Hawk Mimicry and the Evolution of Polymorphic Cuckoos

Home , Accipiter, Apostatic selection, Coccycua, Neomorphus

Chinese Birds 2013, 4(1):39–50 ORIGINAL ARTICLE DOI 10.5122/cbirds.2013.0002

Hawk mimicry and the evolution of polymorphic cuckoos

Rose THOROGOOD , Nicholas B. DAVIES

Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, U.K.

Abstract The resemblance of some parasitic cuckoos to Accipiter hawks has been known since ancient times. Recent experiments show that the hawk-like features of Common Cuckoos (Cuculus canorus) facilitate access to Reed Warbler (Acrocephalus scirpaceus) host nests. However, social information alerts hosts to see through the cuckoo’s mimetic disguise. In turn, this has promoted the evolution of a cuckoo polymorphism to thwart host recognition. Here we show by comparative analyses that parasitic cuckoos with hawk-like features (yellow eyes, barred underparts, yellow legs) are more likely to be polymorphic (29% of species) than those without (8% of species). Phylogenetic analyses conﬁrm correlated evolution of hawk-like features and cuckoo polymorphism. We suggest that mimicry dynamics are particularly likely to promote the evolution of various guises in parasitic cuckoos to beat host defences.

Keywords Batesian mimicry, brood parasitism, phylogenetic analysis, plumage morph, polymorphism

Introduction bergen and Davies, 2011). Reed Warblers (Acrocephalus scirpaceus), a common host in Europe, are much more The uncanny resemblance of Old World cuckoos to likely to mob a Common Cuckoo (Cuculus canorus) Accipiter hawks has been remarked upon since ancient model if its barred underparts are covered (Welbergen times, with the metamorphosis of cuckoos to Accipiter and Davies, 2011), and tits (Paridae) who are naïve to hawks and vice versa prevalent in old Norse, Greek and cuckoo parasitism are as afraid of Common Cuckoos Chinese mythology (Lai, 1998). Most Accipiter hawks as sparrowhawks (Accipiter nisus), unless the cuckoo’s have barred underparts, yellow feet, and yellow eyes or barring is obscured (Davies and Welbergen, 2008). Of yellow eye ring (del Hoyo et al., 1994), and some cuck- course, barring is unlikely to be the only salient cue that oo species share these features (Voipio, 1953; Payne, causes confusion (Trnka et al., 2012), with the yellow 1967; Johnsgard, 1997). However, it was not until Wal- eye ring, legs, and flight behavior all probably part of lace (1889) that this similarity was suggested to be a the disguise. But, for Reed Warblers at least (cf. Trnka mimetic defence. Wallace suggested that a cuckoo mim- and Prokop, 2012), we now know that these hawk-like ics hawks for protection from the hawks themselves, features are an effective disguise that facilitates access to and indeed, cuckoos suffer less from hawk-predation host nests because of the risk to hosts of incorrect iden- than expected by chance (Møller et al., 2012). However, tification of a dangerous predator. recent work has demonstrated that this disguise can Appearing similar to a harmful hawk to gain access also be another feature in the arms race between brood to Reed Warblers’ nests is best thought of as an example parasite and host (Davies and Welbergen, 2008; Wel- of Batesian mimicry (Fig. 1): a species that could be safely attacked uses the disguise of a harmful model to avoid interactions with a dupe. According to definitions Received 11 December 2012; accepted 12 January 2013 developed largely with predators and prey in mind (Ruxton et al., 2004), a dupe avoids a Batesian mimic, Author for correspondence (Rose Thorogood) even though the interaction would be harmless, because E-mail: [email protected] the mimic has copied the signals or cues given by a

