Polymorphism at the nestling stage and host-specific mimicry in an Australasian cuckoo-host arms race
Abstract
- Decades of research have shown that the coevolutionary arms race between avian brood parasites and their hosts can promote phenotypic diversification in hosts and brood parasites. However, relatively little is known about the role of brood parasitism in promoting phenotypic diversification of nestlings.
- We review field data collected over four decades in Australia, New Caledonia and New Zealand to assess potential for coevolutionary interactions between the shining bronze-cuckoo (Chalcites lucidus) and its hosts, and how diversification at the nestling stage may be generating different subspecies.
- The shining bronze-cuckoo is a specialist parasite of a few hosts in the family Acanthizidae. It has diversified into subspecies, of which the nestlings closely mimic the respective host nestlings in each region. Additionally, some cuckoo subspecies have polymorphic nestlings.
- The Acanthizidae hosts have similar breeding and nesting habits and only moderately effective frontline defences against parasitism at cuckoo egg laying or at the egg stages. However, some hosts have developed highly effective defences at the nestling stage by recognising and ejecting cuckoo nestlings from the nest. As with the cuckoo nestlings, some hosts have polymorphic nestlings.
- The coevolutionary interactions in each region suggest different evolutionary stages of the arms race in which either the parasite or the host is currently in the lead. The presence of moderately effective defences at the egg laying and egg stages might explain why some hosts do not have defences at the nestling stage.
- The south-Pacific cuckoo – host systems are excellent models to explore the evolutionary mechanisms driving the diversification at the nestling stage in the coevolutionary arms race between avian brood parasites and their hosts.
1 INTRODUCTION
Obligate brood parasites lay their eggs in nests of other species, their hosts, which are left to care for the foreign nestlings incurring a considerable fitness cost (Davies, 2015; Soler, 2017). This can induce an evolutionary arms race in which hosts are under selection to evolve defences against parasitism while brood parasites are under selection to evolve improved strategies to evade host defences (Dawkins & Krebs, 1979). Antagonistic behavioural interactions can often lead to phenotypic diversification (Vamosi, 2011; Yoder & Nuismer, 2010), and indeed decades of research have shown that brood parasitism can promote phenotypic diversification in several aspects of the biology of brood parasites and their hosts (reviewed in Medina et al., 2020). For example, a brood-parasitic lifestyle is associated with traits that facilitate successful parasitism, such as small body size and small eggs (Davies, 2011; Krüger & Davies, 2002), thick eggshells (Stokke et al., 2002) and plastic mating systems (Louder et al., 2019). Similarly, brood parasitism can influence the evolution of host traits such as egg (Caves et al., 2015; Øien et al., 1995; Soler & Møller, 1996) and nestling (Hauber & Kilner, 2007) phenotypes and breeding systems (Feeney et al., 2013).
Polymorphism is a common outcome of the arms race and is generally thought to result from negative frequency-dependent selection driven by host discrimination of the parasite (Cook et al., 2012; Spottiswoode & Stevens, 2012; Takahashi et al., 2010; Tanaka, 2016). For example, a common parasite morph resembling that of the host is an effective way to escape host detection, however a rare parasite morph could also escape detection if the host has not yet learned to recognise it as a parasite. Relative morph frequencies can thus oscillate over time depending on the host's discrimination ability or perception of local risk (Davies et al., 1996; Ruiz-Raya & Soler, 2020). For example, there is experimental evidence from adult common cuckoos Cuculus canorus that negative frequency dependent selection occurs on polymorphisms in parasite plumage (Mappes & Lindström, 2012). Females have two distinct plumage phenotypes, a grey and a rufous or hepatic form (Honza et al., 2006; Thorogood & Davies, 2012). Hosts react more aggressively towards the locally-common morph (Honza et al., 2006; Trnka & Grim, 2013), but can be induced to switch their behaviour if the alternate form becomes less rare (Thorogood & Davies, 2012). Similar plumage polymorphisms are common across Cuculidae but are more likely to occur in species that are brood parasites, suggesting that this is an adaptation to facilitate successful parasitism (Thorogood & Davies, 2013).
The factors underlying the evolution of egg polymorphism are also well understood (Honza & Cherry, 2017). A great number of host species across different avian clades recognise and reject (either by ejection or nest abandonment) the parasite egg (Medina & Langmore, 2015; Yang et al., 2020), which leads to egg mimicry by the parasite (Brooke & Davies, 1988; Stoddard & Stevens, 2010). Host egg polymorphism can thus evolve in response to parasite mimicry because it facilitates recognition of the parasite egg (Medina & Langmore, 2016; Takasu, 2003; Yang et al., 2020). Host egg polymorphism can occur as variation between clutches (Moskát et al., 2008; Øien et al., 1995; Stokke et al., 2002), discrete polymorphism (two or more discrete egg morphs) within a host population (Liang et al., 2017; Yang et al., 2016) or as a variation in multiple egg features (eggshell colour, luminance and spot patterns) acting as individual signatures (Caves et al., 2015; Spottiswoode & Stevens, 2011). Egg polymorphism in brood parasites can occur when generalist brood parasites exploiting multiple host species specialise into host-specific lineages whose egg phenotype is a close match of the specific host they parasitise (Abernathy & Liang, 2020; Brooke & Davies, 1988; Gibbs et al., 2000; Spottiswoode et al., 2011; Starling et al., 2006; Stoddard & Stevens, 2010).
