THE AUK A QUARTERLY JOURNAL OF ORNITHOLOGY

VOL. 99 APRIL 1982 NO. 2

MIMICRY OF BY ORIOLES

JAREDM. DIAMOND PhysiologyDepartment, University of CaliforniaMedical School, LosAngeles, California 90024 USA

ABsTR•cT.--Friarbirdsand orioles show striking but parallel geographicvariation in plumageon islandsof the Australianregion, such that their populationson the sameisland are often closelysimilar in plumage.This parallelisminvolves mimicry of friarbirds by orioles,which are in turn mimicked by a smallerhoneyeater. The larger the com- pared to the oriole, the more perfectis the mimicry. Friarbirdsand oriolesare convergently similar in ecologyand morphologyand belong to a guild of that feed togetherin fruiting and flowering trees and that display much aggressiontoward each other. Within this guild, mimics are sparedfrom attack by larger models,which attackother smaller .Other possiblysimilar casesexist among tropicalbirds. Received5 January1981, Accepted2 September1981.

BIOLOGICALmimicry involves three roles: a are few well-established cases of visual mim- mimic, which resemblesanother speciesand icry in birds. thereby gains some advantage for itself; a Over a centuryago Wallace (1863, 1869: 305) model, which is imitated by the mimic; and a describedwhat he claimed to be an example signal receiver, which mistakes the mimic for of visual mimicry in birds, involving Old the model and thereby becomesthe victim of World orioles( , family Oriolidae) the mimic'sdeception (Wickler 1968). Mimicry and friarbirds (genus Philemon,family Meli- may be visual, acoustic, chemical, or behav- phagidae) in the Australian region. Wallace's ioral. Familiar examplesof visual mimicry in- claim was disputed by Stresemann (1914a), clude nonpoisonous butterflies or snakes that however,and there has been no further study haveevolved to resemblepoisonous butterflies of this case. The present paper is based on or snakesin order to gain the advantageof study of specimensof all speciesof oriolesand being avoided by predators.Although many friarbirds and on field studiesof a coexisting speciesare famous as vocalmimics, there mimic oriole and friarbird (), two

Frontispiece.Adult plumages of friarbirdsand orioles.Large capital letters are island names, small letters are taxon names.Top row right: a typical oriole outside the range of friarbirds (Oriolusoriolus). Top row left: a typicalfriarbird outside the rangeof orioles(Philemon cockerelli). Top row center:on the LesserSunda Islandsan oriolelarger than the O. [bouroensis]superspecies and of typicalOriolus plumage, O. chinensis, coexistswith the westernmostmember of the P. [moluccensis]superspecies, P. buceroides.Rows 2-6: coex- isting friarbirdsof the P. [moluccensis]superspecies and oriolesof the O. [bouroensis]superspecies. The oriolemimics the sympatricfriarbird to varying degreescorrelated with the friarbird'ssize: Australiaand have the smallestfriarbirds, and the two Australianoriole species and the maleoriole of Timor are not mimics,while the largestfriarbird (that of Ceram)is mimickedalmost perfectly. sticto- cephalusis a smallerhoneyeater that mimics the New Guineaoriole. See Table 1 for details.Painted by H. L. Jonesfrom specimensin the AmericanMuseum of NaturalHistory.

187 The Auk 99: 187-196.April 1982 UMBO LES R SUNDAS EURASIA

P. bucer, -ctus -•

ß' O. chinensis P. cockereHi •,0. oriolus

HALMAHE' CERA ' pious P. sue ßtus

forsteni aeochromu•

BUR

ßens s O. b. bouroensis TENIMB' .decipiens

WETA ' ? TIM R P lidiceps P.b bu , e8

•. v finschi J vi s

P. novae. '•b'e Isis P NEW GUINEA :iu•, stictocephalus

P. I10V * ctns

AUSTRALIA 188 JAREDM. DIAMOND [Auk, Vol. 99

I I IOøN-- /' 140øE 15(Y'E I Philippines.'..'•

Lesser $undas WeTM ___ IOOS -

Fig.1. Mapillustrating overlap of oriolesand friarbirds. Dots: Oriolus but no Philemonpresent. Vertical hatching:Philemon but no Orioluspresent. Solid shading: both Philemon and Oriolus present; solid under- liningof islandname indicates mimic oriole, dashed underlining indicates nonmimic oriole. Blank: neither Philemonnor Orioluspresent.