Fig. 1 Reed Warblers (Acrocephalus scirpaceus) are duped by the Common Cuckoo’s (Cuculus canorus) mimicry of its sparrowhawk (Accipiter nisus) model. Batesian mimicry, like this, can become less effective (dashed lines) if the relative risk of the model changes, or new information becomes available. Alternate forms like the rufous cuckoo morph may then evolve. dangerous model that elicit innate or learned fear in the hawk-like appearance, it occurs because of the model’s dupe. Thus, the cuckoo, which can be safely attacked by deterrence. hosts, is a “parasite in wolf’s clothing” (Welbergen and Davies, 2011). This contrasts with aggressive mimicry, Model-mimic dynamics could lead to alternative when an opponent resembles a non-threatening, or disguises inviting, model to gain access to prey or hosts (Wick- ler, 1968). A cuckoo that gained access to host nests by A cuckoo’s disguise is not only useful for deterring mimicking a harmless dove, for example, would be an mobbing. The sight of a cuckoo at the nest can make aggressive mimic. Unlike many Batesian mimics (e.g. hosts more likely to reject foreign eggs (Davies and tasty butterflies that mimic distasteful models), cuckoos Brooke, 1988; Moksnes et al., 1993), but only when are not harmless to duped hosts; they are predators of cuckoos are common (Brooke et al., 1998). This means eggs and chicks (Gärtner, 1981) and for many hosts that as cuckoos increase in the environment, informa- the loss of reproductive investment that comes from tion about their disguise will spread and alert multiple raising a parasite can severely impact their fitness (re- lines of host defences (Davies and Welbergen, 2009; viewed in Welbergen and Davies, 2011). However, while Campobello and Sealy, 2011; Thorogood and Davies, mistaking a cuckoo for a hawk could result in the loss 2012). Environmental changes may also influence the of the Reed Warbler’s reproductive success that season success of the cuckoo’s mimicry. For example, spar- (unless it later recognizes and correctly rejects an im- rowhawk numbers may vary as prey availability alters, poster’s egg), mistakenly approaching a hawk could be increasing or decreasing the risk of parasitism relative deadly (Newton, 1986). It is not the cost of the mimic to the risk of predation. Therefore, mimetic defences to the host that defines mimicry, but the cost of incor- may become less successful if models become more rare rect identification of the model (Pasteur, 1982; Pough, or less dangerous, or if dupes evolve new behaviors or 1988). Therefore, when hosts are duped by the cuckoo’s acquire more information that allow improved recog- www.chinesebirds.net Rose Thorogood and Nicholas B. Davies. Hawk mimicry and evolution of polymorphic cuckoos 41 nition. How might cuckoos fight back against a loss of cuckoo species and ask if cuckoos with hawk-like fea- mimicry’s effectiveness? tures are more likely to be polymorphic. In this analysis Theory and models suggest that when mimetic de- we mapped the presence of three hawk-like features fences become less effective, alternative forms may be (barred plumage, yellow eye ring, and yellow feet) on selected if they escape the dupe’s recognition (Turner, to the cuckoo phylogeny of Sorenson and Payne (2005) 1978; Vane-Wright, 1979; Holen and Johnstone, 2004). and scored each species as mono- or polymorphic. These alternate forms may then evolve to become mim- Barred plumage in cuckoos could represent convergent ics of other models, and/or vary in frequency depend- evolution for crypsis to avoid host detection rather than ing on the success of the primary morph. Analysis of hawk mimicry (Krüger et al., 2007), but the behavioral evolutionary pathways in Papilionid butterflies provides responses of hosts and non-hosts with cuckoo and good evidence to support this hypothesis; here it seems sparrowhawk mounts suggest that barring is part of the that mimicry promotes the evolution of polymorphic suite of cues required to elicit fearful responses towards species (Kunte, 2009). Polymorphisms may evolve cuckoos (Davies and Welbergen, 2008; Welbergen and whether a mimic is Batesian or aggressive, but models Davies, 2011; Trnka et al., 2012). Bright yellow features show that different forms are more likely to evolve in are unlikely to aid crypsis. Without behavioral tests Batesian mimic species (Holen and Johnstone, 2004). we cannot be sure that hawk-like features are used as Alternatively, multiple forms can evolve in non-mi- mimetic defences by all species scored as such (Grim, metic species simply because rare forms are more likely 2005), however this should only make our analyses to evade detection (apostatic selection, Clarke, 1969; more conservative. We predict that if polymorphisms Endler, 1981; Bond, 2007). are linked to hawk-mimicry, then a greater proportion Although rare in birds (3.5% of species, Roulin, of species exhibiting hawk-like features will be poly- 2004), polymorphisms are common among birds of morphic than species without hawk features. However, prey, including the hawks with which some cuckoos if polymorphisms are only a result of apostatic selec- share features (11 of 46 Accipiter hawk species are poly- tion, then we predict that polymorphisms will be just as morphic, Ferguson-Lees and Christie, 2001). Polymor- likely to arise in cuckoo species that do not show hawk- phisms in these birds may sometimes be adaptive (Rou- like features. lin, 2004) as they are often linked to alternative strategies within species (ecological niches, mating strategies, Methods crypsis, and avoidance of prey or predators; Fowlie and Krüger, 2003; Galeotti et al., 2003). Many cuckoo spe- Data collection cies also come in multiple plumages and most of these species are parasitic (Payne, 1967; Galeotti et al., 2003; We collected data from species descriptions (Payne, Payne, 2005, see Results below). Previous hypotheses 2005) for each of the 141 species of cuckoos listed in have focused on alternate cuckoo forms being selected Sorenson and Payne’s (2005) phylogeny. All variables via apostatic selection (Payne, 1967; Galeotti et al., were treated as binary: (1) barring was recorded pres- 2003). Alternatively, polymorphisms may be more ent if it occurred on the breast or belly, (2) legs and feet likely to evolve in parasitic cuckoos if the effectiveness were scored as yellow if they were described as such, of hawk mimicry declines (Fig. 1). Common Cuckoos and (3) eye colour was scored as yellow if either the iris come in two colour morphs: a grey morph that resem- or eye ring included yellow in the description. Plumage bles sparrowhawks (and is the more commonly studied morphs were assigned as polymorphic if species were form), and a rufous morph (Voipio, 1953; Honza et recorded showing different plumage forms across their al., 2006). Our recent experimental work has shown range (sexually dimorphic plumage was not included), that when Reed Warblers are alerted by social informa- but insufficient data exists to include range size in anal- tion to the presence of either morph, then the alternate yses. For these analyses we only considered variation in morph is more likely to slip past host defences (Thoro- adult plumage as our hypothesis relates to protection good and Davies, 2012). We suggested, therefore, that from hosts during breeding. We took a conservative the Common Cuckoo’s polymorphism is maintained by approach and did not restrict polymorphisms only to frequency-dependent selection. where multiple forms co-occur because of the dynamic Here we use a comparative approach to look across nature of host-parasite interactions — different forms