Despite decades of research, we still do not fully understand the coevolution between host and parasite nestlings. One reason for this is that much of the empirical and theoretical research has historically focused on few well-known avian parasite–host systems such as the common cuckoo from Europe and the cowbirds, genus Molothrus, from the Americas and their respective hosts (Kennerley et al., 2022). Many hosts in these systems have well-developed defences at the frontline (before the cuckoo egg is laid, for example mobbing the adult parasite) or egg (after the cuckoo egg is laid, for example rejection of the cuckoo egg or nest desertion) stages, but always accept the parasite nestlings (but see De Mársico et al., 2012 and Grim et al., 2003 for rejection of parasite fledglings). Therefore, nestling discrimination was assumed to be a costly defence due to the high risk of misrecognising own chicks as a parasite (Lotem, 1993). Additionally, hosts encounter parasite nestlings at lower frequencies than parasite eggs, which might relax selection for the development of host defences at the nestling stage (Britton et al., 2007; Grim, 2006). As a consequence, the nestling stage of the arms race has been less studied than the frontline and egg stages (Grim, 2011, 2017) and the selective factors and evolutionary outcomes underlying phenotypic diversification at the nestling stage of the arms race are less understood.
An example of nestling mimicry and diversification comes from the parasitic Vidua finches and their hosts. Vidua nestlings mimic gape colour and patterns, begging calls and begging postures of their estrildid hosts (Jamie et al., 2020; Payne, 2005a). However, they do not evict host nestlings or affect their survival rates (Schuetz, 2005a) and the hosts do not discriminate Vidua nestlings (Schuetz, 2005a, 2005b). Therefore, in these systems nestling mimicry is not the result of coevolutionary interactions between host and parasite but rather appears to be due to nestling competition to stimulate parental provisioning (Hauber & Kilner, 2007; Kilner et al., 1999). The Australasian bronze-cuckoos (genus Chalcites) and their hosts, on the other hand, show evidence of a coevolutionary arms race at the nestling stage. Some hosts of bronze-cuckoos recognise and reject (Langmore et al., 2003) or eject the parasite nestling from the nest (Sato et al., 2015; Sato, Tokue, et al., 2010; Tokue & Ueda, 2010). Correspondingly, bronze-cuckoo nestlings mimic host nestlings in traits such as colour and luminance of skin, flanges and gape and presence and distribution of down feathers (Attisano et al., 2018; Langmore et al., 2011; Noh et al., 2018). Hosts discriminate the parasite nestlings via cues such as natal down, hatching order or begging calls (Attisano, Sato, et al., 2021; Langmore et al., 2003; Noh et al., 2018, 2021). Host breeding experience and risk of parasitism can also influence rejection of the parasite nestling (Langmore, Cockburn, et al., 2009). Remarkably, some of these systems also present polymorphism in both the host and parasite nestlings (Attisano, Sato, et al., 2021; Langmore et al., 2011; Sato et al., 2015).
We present a synthesis based on 40 years of studies on the biology of the shining bronze-cuckoo (Chalcites lucidus) and its hosts conducted in Australia, New Caledonia and New Zealand. We first describe the biology of the shining bronze-cuckoo and its hosts, highlighting commonalities and differences across the range by reviewing key aspects of these parasite–host relationships. We continue by describing host nestling mimicry by the shining bronze-cuckoo, nestling polymorphism and host defences at the nestling stage. Finally, we discuss the role of brood parasitism in promoting nestling phenotypic diversification and the evolutionary implications of nestling polymorphism in the arms race between the shining bronze-cuckoo and its hosts. The field work in New Caledonia was conducted with permits from the Province Sud Nouvelle Calédonie (3045–2011, 2437–2012, 2532–2013, 2801–2014, 2476–2015, 2372–2017, 3469–2018 and 2720–2019) and did not require ethical approval. Field work in New Zealand was conducted with permits from The Wildlife Service, Marlborough Catchment Board, The University of Canterbury Animal Ethics Committee (AEC 2010/24R) and the New Zealand Department of Conservation, Te Papa Atawhai (AK/15301/RES). Field work in Australia was conducted under the approval of the Australian National University Animal Experimentation Ethics Committee (Protocol Numbers F.BTZ.99.99, A2012/47 and F.BTZ.61.03, A2021/37), and Environment ACT (Licence numbers LT1999021, LT200340, LT2004112, LT2005164, LT2006229, LT2007266, LT2011509, LT2012589, LT2013667, LT2014776, LT2015842 and LT2016895).
2 MODEL SPECIES
2.1 The shining bronze-cuckoo
Four subspecies of the shining bronze-cuckoo are currently recognised (Payne, 2005b). The migratory C. l. plagosus breeds in Tasmania, south-east and south-west Australia and overwinters in the Lesser Sunda Islands and New Guinea (Friedman, 1968; Gill, 1983a). The migratory subspecies C. l. lucidus breeds in New Zealand and the Chatham Islands and overwinters in the Solomon Islands and Bismarck Archipelago (Fell, 1947; Friedman, 1968). These two subspecies might come into contact during migration in Queensland outside the breeding season (Gill, 1983a; Noske, 2019), but genetic admixture is unlikely. The subspecies C. l. layardi forms resident populations in New Caledonia and Vanuatu. The fourth subspecies C. l. harterti is resident on Rennell and Bellona Islands, however little is known about its behavioural ecology, and we exclude it from our synthesis. There is no evidence of contact between migratory and resident subspecies (Friedman, 1968).