pairsof coexistingnonmimic orioles and friar- months, distributed over all months of the year and birds (),friarbirds living outsidethe over 9 yr from 1964 to 1981. range of orioles(, Umboi), and oriolesliving outsidethe rangeof friarbirds PATTERNS OF MIMICRY (,South Africa, Germany). The Meliphagidae, or ,originat- ed and (like marsupials)underwent an adap- STUDY MATERIAL AND STUDy SITES tive radiation in the Australian region and At the American Museum of Natural History, Brit- producedthe morphologicaland ecological ish Museum (Natural History), and Museum Zoo- equivalentsof numerousfamilies elsewhere logicumBogoriense I examinedall specimensof (Keast1976). Friarbirds, the largestmeliphag- friarbirds, all specimensof oriolesof the Australian ids, areconvergent on the predominantlyAfro- region(including ), and selectedspecimens Asian Oriolus in size, diet (, insects, soft of all extralimital oriole species. Field observationscovered the following areasand fruit), nest (woven suspendedcup), and hab- species.New Guinea(Fly River, OriomoRiver, Kar- itat (tree canopy).In adult plumagethe Afro- imui Basin,Astrolobe Bay, Huon Gulf, Torricelliand Asian orioles are mainly yellow and black Bewani mountains, Meervlakte, van Rees Moun- (Frontispiece:right-most two birds of top tains, Gauttier Mountains, Onin Peninsula, and vi- row), very unlike the mainlybrown friarbirds. cinitiesof PortMoresby, Popondetta, Wewak, Van- One speciesgroup of Oriolushas invaded imo, and Jayapura): Oriolus szalayi, Philemon and differentiated in the Australian region, so novaeguineae,P. corniculatus,P. citreogularis,P. mey- eri, and Pycnopygiusstictocephalus; Australia (Ath- that the orioles of the Oriolus [bouroensis] su- erton Tableland and vicinities of Canberra, Sydney, perspeciesand friarbirds of the Philemon Brisbane,and Cairns):O. sagittatus,O. fiavocinctus, [moluccensis]superspecies now shareeight is- Philemonnovaeguineae, P. corniculatus, P. citreogular- lands or island groups.These are Ceram, Hal- is; New Britain (Cape Gloucester)and Umboi: P. mahera, Bum, Tenimbar, , Timor, New cockerelli;Java (Bogor):O. chinensis;South Africa (KrugerPark): O. larvatus;Germany: O. oriolus. Guinea and neighboring islands(Waigeu, Sa- Observationsin the Australianregion totalled 25.5 lawati, Batanta,Misol, Japen,Aru), and north- April 1982] OrioleMimicry of Friarbirds 189

TABLE 1. Adult plumage of coexisting friarbirds TABLE 1. Continued. Philemon [moluccensis])a and orioles (Oriolus [bouroensis]).b AUSTRALIA. P. buceroidesand P. novaeguineae. CERAM. P. subcorniculatus. Olive-brown dorsal- As on New Guinea. O. fiavocinctus(streaky yellow- green, with black in wing and tail) and O. sagittatus ly, yellow-brown ventrally, breast and axillariesyel- (olive back, streaked on white below) dissimilar. low. Head markings: grey color; streaked crown; Dissimilar. small bare or sparselyleathered circumocularpatch. O. forsteni.Same. Virtually identical.½ • Friarbirdpopulations are describedwith respectto thoseof other islands,to illustrategeographic variation. . P. fuscicapillus. Medium-dark • Oriole populationsare describedwith respectto the friarbird of brown, darker dorsally; the darkest population. No the sameisland, to illustratemimicry. streaks and no markings except for small bare cir- •' The wordsin boldfacesummarize the closenessof mimicry. cumocular patch and slightly grey throat; the most uniformly colored population. O. phaeochromus. Same, except: slightly darker ventrally, throat not greyer, lacks circumocularpatch. Closely similar. . P.m. moluccensis.Body as on Halmahera, ern Australia(see Fig. 1). Fromisland to island except paler. Head markings: pale hind-collar, su- perciliary, and throat; streaked crown and throat; these oriolesshow great geographicvariation bare black facial patch. O. b. bouroensis.Same, ex- in plumage,and so do the friarbirds.Some are cept:has black featheredpatch correspondingto bare nearly uniform brown, others paler ventrally; skin of friarbird; side of throat streakJer.Closely darknessof colorationvaries; pale completeor similar. partial collars, dark ear coverts,dark mous- TENIMBAR. P. moluccensis timorlaoensis. As on tachestripes, pale superciliarystripes, streaked Buru, except: lacks pale superciliary; obscure dark crownsand throats,and olive or yellow washes ear-coverts and moustache. O. bouroensisdecipiens. Same, except: has black feathered patch correspond- may be present or absent; and one or more ing to bare black skin of friarbird; crown slightly black patches of varying sizes, consisting streakJer.Closely similar. either of bare skin (Philemon)or of solid feath- WETAR. P. buceroidespallidiceps. As on Buru, ex- ering or streakedfeathering (Oriolus), occur on cept: paler brown; throat paler, almost white; bare various parts of the head (Frontispiece,Table blackfacial patch larger; has bare black patchon side 1). This geographicvariation is largelyparallel, of neck; has black moustachefeathering adjacent to bare black facial skin; has bill knob. This population however, such that on all islands except Aus- and the Timor one are the palest populations. They tralia the oriole and friarbird resemble each are also the ones with the boldest facial pattern, due other, often to such a stunning degree that to the contrast between the black facial skin and specimensheld in the hand can be distin- moustacheand the whitish throat. O. viridifuscusfin- schi. Female same, except:has black featheredpatch guishedonly with difficulty (and were some- correspondingto patchesof bare black skin of friar- times misdescribed by taxonomists and mis- bird; streakiercrown and nape; lacksbill knob. Male labelled in museum collections). At least on as female but slightly darker, upper back and crown New Guinea (my observations)and Buru (Wal- greyer, head pattern more obscure. Quite similar lace 1863, 1869:305), this similarity in plumage (female), somewhat similar (male). is reinforced by similarities in posture, move- TIMOR. P. b. buceroides. As on Wetar. O. v. virid- ments, and flight, so that the oriole and friar- ifuscus.Female same, except: has black feathered patchcorresponding to patchesof bare black skin of bird are even harder to distinguishin the field friarbird; lacks bill knob. Male different (brown low- than one would infer from dead specimens. er back and belly, dull green upper back, grey breast The oriole and friarbird of New Guinea are and throat, red bill). Quite similar (female), dissim- amongthe less closelymatched mimic/model ilar (male). pairs in the hand. Yet the result of their be- NEW GUINEA. P. novaeguineae.As on Buru, ex- havioral similarity, added to their plumagere- cept:bare black facial patch larger; somepopulations semblance,is that not only western ornithol- have bill knob; lacks pale hind-collar and supercili- ary. O. szalayi. Similar, but plumage streaky; has ogists but also New Guinea hunters and streakedblack featheredpatch correspondingto bare gathererswho are expertat field identification black skin of friarbird; lacks bill knob; bill red (adult of birds often confuse the New Guinea oriole only). Quite similar. Pycnopygiusstictocephalus. and friarbird. Other field observers have com- Similar to oriole, except: body not streaky; crown streaked charcoal brown and white, instead of char- mented on the difficulty of distinguishingori- coalbrown and light brown; has white malar streak; oles from friarbirds on Ceram, Timor, and Te- bill black. Quite similar (to oriole). nimbar (Forbes 1885: 421, Stresemann 1914b). 190 JAREDM. DIAMOND [Auk, Vol. 99