© 2013 Beijing Forestry University and China Ornithological Society 42 Chinese Birds 2013, 4(1):39–50 could have arisen because of mimicry but are now were largely restricted to those from the Cuculinae, specialized on different hosts or in different habitats except for the three Centropus spp.: 4/83 non-parasitic (Voipio, 1953). Our approach relied on human-based cuckoos (4.8%) are polymorphic compared to 12/58 categories of plumage morphs. Avian vision differs from parasitic species (20.7%). Similarly, all of the 24 species our own, and so our gross classification of morphs may showing barred plumage with yellow eyes and/or feet have under-estimated the true range of plumage varia- were Cuculinae (17%). Of these hawk-like cuckoos, 8 of tion present in some species (Seddon et al., 2010; Burns the 24 were polymorphic (33.3%), so species with hawk and Schultz, 2012; Stoddard, 2012). Our data were features are more likely to be polymorphic than those compared against datasets used in two previous studies without (8/117, 6.8%, two-tailed Fisher’s exact test: p = of plumage in cuckoos (Galeotti et al., 2003; Krüger et 0.001). al., 2007) and concordance was almost complete. When we took phylogeny into account, we found that for each trait there were clear differences in the Data analysis rates of evolution between states. Looking first at each trait separately, species lost hawk-like features more To assess whether plumage polymorphisms have quickly than they gained them (LR = 20.68, df = 1, p < evolved together or independently from hawk-mimicry, 0.001), and polymorphic species became monomorphic we used the DISCRETE module of BAYESTRAITS (Pa- more quickly than they evolved alternate forms (LR gel, 1999; Pagel and Meade, 2006) with the phylogeny = 14.05, df = 1, p < 0.001, results were similar when of Sorenson and Payne (2005). DISCRETE reconstructs branch lengths were equal). When we tested for corre- the evolution of binary traits on a phylogenetic tree lated evolution between these two traits, we found some by assessing the frequency of transitions between the evidence that the evolution of hawk features and poly- various combinations of the two traits, either linked morphisms are linked in cuckoos (Fig. 3, ultrametric or where they occur at random. The log likelihoods tree: LR = 7.74, df = 4, p = 0.040; equal branch lengths: of these two models can then be compared using a LR = 9.54, df = 4, p = 0.020). Species that did not look likelihood ratio (LR) test, with the model with the similar to hawks lost polymorphisms significantly faster significantly greater likelihood being the one that best than they evolved (LR = 10.19, df = 1, p = 0.0008). This explains the current pattern of traits on a tree. Differ- transition was slower for hawk-like species, although ences in the rates of transition between states can be it was not significantly different to the rate at which assessed by restricting rates of evolution to be equal for polymorphisms were acquired (LR = 1.18, df = 1, p = two transitions of interest, and then comparing the log 0.20). Interestingly, hawk features were acquired more likelihood of this model with that of an unrestricted slowly than they were lost by monomorphic species (LR model. We performed separate analyses with either ul- = 8.34, df = 1, p = 0.002), and never acquired after a trametric branch lengths taken from the phylogeny, or polymorphism evolved (Fig. 3). Polymorphic, hawk-like with equal branch lengths to account for uncertainty species never lost their hawk features. These patterns of about the mode of evolution (Grafen, 1989) and esti- correlated evolution were only found if we included the mated transitions using maximum likelihood methods additional hawk features of yellow eyes and/or yellow (25 optimization attempts per model). Significance for feet (21/24 species identified as ‘hawk-like’ had both LR tests was assessed against a chi-squared distribution yellow eyes and feet: 1 species had yellow eyes but not with four degrees of freedom. feet, and 2 species with yellow feet did not have yellow eyes). If we restricted our analyses to plumage barring Results only, a dependent model of evolution was no better than an independent one (ultrametric tree: LR = 4.01, Of the 141 species of cuckoos that we included in our df = 4, p = 0.14; equal branch lengths: LR = 4.09, df = 4, analysis (Fig. 2, only the Cuculinae are shown for ease p = 0.13). of presentation), 16 exhibited at least two different plumage types (11.3 %), and all but four of these were Discussion parasitic cuckoos (Coccyzus minor, Centropus ateralbus, Centropus senegalensis, and Centropus viridis were poly- Our results show that there is not one explanation for morphic but nesting cuckoos). Polymorphic cuckoos the evolution of polymorphisms among cuckoo spe- www.chinesebirds.net Rose Thorogood and Nicholas B. Davies. Hawk mimicry and evolution of polymorphic cuckoos 43