The shining bronze-cuckoo is a specialist parasite of a small group of closely related hosts in the family Acanthizidae, all of which build suspended, dome-shaped nests. The subspecies on the New Zealand mainland exclusively parasitises the grey warbler, Gerygone igata (Gill, 1983b, 1998), the subspecies from New Caledonia exclusively parasitises the fan-tailed gerygone, G. flavolateralis (Sato et al., 2015) and the subspecies from Australia is a specialist parasite of thornbills (Acanthiza species), particularly the yellow-rumped thornbill A. chrysorroa (Brooker & Brooker, 1989a). Although historical nest records in Australia reported also superb fairy-wrens Malurus cyaneus and splendid fairy-wrens M. splendens as hosts of C. l. plagosus (Brooker & Brooker, 1989b), multiple field studies found no parasitism of fairy-wrens (Brooker & Brooker, 1989a, 1992; Langmore & Kilner, 2007). Field experiments additionally showed that all shining bronze-cuckoo chicks transferred from thornbill nests to superb fairy-wren nests resulted in rejection of the cuckoo chicks by the fairy-wren hosts (Langmore et al., 2003). Historical records of shining bronze-cuckoo nestlings being reared by superb fairy-wrens were probably misidentifications of Horsfield's bronze-cuckoos C. basalis, the primary brood parasite of fairy-wrens (Payne, 2005b).
Female shining bronze-cuckoos produce dark unmarked eggs with colours ranging from olive-green to dark brown (Figure 1), which have been assumed to be cryptic in the poorly illuminated hosts' domed nests (Langmore et al., 2005; Langmore, Stevens, et al., 2009). The contrast with the host eggs is striking. All exploited hosts have eggs that are either immaculate white or whitish with brown speckles (Figure 1), but hosts usually accept foreign eggs irrespective of their coloration (Attisano, Sato, et al., 2021; Thorogood et al., 2017). Female bronze-cuckoos always remove one egg from the nest before laying their own and sometimes they can remove an egg laid by a previous cuckoo female (Gloag et al., 2014). Therefore, the dark coloration might decrease the chances of detection by other competing female cuckoos, rather than by the host (Brooker et al., 1990; Gloag et al., 2014; Thorogood et al., 2017).
The duration of the cuckoo laying season is 18–19 weeks (August – December) in Australia (Brooker & Brooker, 1989b), 13–15 weeks (October – January) in New Caledonia (Attisano et al., 2019) and 10–12 weeks (October – December) in New Zealand (Gill, 1983b). Parasitism rates vary across the three regions (Table 1) and increase over the breeding season in Australia and in New Zealand's South Island (Gill, 1983b, 1998; Medina & Langmore, 2016) are low throughout the season in New Zealand's North Island (Anderson et al., 2013) and relatively constant throughout the breeding season in New Caledonia (Table 1). Multiple parasitism (i.e. when multiple cuckoo females lay single eggs in the same host nest) is common in some bronze-cuckoos (e.g. C. minutillus, Gloag et al., 2014). However, it is rare in the shining bronze-cuckoo: instances of two cuckoo eggs being laid in the same gerygone nest have been reported only once in New Caledonia (Attisano et al., 2019) and once in New Zealand (Briskie, 2007). The average incubation period of the cuckoo egg is shorter than that of the hosts' eggs (Attisano et al., 2019; Brooker & Brooker, 1989a; Gill, 1983b; Table 1); therefore, cuckoo nestlings usually hatch earlier than host nestlings. However, mis-timed laying can result in the cuckoo chicks hatching after the host chicks, a relatively frequent occurrence in New Zealand (Table 1). The shining bronze-cuckoo chick usually evicts all host eggs or nestlings within 1–2 days after hatching (Brooker & Brooker, 1989a; Sato et al., 2015), but eviction can occur as late as 7 days after hatching (Gill, 1983b; Table 1).
| Australia | New Caledonia | New Zealand | |
|---|---|---|---|
| Shining bronze-cuckoo subspecies | C. l. plagosus | C. l. layardi | C. l. lucidus |
| Secondary hosts | A. reguloides | No | No |
| Migratory | Yes | No | Yes |
| Host nestling mimicry | Yes | Yes | Yes |
| Nestling polymorphism (bright: dark) | 91:9 (n = 22) | 94:6 (n = 26) | Only dark (n = 20) |
| Average incubation (days) | 13 ± 1 | 16 ± 1 | 16 ± 5 |
| Average nestling period (days) | 20 ± 1 | 19 ± 1 | 20 ± 1 |
| Cuckoo laying relative to host laying (nests) | |||
| Before | 40% | 0 | 0 |
| During | 53% (n = 15) | 100% (n = 26) | 37% (n = 8) |
| After | 7% | 0 | 63% |
| Cuckoo hatching relative to first host hatching (nests) | |||
| Before | 83% | 80% | 58% |
| Same day | 0 (n = 6) | 10% (n = 19) | 0 (n = 12) |
| After | 17% | 10% | 42% |
| Eviction of host eggs/nestlings relative to cuckoo hatching (nests) | |||
| Day 1 from hatching | 0 | 25% | 0 |
| ≥ Day 2 from hatching | 100% (n = 7) | 75% (n = 4) | 100% (n = 12) |
| Cuckoo nestlings rejected by hosts | 0 | 86% (n = 28) | 0 |
| Cuckoo fledglings | 92% (n = 24) | 7% (n = 28) | 65% (n = 20) |
| Main host | A. chrysorrhoa | G. flavolateralis | G. igata |
|---|---|---|---|
| Breeding system | Facultatively cooperative | Socially monogamous | Socially monogamous |
| Territorial | No | Yes | Yes |
| Rejects cuckoo eggs (ejection or nest abandonment) | Only if cuckoo laying mis-timed | No | No |
| Ejects cuckoo chicks | No | Yes | No |
| Nestling polymorphism (bright: dark) | No | 75:25 (n = 222) | 66:34 (n = 192) |
| Brood ratio (bright:dark:mixed) | All bright (n = 40) | 69:23:8 (n = 132) | 43:15:42 (n = 65) |
| Mean clutch size | 3.2 ± 0.1 (n = 51) | 2.0 ± 0.1 (n = 223) | 3.5 ± 0.1 (n = 117) |
| Average incubation length (days) | 17 ± 1 | 17 ± 1 | 20 ± 1 |
| Average nestling period (days) | 19 ± 1 | 15 ± 1 | 18 ± 1 |
| Parasitism rate | 30% (n = 106) | 20% (n = 344) | 26% (n = 138) |
| Nests with multiple parasitism | 0 | 1 | 0a |
- a Briskie (2007) reported a case of multiple parasitism in a different field study.