The visual resemblanceis reinforced by vocal Timor coexist with, respectively, four, one, mimicry (New Guinea, Buru) and possiblyby and one other friarbird species besides P. interspecificduetting (New Guinea, Ceram). [moluccensis]. Several facts make clear that the detailed (2) Weight variation is slight between O. similarity of plumage involves mimicry of [bouroensis]populations but over twofold be- friarbirds by orioles rather than mimicry of tween P. [moluccensis]populations (population oriolesby friarbirds or convergenceof eachon means 94-109 g for orioles, 95-194 g for friar- the other. The predominantlybrown plumage birds). The greater the size advantageof the of the P. [moluccensis]forms is shared by all friarbird over the sympatric oriole, the more other Philemonspecies and many other meli~ perfectis the mimicry of the friarbird's plum- phagids, whereas the brown plumage of the age by the oriole. Thus, on Ceram, which has 0. [bouroensis]forms is unique within the fam- the P. [moluccensis]population with the largest ily Oriolidae. Within the O. [bouroensis]su- body size (mean weight 194 g, 78% heavier perspeciesthe form most similar to the typical than the Ceram oriole), the oriole is virtually yellow and black orioles of and Africa is a perfect mimic; on Halmahera, Buru, and Te- the most remotegeographically, the nonmimic nimbar, where the weight advantageis about 0. fiavocinctusof Australia. The patches of 50%, the mimicry is closebut not perfect; on bare black facial skin in the P. [moluccensis] New Guinea and Wetar, weight advantage forms are alsoshared by all but one other Phi- about 39%, mimicry fair; on Timor, weight lemon speciesand many other meliphagids, advantage13%, mimicry fair in female oriole, while the black featheringthat duplicatesthis absent in male; and in Australia, weight ad- skin in O. [bouroensis]forms lacksa closepar- vantageabout 12% (averagefor Philemonpop- allel in the Oriolidae. O. [bouroensis] is vir- ulations gordoni,ammitophila, and yorki vs. O. tually confined to islands shared with P. fiavocinctusand O. sagittatus),no mimicry. [moluccensis](the sole exceptionis a race of the Note the paradox: the more dissimilarthe size nonmimic O. fiavocinctuson the SouthwestIs- of the oriole and friarbird, the more similar the lands). In contrast, P. [moluccensis]occurs on plumage. at least 24 islands not shared with O. (3) A larger oriole of a superspeciesdifferent [bouroensis],and someof thesepopulations are from O. [bouroensis],O. chinensisbroderipi, co- identical in plumage to populationson islands exists on seven islands of the Lesser Sundas sharedwith O. [bouroensis].Evidently, coexis- (Lombok through Sumba and Alor) with the tence with orioles has not resulted in a modi- westernmost friarbird, P. buceroides of the P. fication of plumagein P. [moluccensis],where- [moluccensis]superspecies (Frontispiece, top as most O. [bouroensis] populations have row, central two birds). This oriole is yellow diverged from their presumed ancestralplum- and black like most Afro-Asian orioles and to- age to mimic the local population of P. tally unlike the brown P. buceroides,although [moluccensis]. the next two islands to the east of Alor (Wetar Among sympatric friarbird and oriole pop- and Timor) are sharedby P. buceroidesand the ulations, mimicry varies from nearly perfectto smaller O. [bouroensis],and the smaller orioles nonexistent. The details of this variation are do mimic P. buceroides. significant: (4) The New Guinea oriole O. szalayi(102 g), (1) There are six speciesor superspeciesof which mimics the large friarbirds, all sympatric with O. [bouroensis]: P. novaeguineae(135 g), is in turn mimicked P. [moluccensis](species listed in Table 1 plus both in plumageand vocallyby a smallerhon- cockerelli, eichhorni, and albitorques), P. eyeater, Pycnopygiusstictocephalus (43 g). [corniculatus](species corniculatusand dieme- Thus, these casesall conform to a pattern of nensis),P. argenticeps,P. [citreogularis](species smallerbirds mimicking larger birds. citreogularis,brassi, and inornatus),P. meyeri, POSSIBLE EXPLANATIONS OF MIMICRY IN and P. gilolensis,listed here in descendingse- ORIOLES AND FRIARBIRDS quence of size. Only P. [moluccensis]equals or exceedsO. [bouroensis]in weight, and only it What is the selectiveadvantage responsible servesas a modelfor mimicry, eventhough the for the evolution of this size-relatedmimicry? mimic orioles of New Guinea, Halmahera, and Five relevant published theoriesare as follows. April 1982] OrioleMimicry of Friarbirds 191