cies as both hawk features and polymorphisms can sometimes arise independently of the other. However, polymorphisms were lost more quickly than they were gained in lineages without hawk features, but not in those that appear similar to hawks. Furthermore, extant hawk-like cuckoos were more likely to be polymorphic than those without hawk features. Together, this sug- gests that both apostatic selection and mimicry dynamics may underlie polymorphisms in cuckoos. Of course, polymorphisms in cuckoos may sometimes also be a neutral or non-adaptive trait (Galeotti et al., 2003). However, we found that for cuckoos unlike hawks, the rate for losing a polymorphism was much greater than the rate at which it evolved (for hawk-like cuckoos the opposite was true). Polymorphisms may be lost because: (1) one morph has a selective advantage over another so that it gradually disappears, (2) random genetic processes mean that the genetic variation responsible for plumage morphs is exhausted, or (3) color morphs may speciate, leading to two differently colored descendent species (Roulin, 2004). Our results suggest that the most stable state for cuckoos unlike hawks is to appear monomorphic, and thus polymorphisms, when they do occur, are unlikely to be neutral for cuckoo species. Field experiments are needed to determine whether apostatic selection is responsible for maintaining polymorphisms in species without hawk-features or, if like birds of prey, they arise as a result of large population sizes (Fowlie and Krüger, 2003), and are maintained via sexual (Fowlie and Krüger, 2003) or disruptive selection (Galeotti et al., 2003). Although less frequent, polymorphisms also occur in non-parasitic cuckoos, and all but one occur in the Centropodinae (4 species overall, Fig. 2). It would be fascinating to look within this group to ascertain why some are polymorphic while their close relatives are not, and whether their reversal in sex roles is important (Andersson, 1995). We found that hawk features are limited to ﬁve genera of parasitic Old World cuckoos (Cuculus, Hierococ- Fig. 2 Phylogeny of the Cuculinae (Old World cuckoos, malkohas, cyx, Cercococcyx, Cacomantis, Chrysococcyx), and that and allies) from Sorenson and Payne (2005). Symbols show which 60% of all barred cuckoo species also had yellow eyes species are hawk-like (barred plumage underparts, yellow eye or and/or feet. Previous analyses of only barred plum- eye ring, and/or yellow legs and feet) and which occur in multiple age in cuckoos have found this to be a derived trait plumage forms across their range (‘polymorphic’). Species without that most likely arose after the evolution of parasitism symbols are neither hawk-like nor polymorphic. Parasitic species are (Krüger et al., 2007). Analyses with all avian families asterisked. Three parasitic species from the Neomorphinae (Tapera have also shown that sexual selection may play a role naevia, Dromococcyx phasianellus, D. pavoninus; none are hawk-like or polymorphic) and three polymorphic species from the Centropo- in the evolution of barred plumage (Gluckman and dinae (Centropus viridis, C. senegalensis, C. ateralbus; none are para- Cardoso, 2010), but this does not seem to be the case sitic or hawk-like) are not shown. for cuckoos (Krüger et al., 2007). Our results show that