2.2 The hosts
The nests of the Acanthizidae hosts of the shining bronze-cuckoo have a distinctive dome-shaped structure and are suspended in various trees and shrubs at heights ranging from ground level to 20 m in the canopy (Attisano et al., 2019; Ebert, 2004; Gill, 1982). G. igata and G. flavolateralis are the only species building such nests in New Zealand (mainland) and New Caledonia, respectively, and no evidence of parasitism by shining bronze-cuckoos has been found in other common sympatric open-cup nesters that are potentially suitable hosts in these regions (Briskie, 2003; Gill, 1983b). In Australia, the number of potential hosts with open and domed nests is higher, however the shining bronze-cuckoo targets thornbill hosts, in particular the yellow-rumped thornbill A. chrysorrhoa, the nests of which closely resemble gerygone nests (Brooker & Brooker, 1989a; Medina & Langmore, 2016). Therefore, host nest location and structure might have more influence than host density or diet in determining host choice by the shining bronze-cuckoo.
The yellow-rumped thornbill from Australia is a facultative cooperative breeder and does not defend exclusive breeding territories (Ebert, 2004). On the other hand, the gerygones from New Zealand and New Caledonia form monogamous pairs and defend their breeding territory against conspecific trespassers (Attisano et al., 2019; Gill, 1982). Nest predation is the major cause of nest loss in all the hosts, with predation rates ranging from about 30% in New Zealand (Gill, 1982) to over 80% in New Caledonia and Australia (Attisano et al., 2020; Ebert, 2004). Besides nest losses due to host egg and chick eviction by the cuckoo nestling, adult female shining bronze-cuckoos also destroy 13%–16% of the active non-parasitised nests by removing host eggs or killing host nestlings (Attisano et al., 2020; Briskie, 2007). This behaviour targets nests that are too advanced in the breeding cycle and stimulates re-nesting in the hosts, thereby increasing the availability of new nests for parasitism (‘farming strategy’; Soler, 2017).
The clutch size is significantly smaller in the New Caledonian host than in the hosts from New Zealand and Australia (Table 1). All the hosts lay eggs at 2-day intervals; thus, the host laying period is between 4 and 6 days in New Caledonia and 6–8 days in New Zealand and Australia. Female cuckoos increase the survival chances of their nestling by laying their egg during the host incubation period (Davies, 2015). The laying window is thus significantly shorter for the New Caledonian cuckoo, which might help explain the strong synchronicity with the host laying period (Table 1). Re-nesting after nest failure occurs in all hosts, however consecutive successful breeding attempts within the same season have been observed only in Australia and in New Zealand's South Island, whereas in the North Island and in New Caledonia host pairs have typically only one successful breeding attempt per season (Anderson et al., 2013; Attisano et al., 2019; Ebert, 2004; Gill, 1982).
All the hosts show some defences at the frontline stage. They react aggressively to adult shining bronze-cuckoos approaching the nest (Briskie, 2007) as well as to stuffed specimens of shining bronze-cuckoos presented at the nest (Attisano, Hlebowicz, et al., 2021; Medina & Langmore, 2016). The yellow-rumped thornbill begins nesting in mid-winter, substantially earlier than congeneric and sympatric insectivorous species and before migratory cuckoos arrive in their breeding ground, which is consistent with selection for earlier breeding to escape parasitism (Medina & Langmore, 2016). A similar process might occur in the grey warbler as in the South Island breeding begins before arrival of migratory cuckoos (although not earlier than sympatric insectivores that are not parasitised); thus, earlier clutches suffer lower parasitism rates than later ones (Anderson et al., 2013).
Yellow-rumped thornbills sometimes reject cuckoo eggs laid before the host starts laying by burying these in the nest lining or they might abandon the nest if the cuckoo lays after the host has started incubation (N. Langmore, pers. obs.). They also show low rejection rates of blue artificial eggs introduced in the nest (Medina & Langmore, 2019). By contrast, gerygone hosts always accept cuckoo and artificial eggs mimicking size and colour of real host and cuckoo eggs (Attisano, Sato, et al., 2021; Gloag et al., 2014; Thorogood et al., 2017). A possible explanation for gerygone hosts being egg acceptors is that they might be unable to grasp and eject the large cuckoo egg (Gill, 1998; Guigueno et al., 2014; Moksnes et al., 1991). In New Caledonia, the cuckoo egg is about 50% larger in volume than the host egg (Figure 1) and the fan-tailed gerygone host has a bill size comparable with small acceptor hosts of the common cuckoo in Europe and cowbirds in North America (Moksnes et al., 1991; Rasmussen et al., 2010). Indeed, fan-tailed gerygones always accept artificial eggs matching the size of real eggs but frequently eject artificial eggs smaller than real eggs even if these mimic the colour of host or parasite eggs, thus suggesting that the small bill size might prevent ejection of the large cuckoo eggs (Attisano, Sato, et al., 2021). Another explanation is that by accepting the cuckoo egg the host reduces the chances of losing an additional host egg during multiple parasitism attempts by different cuckoo females (Sato, Mikami, & Ueda, 2010).
