que, 297. RAITT,R. J., JR. 1961. Plumage development and molts WINKLER, H., AND D. A. CHRISTIE.2002. Family Pici- of CaliforniaQuail. Condor63:294-303. dae 296-555. In J. del 1986. A unknown of (woodpeckers),p. Hoyo, ROHWER, S. previously plumage A. Elliott, and J. Sargatal[EDS.], Handbook of the first-yearIndigo Buntingsand theoriesof delayed birds of the World. Vol. 7. Lynx Edicions, Bar- plumagematuration. Auk 103:281-292. celona. ROHWER, S. 1999. Time constraintsand moult-breed- YOUNG,B. E. 1991. Annual molts and interruptionof ing tradeoffsin large birds.Proceedings of the In- the fall migrationfor molting in Lazuli Buntings. ternationalOrnithological Congress 22:568-581. Condor93:236-250.