Fig. 3 The most likely evolutionary pathways of hawk-like features and polymorphic plumage. Bold and dashed lines show where the rates of transition (log likelihoods, shown next to arrows) were significantly different (p < 0.05, as determined by LR tests), whereas lighter weight lines mean that transitions were as likely to occur in either direction. No state was identified as ancestral or more derived (see results). barred plumage alone does not lead to the evolution of (Johnsgard, 1997; Payne, 2005), and that this genus too polymorphisms, suggesting that yellow hawk-like fea- may use mimicry of aggressive models to intimidate or tures play an additional role (Trnka et al., 2012). distract hosts (Johnsgard, 1997). Alternatively, adults Hawk mimicry is a useful deterrent that aids parasit- of parasitic species unrelated to the Cuculidae might ism, at least for Common Cuckoos and their Reed War- use aggressive mimicry to fool their hosts. Honeyguides bler hosts (Davies and Welbergen, 2008; Welbergen and (Prodotiscus regulus) and Cuckoo Finches (Anomalospi- Davies, 2009, 2011; Thorogood and Davies, 2012). So za imberbis) share visual similarities with innocuous why are hawk features not more widespread amongst species, and flocking with these and their hosts may cuckoo species? Indeed, why do other groups of brood help to disguise their threat (Payne, 1967). Field experi- parasites (e.g. cowbirds) not also take advantage of ments are needed to determine the role of this apparent hawk mimicry? Great Reed Warblers (Acrocephalus mimicry in these groups. arundinaceus) are another frequent host of the Com- Comparative analyses exploring patterns of evolution mon Cuckoo, but they do not seem to be deterred by rely on the assumption that extant species represent the cuckoo’s disguise (Trnka and Prokop, 2012). Instead, diversity randomly. However, if hosts evolve behaviors Great Reed Warblers will attack any threat at their nest, that allow better discrimination (e.g. through social even harmless doves (Honza et al., 2010). This shows learning) and defeat mimetic defences, then the spe- that hawk mimicry may not be an effective disguise cies that represent the tips of our phylogeny’s branches against all hosts, or that some hosts may have learned to may not be representative of the evolutionary past. see past this disguise. Furthermore, hawk-like features When mimetic defences are beaten and alternate forms may not always benefit cuckoos, but entail costs. Swal- arise, species may maintain the original mimetic form. lows (Hirundo rustica) and bushtits (Psaltriparus mini- Alternatively, they may revert to monomorphism by mus) mob cuckoos in a similar manner to raptors (Lyon losing either the mimetic or alternate form (Vane- and Gilbert, in press), so hawk-like features might lead Wright, 1979). These changes in host discrimination to increased aggression throughout the year. and cuckoo appearance may well be microevolutionary, Although not hawk-like, adults of other cuckoo spe- and hence not visible in our phylogeny (Kunte, 2009). cies may also use mimicry to their advantage. Among Therefore, to best understand how mimicry dynam- the most derived genera of cuckoos, it is notable that ics might explain parasitic cuckoo polymorphisms, we Surniculus species lack hawk features. Commonly need to know more of host discrimination across the known as drongo-cuckoos because of their close re- cuckoo phylogeny. By looking at the distribution of semblance to the unrelated drongo, it was previously hawk features and polymorphisms across the cuckoo thought that this was a case of host mimicry (Payne, phylogeny, our results suggest that the relationship be- 1967). However, we now know that drongo-cuckoos tween Reed Warbler dupes, sparrowhawks, and the mul- primarily parasitise small babblers rather than drongos tiple forms of hawk-like Common Cuckoos may not be www.chinesebirds.net Rose Thorogood and Nicholas B. Davies. Hawk mimicry and evolution of polymorphic cuckoos 45 unique to this brood parasite. Future research should Ferguson-Lees J, Christie DA. 2001. Raptors of the World. Chris- explore similarities and differences between hawk-like topher Helm, London, U.K. and not hawk-like cuckoos by asking if some hosts are Fowlie MK, Krüger O. 2003. The evolution of plumage poly- more easily fooled by hawk disguises, if some use differ- morphism in birds of prey and owls: the apostatic selection ent strategies for parasitism (e.g. distraction tactics by hypothesis revisited. J Evolution Biol, 16:577–583. Clamator cuckoos (Soler and Soler, 2000)), or if some Galeotti P, Rubolini D, Dunn PO, Fasola M. 2003. Colour poly- cuckoos simply have no suitable models to copy. morphism in birds: causes and functions. J Evolution Biol, 16:635–646. Acknowledgements We thank all participants of the symposium Gärtner K. 1981. Das Wegnehmen von Wirtsvogeleiern durch for their informative discussions and providing the inspiration den Kuckuck Cuculus canorus. Ornithologische Mitteilungen, for this paper. In particular, we thank Mary Caswell Stoddard 33: 115–131. and Martin Stevens for stimulating debate, three anonymous Gluckman T-L, Cardoso GC. 2010. The dual function of barred reviewers’ helpful comments and Wei Liang, Eivin Røskaft and plumage in birds: camouflage and communication. J Evolution their committee for organizing the symposium and facilitating Biol, 23:2501–2506. these discussions. This research was funded by the Natural Envi- Grafen A. 1989. The phylogenetic regression. Philos T Roy Soc B, ronment Research Council. 326:119–157. Grim T. 2005. Mimicry vs. similarity: which resemblances be- References tween brood parasites and their hosts are mimetic and which are not? Biol J Linn Soc, 84:69–78. Andersson M. 1995. Evolution of reversed sex roles, sexual size Holen Ø, Johnstone R. 2004. The evolution of mimicry under dimorphism, and mating system in coucals (Centropodidae, constraints. Am Nat, 164:598–613. Aves). Biol J Linn Soc, 54:173–181. Honza M, Procházka P, Šicha V, Požgayová, M. 2010. Nest defence Bond AB. 2007. The evolution of color polymorphism: crypticity, in a cuckoo host: great reed warblers risk themselves equally searching images, and apostatic selection. Annu Rev Ecol Evol for their own and parasitic chicks. Behaviour, 147, 741–756. S, 38:489–514. Honza M, Šicha V, Procházka P, Ležalová R. 2006. Host nest Brooke M de L, Davies NB, Noble DG. 1998. Rapid decline of defense against a color-dimorphic brood parasite: great reed host defences in response to reduced cuckoo parasitism: be- warblers (Acrocephalus arundinaceus) versus Common Cuck- havioural flexibility of Reed Warblers in a changing world. P R oos (Cuculus canorus). J Ornithol, 147:629–637. Soc Lond B, 265:1277–1282. Johnsgard PA. 1997. The Avian Brood Parasites: Deception at the Burns K, Shultz A. 2012. Widespread cryptic dichromatism and Nest. Oxford University Press, Oxford, U.K. ultraviolet reflectance in the largest radiation of Neotropical Krüger O, Davies NB, Sorenson MD. 2007. The evolution of songbirds: implications of accounting for avian vision in the sexual dimorphism in parasitic cuckoos: sexual selection or study of plumage evolution. Auk, 129:211–221. coevolution? P R Soc B, 274:1553–1560. Campobello D, Sealy SG. 2011. Use of social over personal infor- Kunte K. 2009. The diversity and evolution of batesian mimicry mation enhances nest defense against avian brood parasitism. in Papilio swallowtail butterflies. Evolution, 63:2707–2716. Behav Ecol, 22:422–428. Lai C. 1998. Messenger of Spring and morality: cuckoo lore in Clarke BC. 1969. The evidence for apostatic selection. Heredity, Chinese sources. J Am Oriental Soc, 118:530–542. 24:347–352. Lyon BE, Gilbert GS. In press. Unparasitized species mob and Davies NB, Brooke M de L. 1988. Cuckoos versus reed warblers: alarm call to cuckoos: implications for sparrowhawk mimicry adaptations and counteradaptations. Anim Behav, 36:262–284. by brood parasitic cuckoos. Wilson J Ornithol. Davies NB, Welbergen JA. 2008. Cuckoo-hawk mimicry? An ex- Moksnes A, Røskaft E, Korsnes L. 1993. Rejection of cuckoo (Cu- perimental test. P R Soc B, 275:1817–1822. culus canorus) eggs by meadow pipits (Anthus pratensis). Behav Davies NB, Welbergen JA. 2009. Social transmission of a host de- Ecol, 4:120–127. fense against cuckoo parasitism. Science, 324:1318–1320. Møller AP, Solonen T, Byholm P, Huhta E, Nielsen JT, Tornberg del Hoyo J, Elliott A, Sargatal J. 1994. Handbook of the Birds of R. 2012. Spatial consistency in susceptibility of prey species to the World, Vol. 2. New World Vultures to Guinea Fowl. Lynx predation by two Accipiter hawks. J Avian Biol, 43:390–396. Edition, Barcelona, Spain. Newton I. 1986. The Sparrowhawk. T & AD Poyser, London, U.K. Endler J. 1981. An overview of the relationships between mim- Pagel M, Meade A. 2006. Bayesian analysis of correlated evolu- icry and crypsis. Biol J Linn Soc, 16:25–31. tion of discrete characters by reversible-jump Markov chain