Frontline and egg defences in these hosts are thus only partly or not at all effective in preventing parasitism. This might have been instrumental in promoting an escalation to the nestling stage as the next step in these arms races (Grim, 2017; Langmore et al., 2003).
3 NESTLING MIMICRY, POLYMORPHISM AND DISCRIMINATION
3.1 Nestling mimicry and polymorphism in parasite nestlings
Discrimination of the parasite egg by the hosts favours selection of host egg mimicry in brood parasites (Honza & Cherry, 2017). Similarly, discrimination of the parasite nestling by hosts of bronze-cuckoos (Langmore et al., 2003; Sato, Tokue, et al., 2010; Tokue & Ueda, 2010) has favoured selection of host nestling mimicry in several bronze-cuckoo species (Langmore et al., 2011). Some hosts of bronze-cuckoos also discriminate the begging call of the parasite nestling (Colombelli-Négrel et al., 2012; Langmore et al., 2003; McLean & Waas, 1987), which in turn has favoured selection of begging call mimicry by the parasite (Anderson et al., 2009; Langmore et al., 2008; Noh et al., 2021; Ranjard et al., 2010).
The shining bronze-cuckoo also has geographic variation in visual mimicry with nestlings of each subspecies closely resembling the respective local host in traits such as colour and luminance of the skin and mouth flanges (Figure 2 and Supplementary Material). McGill and Goddard (1979) found a similar difference in cuckoo nestling coloration among subspecies of the little bronze-cuckoo C. minutillus. Besides coloration, the shining bronze-cuckoo nestlings of each subspecies also usually mimic their respective hosts in regard to the presence of natal down; cuckoo nestlings from New Caledonia and New Zealand have sparse down feathers which mimic the natal down of the host nestlings (Figure 3). However, cuckoo nestlings in Australia are usually naked or have a small number of short, fine filaments on the head, whereas host nestlings have a longer, fine down on the head and back (Langmore et al., 2011). Usually, Cuculinae nestlings are naked at hatching (Payne, 2005b), however natal down is a common trait of bronze-cuckoos parasitising gerygone hosts (Langmore et al., 2011; Tokue & Ueda, 2010). Remarkably, shining bronze-cuckoo nestlings can also occur in two distinct skin colour morphs within a subspecies, typically a common ‘bright’ and a rare ‘dark’ (melanic) morph (Figure 3). The occurrence of polymorphism in shining bronze-cuckoos varies across its range, as it has been observed only in Australia and New Caledonia (Langmore et al., 2011; Sato et al., 2015), but not in New Zealand where Gill (1983b) only observed a dark morph (Table 1). There is therefore evidence of the formation of host-specific lineages in the shining bronze-cuckoo, as the nestlings of each cuckoo subspecies appear more similar to their main hosts than they are to each other (Figures 2 and 3 and Supplementary Material).
3.2 Polymorphism in host nestlings
One of the most intriguing aspects of the shining bronze-cuckoo – host systems is that hosts also have polymorphic nestlings with a common ‘bright’ and a rare ‘dark’ (melanic) morph (Figure 3). As with the cuckoo, polymorphism varies regionally: gerygone hosts from New Caledonia and New Zealand have polymorphic nestlings, whereas yellow-rumped thornbill nestlings from Australia are monomorphic (Figure 3). In addition, while the New Caledonian host is distinctly dimorphic, colour variation in the New Zealand host is less evident yet clearly noticeable to a human observer. In New Zealand, the two host nestling morphs progressively darken with age (Gill, 1983c), whereas in New Caledonia only the bright morph becomes progressively darker until the two morphs are indistinguishable from each other around the age of 9–10 days (Attisano et al., 2018). The two morphs of the New Caledonian host also have similar growth and survival rates, with no evident effect of skin coloration on nestling condition (Attisano et al., 2019).
The two nestling morphs can co-exist in the same brood (Gill, 1983c; Sato et al., 2015). The overall ratio of bright to dark host nestlings is similar in New Caledonia and New Zealand, but mixed broods are relatively more common in the New Zealand host and the brood coloration varies across sites within both regions (Table 1 and Supplementary material). At least for the New Caledonian host, extra-pair copulation does not influence the proportion of mixed broods (Bojarska et al., 2018) and population genetic analyses confirm that the observed host chick polymorphism does not diverge from a Hardy–Weinberg equilibrium, particularly when assuming a dominant dark-morph allele (Sato et al., 2015).
3.3 Host defences at the nestling stage
In Australia, all cuckoo nestlings that hatched in nests of yellow-rumped thornbills evicted host eggs or nestlings and were reared by the hosts. There is thus no indication of discrimination of the parasite nestling (Table 1).