(1) Stresemann(1914a) dismissed the resem- both New Guinea (pers. obs.) and Buru (Stre- blances as a coincidencedue to parallel evo- semann 1914a), however, the mimics and lution of orioles and friarbirds in the same en- modelsnever travel together,even when feed- vironments, such as evolution of darker forms ing in the same tree; they fly in and out in- in wetter climates(Gloger's rule). Of the eight dependently. islands shared by O. [bouroensis]and P. (4) Is it possiblethat friarbirds are distasteful [moluccensis],seven have elevations exceeding and that orioles are Batesian mimics? This 1,400 m, while one (Tenimbar) is a low coral seems unlikely, because New Guinea men island. Annual rainfall in the lowlands, where working with me broiled and ate without ob- oriolesand friarbirdslive, is 200 cm or higher jection several dozen friarbirds and orioles, on New Guinea, Halmahera, Ceram, and Buru, althoughthey did objectto the tasteof certain lower on Timor, Wetar, and Tenimbar, and other bird species(e.g. Megapodiusfreycinet, lowest (down to 60 cm) in some areas of Aus- Centropusmenbeki). In addition, if friarbirds tralia occupiedby these species.Stresemann's were distasteful and orioles not distasteful, it explanationcould perhaps contribute to an un- would remain a puzzle why Pycnopygiusstic- derstandingof interislanddifferences in over- tocephalusmimics an oriole. all darknessof plumage,though the correlation (5) Wallace (1863, 1869) suggestedthat ori- is quite imperfect: the palest forms are on two olesmimic the larger,pugnacious friarbirds to of the drier islands (Wetar and Timor), but the escapeattack by hawks. Stresemann(1914a, b) New Guinea forms are as pale as or paler than found bird-eating hawks so rare on Buru and those on much drier Tenimbar, and the driest Ceram that he consideredthis explanation ab- environment(Australia) has one of the palest surd. I draw the same conclusion for New P. [moluccensis]forms but the most colorful O. Guinea. In the New Guinea lowlands there are [bouroensis]form. Apart from these doubtful few widespread speciesof hawks likely to at- correlationsbetween aridity and paleness,the tack adult birds, and most of them are very detailedsimilarities in head patternsof orioles uncommon or rare. Only once did I see a bird- and friarbirds, and the correlation between eating hawk (the uncommonMegatriorchis do- degree of mimicry and size, make an expla- riae) near a tree frequented by orioles, friar- nation basedon coincidenceimplausible. birds, and other honeyeaters,and I never saw (2) Cody (1969) described numerous in- a hawk attack a bird in such a tree. stances in which unrelated but ecologically Nevertheless,one can observeattacks among similar bird speciesconverge on each other orioles, friarbirds, and other honeyeaterspe- and thereby can more readily maintain inter- cies dozens of times daily, and these attacks specific territories. The mimics and models of provide a strong and obvious selectiveforce New Guinea, Timor, and Tenimbar and the for evolution of size-related mimicry. Orioles, nonmimic orioles and friarbirds of Australia, friarbirds, and Pycnopygiusstictocephalus reg- however, are neither interspecificallynor in- ularly occur together, often in the same tree, traspecificallyterritorial. I often observed P. becausethey and other honeyeatersbelong to novaeguineae,O. szalayi, and Pycnopygiusstic- a guild of bird speciesthat gatherin flowering tocephalusfeeding and singingtogether in the and fruiting trees to feed. The three New same tree and up to 10 individuals of P. no- Guinea specieshave largelycoincident habitat vaeguineaesinging simultaneouslyin the same preferences(forest edge, clearingswith scat- tree. Finally, the oriole-friarbird casedoes not tered trees, occasionallyforest), perch height involve convergence(evolution of two taxa to- ranges (canopydown to about 10 m), and al- ward eachother), as in the casesdiscussed by titude ranges(P. novaeguineaeand O. szalayi Cody, but mimicry (evolutionof oriolestoward from 0 up to 600-900 m, occasionallyhigher in friarbirds, without evolution of friarbirds to- disturbed areas;Pycnopygius stictocephalus up ward orioles). to 400-600 m). The same holds for the mimic (3) Moynihan (1968)has noted numerousin- orioles and friarbirds of Ceram and Buru. Ori- stanceswherein neotropical bird species that oles, friarbirds, and Pycnopygiusstictocephalus habituallytravel together in foragingflocks re- also show broad overlap in their diets of fruit, semble each other. He postulatedthat this re- insects,and nectar, althoughthere are quan- semblance contributes to flock cohesion. On titative differencesin diet (possiblyrelated to 192 JAREDM. DIAMOND [Auk, Vol. 99 differencesin bill shape)that make coexistence zomelaadolphinae) that they couldbarely alight ecologicallypossible (e.g. analysesof stomach in the tree before being attacked; the three contentsby me and other collectorsyielded: 72 honeyeaters spent most of this time in another P. novaeguineae,insects in 33%, fruit in 67%; tree only 5 m away. At the same time Pycno- 20 O. szalayi,insects in 10%, fruit in 90%; nec- pygiusstictocephalus was ignoring five species tar not detectable in stomach). with quite different diets (the similar-sized There have been numerous studies of this kingfisher Halcyon sanctaand whistler Collur- guild of fruiting and flowering tree specialists icincla harmonica and the smaller warbler Ge- in New Guinea (e.g. Mayr and Rand 1937;Rip- rygone olivacea, flycatcher Myiagra rubecula, ley 1959, 1964; Terborgh and Diamond 1970; and mistletoe-bird Dicaeumpectorale), as well Beehler 1980; Pratt in press). The latter three as five larger speciesthat do overlapwith it in studies included systematic measurement of diet (O. szalayi,P. novaeguineae,the bird usageand of interspecificand intraspecif- Sphecotheresvieilloti, pigeon Ptilinopusiozonus, ic aggressionin the trees. Every one of these and bowerbird Chlamyderacerviniventris). studies stressedthe high frequencyof fighting On each island shared by P. [moluccensis] among guild members gathered in trees to and O. [bouroensis],P. [moluccensis]is the larg- feed. The size sequenceof bird consumersde- est member of the guild, and O. [bouroensis]is fines a dominance hierarchy: large species the secondlargest, while Pycnopygiussticto- often drive smaller speciesaway from food re- cephalusis the third largest on New Guinea. sources.The fighting is adaptive, in the sense Virtually all authors who have commented on that each speciesdevotes time and energy to the habits of friarbirds emphasizetheir pug- driving off only those other speciesthat over- nacity in driving off other birds (see e.g. Hill lap with it in diet and that are not larger(hence 1967, Bell 1969, Frith 1976, Peckover and File- can be safely attacked). Beehler (1980: 516) de- wood 1976, and Pizzey 1980 for P. novaegui- scribesthese New Guinea feedingassemblages neae, and Wallace 1863; Layard and Layard as "veritable riots of interindividual aggres- 1882; Siebers1930; Ripley 1959;Rand and Gil- sion. In general, the organizing factor among llard 1967; Hill 1967; Frith 1969, 1976; Pizzey the New Guinea birds seemed to be domi- 1980; and Stokes 1980 for P. moluccensis,P. nance hierarchy, based on size and aggres- corniculatus,P. citreogularis,P. diemenensis,P. siveness.The larger specieswere usually more gilolensis,and P. meyeri).Although I too often successful;thus, they occupied the favored saw P. novaeguineae,as well as O. szalayiand feeding spotswith minimum harassment.But, Pycnopygiusstictocephalus, driving off smaller even under the best of circumstances, the dom- species, it is remarkable that ! never saw P. inant speciesin a tree spent most of the time novaeguineaeattack the smaller O. szalayi, nor supplanting and chasing conspecifics and O. szalayi attack the smallerPycnopygius stic- smaller heterospecifics."Pratt (in press) tabu- tocephalus.This is despite the fact that I looked lated attacks by color-banded frugivores on particularly for such an interaction during a other birds in fruiting trees;the attackersregu- total of about 500 h of observationat fruiting larly succeededin chasingoff other birds that and flowering trees, spreadover 19 months of were not too large. fieldwork in many different areas of New As an illustration of the adaptiveness of Guinea, where these three specieswere com- fighting, considerthe following caseinvolving mon and observed daily. These speciesregu- Pycnopygiusstictocephalus, as it has been the larly shared the same feeding trees with each subjectof many fewer publishedstudies than other, foragedwithin 1-2 m of each other, and the more common P. novaeguineaeor O. sza- attacked other speciesfrequently, but ignored layi. In a fruiting tree at Sogeri(450 m, south- each other. In contrast, in northeast Australia, east New Guinea), which I observed for a total where the nonmimic O. fiavocinctusmeets the of 4.5 h on 6-7 September 1979, Pycnopygius southern edge of the range of the somewhat stictocephalusrepeatedly and so successfully larger P. novaeguineae,I did observethe friar- drove off four species overlapping with it in bird attack and drive off the oriole. [These at- diet (the similar-sized -shrike Coracina tacks may not interfere seriouslywith the ori- papuensisand the smaller honeyeaters Meli- ole, becausethis friarbird is considerablyless phagafiavescens, obscura, and My- common than the oriole in northeast Australia April1982] OrioleMimicry ofFriarbirds 193 and differs somewhat in preferred foraging the orioles were able to defend themselves and stratum and diet, whereas in New Guinea P. maintain their place in feeding trees. Where novaeguineaeis more commonthan its mimic smaller orioles encounteredlarger friarbirds, O. szalayi,which is in turn morecommon than however, the orioles, as the member of the its mimic Pycnopygiusstictocephalus. The Aus- feeding guild next in size to the friarbirds, tralian orioles may instead be mimics of fig- were the prime target of attack by friarbirds, birds (Beland 1977)]. In southeast Australia, and this must have interfered seriously with where the nonmimic O. sagittatus broadly their accessto food. By convergingin appear- sharesits rangewith the similar-sizedfriarbird ance on the larger friarbirds, the orioles re- P. corniculatus,the oriole stands its ground duced the risk of these attacks and may also when attackedby the friarbird, and sometimes have gained status with respect to smaller the orioleis the aggressorand attacksthe friar- birds. The larger the size advantage of the bird. friarbird over the oriole, the more perfect had It seems clear, then, that mimicry serves at to be the plumageresemblance of the oriole to least one purpose and possibly two. First, the friarbird, as the oriole becameincreasingly mimics escapeattack by larger model species unableto assertitself against increasingly large that might otherwisedrive themoff. It remains aggressors.