Monte Carlo. Am Nat, 167:808–825. Thorogood R, Davies NB. 2012. Cuckoos combat socially trans- Pagel M. 1999. Inferring the historical patterns of biological evo- mitted defenses of reed warbler hosts with a plumage poly- lution. Nature, 401:877–884. morphism. Science, 337:578–580. Pasteur G. 1982. A classification review of mimicry systems. Trnka A, Prokop P, Grim T. 2012. Uncovering dangerous cheats: Annu Rev Ecol Evol S, 13:169–199. how do avian hosts recognize adult brood parasites? PloS Payne RB. 1967. Interspecific communication signals in parasitic ONE, 7:e37445. birds. Am Nat, 101:363–375. Trnka A, Prokop P. 2012. The effectiveness of hawk mimicry Payne RB. 2005. The Cuckoos. Oxford University Press, Oxford, in protecting cuckoos from aggressive hosts. Anim Behav, U.K. 83:263–268. Pough F. 1988. Mimicry of vertebrates: are the rules different? Turner JRG. 1978. Why male butterflies are non-mimetic: natural Am Nat, 131:S67–S102. selection, sexual selection, group selection, modification and Roulin A. 2004. The evolution, maintenance and adaptive func- sieving. Biol J Linn Soc, 10:385–432. tion of genetic colour polymorphism in birds. Biol Rev, Vane-Wright RI. 1979. Towards a theory of the evolution of but- 79:815–848. terfly colour patterns under directional and disruptive selec- Ruxton GD, Sherratt TN, Speed MP. 2004. Avoiding Attack: the tion. Biol J Linn Soc, 11:141–152. Evolutionary Ecology of Crypsis, Warning Signals and Mim- Voipio P. 1953. The hepaticus variety and the juvenile plumage icry. Oxford University Press, New York. types of the cuckoo. Ornis Fennica, 30:99–117. Seddon N, Tobias JA, Eaton M, Ödeen A, Byers BE. 2010. Human Wallace AR. 1889. Darwinism: An Exposition of the Theory of vision can provide a valid proxy for avian perception of sexual Natural Selection with Some of Its Applications. Macmillan, dichromatism. Auk, 127:283–292. London, U.K. Soler M, Soler J. 2000. Brood-parasite interactions between great Welbergen JA, Davies NB. 2009. Strategic variation in mobbing spotted cuckoos and magpies: a model system for studying co- as a front line of defense against brood parasitism. Curr Biol, evolutionary relationships. Oecologia, 125:309–320. 19:235–240. Sorenson MD, Payne RB. 2005. A molecular genetic analysis of Welbergen JA, Davies NB. 2011. A parasite in wolf’s clothing: the cuckoo phylogeny. In: Payne RB (ed) The Cuckoos. Oxford hawk mimicry reduces mobbing of cuckoos by hosts. Behav University Press, Oxford, U.K. pp 68–94. Ecol, 22:574–579. Stoddard MC. 2012. Mimicry and masquerade from the avian Wickler W. 1968. Mimicry in Plants and Animals. World Univer- visual perspective. Curr Zool, 58:630–648. sity Library, London, U.K.