In New Zealand, a total of 20 cuckoo nestlings hatched in nests of grey warblers and evicted host eggs or nestlings, and 13 cuckoos were reared by the host (Table 1). Of the seven cuckoo nestlings that failed to fledge, four died from nest depredation but the cause of death of the other three cuckoo nestlings was uncertain. In one case, a cuckoo nestling was dead in the nest after a sudden change to cold and wet weather and with no signs of injury, however we cannot rule out that the death resulted from nest abandonment by host parents. Finally, two other nests, each previously containing a cuckoo nestling (3–7 days old), were found intact but empty at last inspection. It is possible that these nestlings were depredated by adult shining bronze-cuckoos (Briskie, 2007), but we also cannot exclude the possibility that host parents removed the parasite from the nest. In summary, field data show that the grey warbler host does not usually discriminate and eject the cuckoo nestling, but any rejection of the cuckoo nestling cannot yet be ruled out without an experiment manipulating nestling mimicry.
By contrast, in New Caledonia most cuckoo nestlings that hatched in fan-tailed gerygone nests were ejected by the host within 24–48 h from hatching before they could evict any host eggs or nestlings (Table 1). The discrimination ability of the host is very accurate: they ejected the parasite even when it was the only chick in the nest and never ejected host chicks in non-parasitised or parasitised nests (Attisano et al., 2018; Sato et al., 2015). Both parents can eject the parasite nestling: in three cases, this was the host mother, in another three cases the host father and in one case both parents seemingly collaborated in ejecting the parasite nestling. The parasite nestling was ejected whether or not it matched the host brood colour (Attisano et al., 2018; Sato et al., 2015). Field experiments show that nestling down, hatching order and begging calls might be more important cues for the recognition of the parasite than chick colour (Attisano, Sato, et al., 2021). Host acceptance of the shining bronze-cuckoo nestlings in New Caledonia is relatively low, as only 14% of the cuckoo hatchlings (n = 28) were accepted by the host parents, of which only two fledged (Table 1).
4 DISCUSSION
4.1 Host-specific lineages in the shining bronze-cuckoo
Chalcites bronze-cuckoos are already known for the exquisite mimicry exhibited by their nestlings (Langmore et al., 2011), but our synthesis adds a further level of complexity in these intriguing arms races by highlighting the remarkable phenotypic diversification of the shining bronze-cuckoo's nestlings across its range. The shining bronze-cuckoo is a specialist brood parasite of a small number of passerines, some of which discriminate and reject the parasite nestling. The choice of hosts might be constrained by their availability, for example only one host exists in New Caledonia and mainland New Zealand, by competition for hosts with other bronze-cuckoo species, for example between C. minutillus and C. basalis in Australia, or by specific habitat or diet requirements (Payne, 2005b). It is well known that discrimination of the parasite egg by the host can lead to parasite egg mimicry and diversification into host-egg-specific lineages (Abernathy & Liang, 2020; Gibbs et al., 2000; Spottiswoode et al., 2011; Starling et al., 2006). Therefore, the shining bronze-cuckoo – host systems fulfil the expectations of selection for the evolution of specialised host-specific lineages at the nestling stage of the arms race much in the same way that host discrimination drives parasite egg mimicry and diversification.
No study yet has looked at the genetic relationships among the shining bronze-cuckoo subspecies, whereas the population genetic structure and host-specific lineages of two closely related species, the Horsfield's bronze-cuckoo (C. basalis) and the little-bronze-cuckoo (C. minutillus), have received more attention. The Horsfield's bronze-cuckoo shows no evidence of host-specific diversification (Joseph et al., 2002). It has instead evolved a ‘Jack-of-all-trades’ strategy that appears to have relaxed selection for host-specific lineages: it has evolved an egg that is somewhat similar to the eggs of several hosts (Feeney et al., 2014), a two-tone nestling colour that is somewhat similar to nestlings of various hosts (Langmore et al., 2011) and a begging call that can be modified to match that of the relevant host (Langmore et al., 2008). Australian populations of the little bronze-cuckoo are not completely genetically separated (Joseph et al., 2011), but present nestling diversification that suggests the presence of host-specific lineages (McGill & Goddard, 1979). The shining bronze-cuckoo has a wider geographic range and subspecies that are spatially separated in distinct geographic regions, which could result in genetic isolation and a more marked differentiation than in other closely related species.
4.2 Coevolutionary arms race and nestling polymorphism
Mimicry and host-specific lineages at the egg and nestling stages might be shaped by similar selective factors and have similar roles in the coevolutionary interactions between brood parasites and their hosts (Honza & Cherry, 2017; Tanaka, 2016). However, the role of nestling polymorphism in the arms race is not as clear as that of egg polymorphism. Nestling polymorphism is rare in birds (Kilner, 2006), whereas egg polymorphism is relatively common (Yang et al., 2020). A few studies have reported discrete polymorphism in melanin or carotenoid-based plumage coloration in older nestlings or juveniles (Galván et al., 2010; Kapun et al., 2011; Rohwer et al., 2012; Roulin et al., 2016), however, the case of shining bronze-cuckoos and their gerygone hosts is, to our knowledge, the only known example of discrete skin, but not plumage, polymorphism in hatchlings, and of nestling polymorphism occurring in both the brood parasite and the host.