(Even the largestfriarbird is only unclear, however, just why the model refrains 25% greaterin linear dimensionsthan its mim- from attackingthe mimic. It may be that the ic oriole, so that this size difference does not model regardsconspecifics as equal opponents make it easyto distinguishthe oriolefrom the and hence more dangerous to attack than friarbird, althoughthe twofold weight differ- smaller birds recognized as heterospecific. encedoes seal the outcomeof fights.)Evolution Both friarbirds and orioles often tolerate con- of mimicry may have proceededvia retention specificsin the vicinity. Nevertheless,attacks of juvenal plumage as an adult: among Afro- on conspecificsdo sometimesoccur, and one Asian orioles, the dull, streakedjuveniles are might thus expectmimics to enjoy only partial more similar than are the bright yellow and rather than nearly completeimmunity. A clue black adults to friarbirds, especiallyto juvenile to this puzzle is that O. szalayiis more similar friarbirds, which sometimes have streaks and to juvenile than to adult P. novaeguineae,be- a yellow wash. causejuveniles are more streakedand lack a From the perspectiveof biologicalmimicry, bill knob, and that the black bill of Pycnopygius what is interestingin this caseis that the mod- stictocephalusmakes it more similar to black- el itself is the signalreceiver: the victim of the billed juvenilesthan to red-billedadults of O. deceit is the friarbird itself (or the oriole in the szalayi. Adult birds of many speciesare more case of Pycnopygiusstictocephalus as mimic), tolerant of conspecificsin juvenal or subadult rather than some third species. Below I men- plumag•than in adult plumage(Hardy 1974; tion other suggestivelysimilar cases. Rohwer 1978; Immelmann 1980:160 and pers. comm.; Rohwer et al. 1980). A secondpossible benefit of mimicry is that the mimic, by resem- TESTS AND UNSOLVED PROBLEMS bling a larger bird, may derive higher status in the eyesof smallerspecies and may succeed I believe that study of specimensand field in scaringthem off with lesseffort by its mere observationsin New Guinea presentlywarrant appearance. two conclusions. First, orioles are indeed mim- To summarizethe hypotheticalevolutionary ics of friarbirds, as Wallace postulated over a historyof thesespecies, the Afro-Asiangenus century ago. The case for mimicry is much Oriolusand the genusPhilemon of the Austra- stronger than Wallace realized: he had seen lian region convergentlyevolved ecological only two of the eight setsof populationsthat and morphologicalsimilarities, although their we now know. Second, at least in New Guinea adult plumageswere quite different.One Ori- a plausibleselective force is the "riots of in- olus stock then invaded the Australian region terindividual aggression" (to use Beehler's and encountered friarbirds. All friarbird phrase) that occur in the trees where orioles species are notoriously belligerent. Where and friarbirds feed together. It is obvious, larger oriolesencountered smaller friarbirds, however, that a host of questions remains un- 194 JAREDM. DIAMOND [Auk, Vol. 99 solved and that critical tests remain to be per- find orioles and friarbirds more confusingly formed. Some of these are as follows: similar in the field than one would guessfrom (1) Most important, we need quantitative the frontispiecebecause of similarity in pos- studies of the interactions (and noninterac- ture, movements,and flight, and Wallacemade tions) of orioles, friarbirds, and other guild the same comment for Bum. Are mimic orioles members. There are no published field obser- more similar to friarbirds in theserespects than vations at all from Halmahera and Wetar for are nonmimic orioles or orioles outside the these species; from Buru, Ceram, Tenimbar, range of friarbirds? How extensive is vocal and Timor, only scanty ones. Even for New mimicry, reported for New Guinea and Buru? Guinea and Australia, the most accessible of (5) Are there other examples of visual mim- these eight islands, further quantitative stud- icry within the guild that includes orioles and ies are needed. friarbirds? Possible examples involve Austra- (2) I postulatedthat mimicryyields two dis- lian oriolesand (Beland1977), and the tinct advantages: in reducing attacks by the New Guinea honeyeatersMeliphaga fiaviventer larger model, and in more easilycowing small- philemonand Philemonmeyeri. er guild members.What is the relative impor- tance of these two factors? OTHER POSSIBLE EXAMPLES OF (3) How do severalnonmimic orioles manage SIZE-RELATED VISUAL MIMICRY to coexistwith large friarbirds (seenext para- Ii',l BIRDS graph)? We have only scanty information for one area, northeast Australia. Especially in the tropics, there are other Questions2 and 3 could be approachedby suggestivecases involving larger and smaller manipulative experimentsand by natural ex- birds that have strikingly similar plumage and periments. Examples of manipulative experi- that may conceivablybe involvedin mimicry. ments are to trap mimics, modify them to re- Previous authors have noted, and been puz- semble models, and release them, as has been zled by, severalof these cases(Hall et al. 1966, done for mimic butterflies. If one paints the Diamond 1972, Haffer 1974). My purpose in bill of Pycnopygiusstictocephalus reddish, or assemblingthe following list of such casesis glues a knob onto the bill of an oriole coexist- to stimulate evaluation of them. This list is un- ing with a friarbird whose adult but not ju- doubtedlya heterogeneousone requiring case- venile is knob-billed, or bleachesthe head pat- by-case study to decide whether mimicry is tern of Pycnopygiusor of an oriole (cf. Rohwer involvedat all and, if so,what the underlying 1978), does the former mimic become the sub- selective force is. ject of friarbird attacks,and do smaller guild Instances that involve parallel geographic members avoid it less than before? As one ex- variation, as in the case of Philemon and Ori- ample of a natural experiment, do the mimic olus, include the following (the larger species female and nonmimic male of O. viridifuscus is named first in each case): the New Guinea on Timor differ in their spatial overlap with birds of paradiseParotia [sefilata]and Lophor- friarbirds and in the frequency with which ina superba [Diamond 1972: 313; cf. Strese- they