www.chinesebirds.net Rose Thorogood and Nicholas B. Davies. Hawk mimicry and evolution of polymorphic cuckoos 47

Appendix 1 List of species included in analysis and presence of hawk features (barred plumage, yellow feet or legs, and yellow eye or ring) or polymorphisms. Hawk-like species were barred with at least one of the yellow hawk features. Species Barred plumage Yellow feet or legs Yellow eye or ring Hawk-like Polymorphic Cuculus saturatus 1 1 1 1 1 Cuculus optatus 1 1 1 1 1 Cuculus canorus 1 1 1 1 1 Cuculus rochii 1 1 1 1 0 Cuculus gularis 1 1 1 1 0 Cuculus micropterus 1 1 1 1 0 Cuculus crassirostris 1 1 1 1 0 Cuculus poliocephalus 1 1 1 1 1 Cuculus solitarius 1 1 1 1 0 Cuculus clamosus 1 0 0 0 0 Hierococcyx nisicolor 1 1 1 1 0 Hierococcyx fugax 1 1 1 1 0 Hierococcyx pectoralis 0 1 1 0 0 Hierococcyx hyperythrus 0 1 1 0 0 Hierococcyx sparverioides 1 1 1 1 0 Hierococcyx varius 1 1 1 1 0 Hierococcyx bocki 1 1 1 1 0 Hierococcyx vagans 1 1 1 1 0 Surniculus lugubris 0 0 0 0 0 Surniculus musschenbroeki 0 0 0 0 0 Surniculus velutinus 0 0 0 0 0 Surniculus dicruroides 0 0 0 0 0 Cercococcyx olivinus 1 1 0 1 0 Cercococcyx montanus 1 1 1 1 0 Cercococcyx mechowii 1 1 1 1 0 Cacomantis passerinus 1 1 1 1 1 Cacomantis variolosus 1 1 1 1 1 Cacomantis merulinus 1 1 1 1 1 Cacomantis sonneratii 1 0 1 1 0 Cacomantis castaneiventris 0 1 1 0 0 Cacomantis flabelliformis 0 1 1 0 1 Cacomantis pallidus 0 0 1 0 0 Cacomantis leucolophus 0 0 0 0 0 Chrysococcyx minutillus 1 0 0 0 0 Chrysococcyx meyeri 1 0 0 0 0 Chrysococcyx lucidus 1 0 0 0 0 Chrysococcyx ruficollis 1 0 0 0 0 Chrysococcyx osculans 0 0 0 0 0 Chrysococcyx basalis 1 0 0 0 0 Chrysococcyx megarhynchus 1 0 0 0 0 Chrysococcyx cupreus 1 0 0 0 0 Chrysococcyx flavigularis 1 1 1 1 0 Chrysococcyx klaas 1 0 0 0 0 Chrysococcyx caprius 1 0 1 1 1 Chrysococcyx maculatus 1 0 0 0 0 Chrysococcyx xanthorhynchus 1 0 0 0 0 Urodynamis taitensis 0 0 0 0 0 Scythrops novaehollandiae 1 0 0 0 0

Appendix 1 (continued)

Species Barred plumage Yellow feet or legs Yellow eye or ring Hawk-like Polymorphic Eudynamys scolopacea 1 0 0 0 1 Microdynamis parva 0 0 0 0 0 Pachycoccyx audeberti 0 0 0 0 0 Coccyzus merlini 0 0 0 0 0 Coccyzus longirostris 0 0 0 0 0 Coccyzus vieilloti 0 0 0 0 0 Coccyzus vetula 0 0 0 0 0 Coccyzus pluvialis 0 0 0 0 0 Coccyzus ruﬁgularis 0 0 0 0 0 Coccyzus lansbergi 0 0 1 0 0 Coccyzus erythropthalmus 0 0 1 0 0 Coccyzus americanus 0 0 1 0 0 Coccyzus euleri 0 0 0 0 0 Coccyzus minor 0 0 1 0 1 Coccyzus ferrugineus 0 0 1 0 0 Coccyzus melacoryphus 0 0 1 0 0 Piaya cayana 0 0 0 0 0 Piaya melanogaster 0 0 0 0 0 Coccycua pumila 0 0 0 0 0 Coccycua cinerea 0 0 0 0 0 Coccycua minuta 0 0 0 0 0 Clamator jacobinus 0 0 0 0 1 Clamator levaillantii 0 0 0 0 1 Clamator glandarius 0 0 0 0 0 Clamator coromandus 0 0 0 0 0 Dasylophus superciliosus 0 1 1 0 0 Dasylophus cumingi 0 0 0 0 0 Rhamphococcyx calyorhynchus 0 0 0 0 0 Phaenicophaeus diardi 0 0 0 0 0 Phaenicophaeus tristis 0 0 0 0 0 Phaenicophaeus viridirostris 0 0 0 0 0 Phaenicophaeus pyrrhocephalus 0 0 0 0 0 Phaenicophaeus sumatranus 0 0 0 0 0 Phaenicophaeus curvirostris 0 1 1 0 0 Zanclostomus javanicus 0 0 0 0 0 Taccocua leschenaultii 0 0 0 0 0 Ceuthmochares aereus 0 0 1 0 0 Ceuthmochares australis 0 0 1 0 0 Rhinortha chlorophaea 0 0 0 0 0 Centropus phasianinus 0 0 0 0 0 Centropus bernsteini 0 0 0 0 0 Centropus violaceus 0 0 1 0 0 Centropus bengalensis 0 0 0 0 0 Centropus viridis 0 0 0 0 1 Centropus grillii 0 0 0 0 0 Centropus goliath 0 0 0 0 0 Centropus toulou 0 0 0 0 0 Centropus sinensis 0 0 0 0 0