Phenotypic diversification and polymorphism are known in brood parasites as a consequence of their parasitic habits (Medina & Langmore, 2015). The rare dark morph of the Australian cuckoo subspecies might indicate a specialisation to a host with unknown nestling colour (e. g. Tasmanian thornbill, Acanthiza ewingii). The New Caledonian cuckoo subspecies started parasitising its single host within the last 1.7 my (Nyári & Joseph, 2012) and the nestling polymorphism might have originated before the separation from the Australian subspecies or have parallelly been developed in more recent times. The absence of polymorphism in the New Zealand cuckoo might be due to the loss of a second morph or because this was not observed during field studies due to small samples and its rarity. However, in all regions we found that nestlings of each cuckoo subspecies have a remarkable resemblance to the main host (Figures 2 and 3 and Supplementary Material). Furthermore, the common bright cuckoo morph in New Caledonia mimics several visual features of both host morphs (Attisano et al., 2018), which might indicate selection for an intermediate parasite morph similar to the intermediate egg and nestling phenotypes of the generalist common cuckoo (Stoddard & Stevens, 2010) and Horsfield's bronze-cuckoo (Feeney et al., 2014; Langmore et al., 2011). Such a strategy should return higher pay-offs due to the overall high frequency of bright host nestlings, of which the bright cuckoo is a closer match, and still offer some competitive advantage when the parasite does not match the host brood colour. Therefore, the rarity of the dark morphs in Australia and New Caledonia and the monomorphism in New Zealand might indicate selection towards improved mimicry of the host nestlings rather than polymorphism.
Host nestling polymorphism fits with theoretical predictions that offspring phenotypes dissimilar from the mimetic parasite should improve the capability for discrimination by the host (Takasu, 2003; Tanaka, 2016); however, the nestling phenotype might be a less reliable discrimination cue than the egg phenotype. First, the egg phenotype is maternally inherited (Gosler et al., 2000), whereas the nestling phenotype is a result of the genetic contribution of both parents (Roulin & Dijkstra, 2003). For example, in New Caledonia each adult host can experience a brood colour change across seasons due to partner replacement or extra-pair copulations (Bojarska et al., 2018), which also means that the frequency of mixed broods can be low but cannot reach zero in this host. Second, egg discrimination in rejector hosts is improved by increased within-nest uniformity and between-nest variability of egg phenotypes (Moskát et al., 2008; Stokke et al., 2002), however mixed broods of the two gerygone hosts challenge this assumption and should in theory lead to more recognition errors and rejection of own offspring by mistake. The frequency of mixed broods varies among and within regions which might suggest a disadvantage of mixed broods in areas with high parasitism pressure and thus selection for monomorphic broods of either morph. Third, the female is usually the sex responsible for the discrimination and rejection of the parasite egg as she can imprint on her own egg phenotype or compare the foreign egg with her own (Davies, 2015; Grim, 2017). However, in the fan-tailed gerygone both sexes can discriminate and eject the parasite nestling. Thus, imprinting on their own nestling phenotype is advantageous only if the pair produces the same nestling phenotype, which is not always the case in this host. Therefore, nestling coloration should not be the main cue for the discrimination of the parasite. Experimental evidence indeed shows that fan-tailed gerygones discriminate foreign nestlings mainly based on cues such as down feathers, hatching order and begging calls (Attisano, Sato, et al., 2021).
Nestling polymorphism might also be related to environmental factors, for example skin coloration might be influenced by UV radiation levels that are potentially harmful for the nestlings. Testing this hypothesis will require more work as little is known about the possible protective role of skin melanisation against UV radiation in nestlings (Galván & Solano, 2016). Melanisation is also linked to immune response and fitness in birds (Chakarov et al., 2008; Gangoso et al., 2015), however the condition of nestlings does seem to be unrelated to skin melanisation in the fan-tailed gerygone (Attisano et al., 2019). Further work is thus needed to understand if and how environmental factors and physiology contribute to nestling polymorphism in these host species.
In New Caledonia, nestling polymorphism might have decreased the rate of recognition errors of the parasite during early stages of the arms race. However, it gradually lost its importance as the main discrimination cue due to the cuckoo achieving improved mimicry and polymorphism, leading to a finer discrimination based on a combination of multiple cues (Attisano et al., 2018; Attisano, Sato, et al., 2021). Polymorphism can then persist in the host population because it bears no additional costs to nestling growth and survival. In New Zealand, the arms race might be at an early stage as the host morphs are not yet sufficiently dissimilar from the parasite nestling to allow host parents to reliably discriminate the parasite (Figure 3 and Supplementary Material). Further support for different evolutionary stages of the arms race also come from the fact that the cuckoo accurately mimics the begging calls of the grey warbler in New Zealand (Anderson et al., 2009; McLean & Waas, 1987; Ranjard et al., 2010), whereas the mimicry is imperfect in New Caledonia (Attisano, Sato, et al., 2021).
4.3 Host defences at the nestling stage
The nestling mimicry by the Australian cuckoo subspecies suggests the possibility that a thornbill host was able to discriminate the parasite nestling at some point in the past but nowadays it can no longer accurately discriminate the parasite from its own chicks. Instead, it developed moderately effective anti-parasitism defences at the nest-building and egg stages (Medina & Langmore, 2016), which might have further relaxed selection for improved nestling recognition (Britton et al., 2007; Grim, 2006).