are attackedby friarbirds and avoided by mann's (1934) misidentification and subse- smaller guild members?As further natural ex- quent descriptionof the new Lophorinarace L. periments, one could ask the same questions s. pseudoparotia];African bush-shrikesof gen- in coastal south-central New Guinea, where era Malaconotusand Chlorophoneus(Hall et al. one mimic oriole (O. szalayi)and two nonmim- 1966); the neotropical toucans Ramphastos ic orioles(O. sagittatusand O. fiavocinctus)co- [tucanus] and R. [dicolorus](Haffer 1974: 252); exist with one larger and three small friarbird and the neotropicalcotingids Lipaugus vocifer- species, and in most of New Guinea, where ans-L.unirufus, tyrannids Rhytipterna simplex- the mimic Pycnopygiusstictoce- R. holerythra,and tyrannids Laniocerahypo- phalusand the similar-sized,nonmimic honey- pyrrha-L. rufescens.Instances not involving eater Meliphagafiaviventer often sharefeeding parallel geographicvariation include: the neo- trees with larger orioles. tropical flycatchersMegarhynchus pitangua, Pi- (4) The frontispiece documents plumage tangussulphuratus, P. lictor, Myiozetetescaya- mimicry. I and other observersin New Guinea nensis, M. similis, and Conopias parva (for April 1982] OrioleMimicry of Friarbirds 195 illustrationssee plate 19 of Ridgley 1976 or ACKNOWLEDGMENTS plate 28 of Meyer de Schauenseeand Phelps It is a pleasureto expressmy debt to the National 1978);the neotropicalmotmots Baryphthengus GeographicSociety, Lievre Fund, and World Wild- ruficapillusand Electronplatyrhynchum; the life Fund for support;to Martin Cody, ThomasHow- neotropical Cacicuscela and tanager ell, and ThomasSherry for suggestionson the manu- Ramphoceluspasserinii, and the icterid C. uro- script;and to moreresidents of the areasvisited than pygialisand tanagerR. icteronotus;the African can be mentioned by name for making the field work drongo Dicrurusadsimilis and flycatcherMe- possibleand for information. laenornispammelaina ; the African shrike Lanius collarisand flycatcherSigelus silens; and the LITERATURE CITED Melanesianparrots Domicella chlorocercus and BEEHLER,B. 1980. A comparisonof avian foraging Vini margarethae. at flowering trees in Panamaand New Guineaß Two factorsmake the argumentthat the de- Wilson Bull. 92: 513-519. tailed resemblance of orioles and friarbirds in- BELAND,P. 1977. Mimicry in oriolesof southeastern volves mimicry rather than coincidence a Queensland. Emu 77: 215-218. BELL, H. L. 1969. Field notes on the birds of the Ok strongone: the resemblanceis betweentaxa in Tedi River drainage, New Guinea. Emu 69: two distant families, and the resemblance is 193-211. maintained through parallel geographic.varia- COPY, M. L. 1969. Convergent characteristicsin tion of sevenpairs of forms. None of the cases sympatric species:a possible relation to inter- mentioned in the preceding paragraph is as specificcompetition and aggressionßCondor 71: compelling.Perhaps the mostsuggestive is the 222-239. example involving cotingids and tyrannids, DIAMOND,J. M. 1972. Avifauna of the EasternHigh- becauseit involvesnot only formsfrom differ- lands of New Guinea. Cambridge, Massachu- ent families but also parallel variation (in two setts, Nuttall Ornithol. Club. rather than seven areas) and is further FORBES,H. O. 1885. A naturalist'swanderings in the strengthenedby the fact that the parallelvari- EasternArchipelagoß New York, Harper. FRITH, H. J. 1969. Birds in the Australian High ation involves three setsof taxarather than just Country. Sydney,Reed and Reed. two, as with orioles and friarbirds. The ex- ß 1976. Complete book of Australian birds. amplesinvolving bush-shrikes,birds of para- Sydney, ReadersDigest Services. dise, and toucansare strengthenedby parallel HAFFER,J. 1974. Avian speciationin tropical South geographicvariation (over at leastseven pairs America. Cambridge, Massachusetts, Nuttall of forms in the bush-shrikes, four in the tou- Omithol. Club. cans),but the formsbelong to the samefamily HALL, B. P., R. E. MOREAU, & I. C. J. GALBRAITH. (or genus in the case of the toucans).The re- 1966. Polymorphism and parallelism in the Af- maining nine exampleslack supportfrom par- rican bush-shrikesof the genusMalaconotus (in- allelgeographic variation; four involvespecies cluding Chlorophoneus).Ibis 108: 166-182. HARDY, J. W. 1974. Behavior and its evolution in in different families, five involve confamilial neotropical jays (Cissilopha).Bird-banding 45: species.One could argue that the confamilial 253-268. examplesneed not involve biologicalinterac- HILL, R. 1967. Australian birds. Melbourne, Nelson. tions: the two formsmay merelyhave retained IMMELMANN, K. 1980. Introduction to ethology. a pattern that is their shared inheritance. For New York, Plenum. instance,in a similar exampledrawn from Af- KEAST,A. 1976. The origins of adaptive zone utili- rican mammals, the resemblance between the zationsand adaptive radiations, as illustratedby hyaenaHyaena hyaena and the aardwolfProte- the AustralianMeliphagidae. Pp. 714•2 in Proc. 16th Intern. Ornithol. Congr. (H. J. Frith and J. les cristalus is sometimes dismissed on these H. Calaby, Eds.). Canberra, Australian Acad. grounds.Whether the interpretationbased on Sci. sharedinheritance is moreplausible than that LAYARD, E. L., & E. L. C. LAYARD. 1882. Notes on basedon biologicalinteraction requires careful the avifauna of . Ibis 6: 493. evaluation,based in part on whetherthe close- MAYR, E., & A. L. R•ND. 1937. Results of the Arch- bold Expeditions. 14. Birds of the 1933-1934 ness in field charactersis greater than one Papuan expedition. Bull. Amer. Mus. Nat. Hist. would expectfrom the closenessin geneticher- 73: 1-248. itage. MEYER DE SCI-IAUENSEE,R., & W. H. PHELPS, JR. 196 JAREDM. DIAMOND [Auk, Vol. 99