www.chinesebirds.net Rose Thorogood and Nicholas B. Davies. Hawk mimicry and evolution of polymorphic cuckoos 49

Appendix 1 (continued)

Species Barred plumage Yellow feet or legs Yellow eye or ring Hawk-like Polymorphic Centropus andamanensis 0 0 0 0 0 Centropus nigrorufus 0 0 0 0 0 Centropus cupreicaudus 0 0 0 0 0 Centropus superciliosus 0 0 0 0 0 Centropus monachus 0 0 0 0 0 Centropus senegalensis 0 0 0 0 1 Centropus leucogaster 0 0 0 0 0 Centropus anselli 0 0 0 0 0 Centropus c celebensis 0 0 0 0 0 Centropus steerii 0 0 0 0 0 Centropus rectunguis 0 0 0 0 0 Centropus melanops 0 0 0 0 0 Centropus chlororhynchos 0 0 0 0 0 Centropus unirufus 0 0 1 0 0 Centropus menbeki 0 0 0 0 0 Centropus chalybeus 0 0 1 0 0 Centropus ateralbus 0 0 0 0 1 Centropus milo 0 0 0 0 0 Coua serriana 0 0 0 0 0 Coua delalandei 0 0 1 0 0 Coua gigas 0 0 0 0 0 Coua coquereli 0 0 0 0 0 Coua cursor 0 0 0 0 0 Coua reynaudii 0 0 0 0 0 Coua ruﬁceps 0 0 0 0 0 Coua caerulea 0 0 0 0 0 Coua verreauxi 0 0 0 0 0 Coua cristata 0 0 0 0 0 Carpococcyx renauldi 0 0 0 0 0 Carpococcyx radiatus 1 0 0 0 0 Carpococcyx viridis 1 0 0 0 0 Neomorphus geoffroyi 0 0 0 0 0 Neomorphus radiolosus 1 0 0 0 0 Neomorphus ruﬁpennis 0 0 0 0 0 Neomorphus pucheranii 0 0 0 0 0 Geococcyx californianus 0 0 0 0 0 Geococcyx velox 0 0 0 0 0 Morococcyx ertyhropygus 0 0 0 0 0 Dromococcyx phasianellus 0 0 0 0 0 Dromococcyx pavoninus 0 0 1 0 0 Tapera naevia 0 0 1 0 0 Crotophaga ani 0 0 1 0 0 Crotophaga sulcirostris 0 0 0 0 0 Crotophaga major 0 0 0 0 0

杜鹃模拟鹰及杜鹃外形多态性的进化

Rose THOROGOOD, Nicholas B. DAVIES

（英国剑桥大学动物学系）

摘要：很早人们就已发现一些寄生性繁殖的杜鹃，其外形与 Accipiter 属的鹰类非常相似。最近的实验表明，大杜鹃（Cuculus canorus）模拟鹰类有助于其接近宿主大苇莺（Acrocephalus scirpaceus）的巢。不过，大苇莺却能够通过公众信息传递来识别大杜鹃的这一计谋。反过来，这也导致了杜鹃外形多态性的进化，以避免宿主对杜鹃的识别。通过比较研究，我们发现在寄生性的杜鹃中，模拟鹰类外形（如虹膜黄色，胸腹部具斑纹，脚黄）的杜鹃种类具有明显的外形多态性（29% 的种类），远高于不模拟鹰类外形的杜鹃（仅 8% 的种类具有多态性）。系统发生分析进一步证实了模拟鹰类外形与杜鹃多态性进化之间的相关性。我们认为，杜鹃对鹰类的动态模拟很可能促进了杜鹃外形多态性的进化，以避免宿主对杜鹃的识别。关键词：贝式拟态，巢寄生，系统发生分析，羽毛形态，多态性