1978. A guide to the birds of Venezuela. Prince- , S. D. FRETWELL, & D. M. NILES. 1980. De- ton, Princeton Univ. Press. layed maturationin passerineplumages and the MOYNIHAN,M. 1968. Social mimicry; charactercon- deceptiveacquisition of resources.Amer. Natur. vergenceversus characterdisplacement. Evolu- 115: 400-437. tion 22: 315-331. SIEBERS, H. C. 1930. Fauna Buruana. Treubia 7 PECKOVER,W. S., & L. W. C. FIREWOOD.1976. Birds Suppl.: 164-303. of New Guinea and tropical Australia. Sydney, STOKES, T. 1980. Notes on the landbirds of New Reed & Reed. Caledonia. Emu 80: 81-86. PIzzE¾,G. 1980. A field guide to the birds of Aus- STRESEMANN,E. 1914a. Beitr•ge zur Kenntnis der tralia. Princeton, Princeton Univ. Press. Avifauna von Buru. Novitates Zool. 21: 358-400. PRATT,T. K. In press. Examplesof resourcedefense --. 1914b. Die V6gel von Seram(Ceram). Nov- by tropical frugivorous birds. Condor. itates Zool. 21: 25-153. RAND, A. L., & E. T. GILLIARD. 1967. Handbook of --. 1934. Vier neue Unterarten von Paradies- New Guinea birds. London, Weidenfeld, Nicol- vogeln.Ornitol. Monatsber.42: 144-147. son. TERBORGH,J., •r J. M. DIAMOND. 1970. Niche overlap RIDGEL'V,R. S. 1976. A guide to the birds of Panama. in feeding assemblagesof New Guinea birds. Princeton, Princeton Univ. Press. Wilson Bull. 82: 29-52. RIPLE¾,S. D. 1959. Competition between WALLACE, A. R. 1863. collected in the and honeyeaterspecies in the Moluccanislands. island of Bouru (one of the Moluccas), with de- Amer. Natur. 93: 127-132. scriptions of the new species. Proc. Zool. Soc. --. 1964. A systematicand ecologicalstudy of London: 18-28. birds of New Guinea. PeabodyMus. Nat. Hist. --. 1869. .New York, Bull. 19. Dover Publ. (1962 reprint). ROHWER,S. 1978. Passerinesubadult plumages and WICKLEa,W. 1968. Mimicry in plants and . the deceptiveacquisition of resources:test of a London, Weidenfeld & Nicolson. critical assumption. Condor 80: 173-179.