Besides the fan-tailed gerygone from New Caledonia, two other gerygone species from Australia, the large-billed G. magnirostris and the mangrove gerygone, G. laevigaster, which are parasitised by the little bronze-cuckoo, recognise and eject parasite nestlings (Sato, Tokue, et al., 2010; Tokue & Ueda, 2010). These three gerygone nestling ejector species belong to the same evolutionary branch (Nyári & Joseph, 2012), thus suggesting a common origin for this host defence behaviour. True recognition of the parasite occurs in the large-billed gerygone and fan-tailed gerygone hosts (Attisano, Sato, et al., 2021; Noh et al., 2018). However large-billed gerygones sometimes eject their own nestlings by mistake (Noh et al., 2018; Sato, Tokue, et al., 2010; Tokue & Ueda, 2010), whereas fan-tailed gerygones do not (Attisano et al., 2018; Sato et al., 2015). The misrecognition errors of the large-billed gerygone host might depend on the fact that the little bronze-cuckoo is the gerygone specialist in Australia and nestling mimicry can be so accurate as to sometimes fool the host. On the other hand, the shining bronze-cuckoo is not a gerygone specialist in its Australian range meaning the fan-tailed gerygone might only be a suboptimal host, yet the only one building a dome-shaped nest in New Caledonia, and the less accurate mimicry allows the host to always recognise its own nestlings. The New Caledonian cuckoo subspecies is also a year-round resident; thus the parasitism pressure is evenly spread across the entire breeding season and the host cannot escape parasitism via changes in the breeding phenology. Coupled with the lack of effective defences at the egg laying stage, these factors might have strengthened selection for accurate nestling discrimination in the fan-tailed gerygone host.
The grey warbler from New Zealand is closely related to the three gerygone ejector species and separated from them about 100,000 years ago (Nyári & Joseph, 2012). Therefore, the arms race in New Zealand might be at a relatively early stage in which enough selective pressure promoted the parasite's visual and auditory mimicry but not nestling dimorphism and recognition of the parasite nestling by the host. It is also possible that grey warblers have some defences at the nestling stage that do not involve ejection of the parasite, for example they might abandon nests containing a single parasite nestling as other bronze-cuckoo hosts do (Langmore, Cockburn, et al., 2009), but our field evidence does not conclusively support this. Grey warblers also escape parasitism thanks to their breeding phenology as populations on the North Island and earlier breeders in the South Island of New Zealand suffer low parasitism rates (Anderson et al., 2013; Gill, 1983b). This might suggest relaxed selection for nestling discrimination in this host.
5 CONCLUSIONS
The shining bronze-cuckoo – host systems are promising models for coevolutionary research as they can contribute to significant advancements of current theories of the nestling stage of the arms race. We suggest some possible research avenues that could contribute greatly to our understanding of these systems. First, there is the need to clarify the phylogenetic relationships within the shining bronze-cuckoo clade and understand if the variation in nestling appearance is due to the presence of host-specific lineages or to genetically distinct cuckoo species. Second, detailed field studies of other members of the Gerygone clade are needed to assess the variation in host defences and the occurrence of nestling polymorphism across members of the group. The ancestral nature of the parasite ejection behaviour can only be supported by confirming its occurrence in a larger number of Gerygone species. Additionally, gerygones show specialisation to different types of habitats over their wide geographic range (Keast & Recher, 1997), thus factors dependent on local ecological conditions (for example temperature, solar radiation or nest microclimate environment) could potentially influence the nestling phenotype. Therefore, gerygones are excellent models for comparative analyses of the wider ecological factors shaping the evolution of nestling polymorphism and defences at the nestling stage of the arms race. Third, the study of the genetic and physiological mechanisms underlying nestling skin polymorphism could help in understanding the selective factors contributing to the expression of the trait. Molecular (genomics, proteomics) and physiological (stress response assays) techniques have so far been employed in the study of nestling plumage polymorphism (Gangoso et al., 2015; Roulin & Dijkstra, 2003), yet no study has been published on species with skin polymorphic nestlings. Finally, a more careful investigation on the presence of nestling polymorphism in hosts, including those not belonging to the genus Gerygone, and non-hosts is also needed. Nestling polymorphism might have been overlooked due to an historical imbalance of studies focusing on the egg stage rather than the nestling stage of the arms race. Examples of host defences at the nestling stage have been found in multiple hosts (reviewed in Grim, 2017), thus researchers should be aware that nestling polymorphism could be a possible outcome of the arms race. At the same time, evidence of nestling polymorphism in non-hosts would allow to confirm or deny its role within an arms race and better understand its evolutionary causes.
AUTHOR CONTRIBUTIONS
Alfredo Attisano, Brian J. Gill, Michael G. Anderson, Roman Gula, Naomi E. Langmore, Nozomu J. Sato, Keita D. Tanaka, Rose Thorogood, Keisuke Ueda and Jörn Theuerkauf designed the studies; Alfredo Attisano, Brian J. Gill, Michael G. Anderson, Naomi E. Langmore, Yuji Okahisa, Nozomu J. Sato, Rose Thorogood collected the data; Alfredo Attisano led the writing of the manuscript; Alfredo Attisano, Brian J. Gill, Michael G. Anderson, Roman Gula, Naomi E. Langmore, Keita D. Tanaka, Rose Thorogood and Jörn Theuerkauf contributed to the writing; all authors gave final approval for publication.
ACKNOWLEDGEMENTS
The research was funded by the National Science Centre, Poland: NCN 2012/05/E/NZ8/02694 and NCN 2016/23/B/NZ8/03082; by the Japan Society for Promotion of Science (JSPS): grant no. 24-4578 (to NJS), 24770028 (to K.D.T.), 23255004 (to K.U.); by Rikkyo University: SFR 11-54 (to N.J.S.); by the Australian Research Council: grant DP180100021 (to N.E.L.); by a University Grants Committee Postgraduate Scholarship (to B.J.G.) and institutional support from the University of Canterbury; and by a Phyllis and Eileen Gibbs Travelling Fellowship from Newnham College, Cambridge, UK (to R.T.). For data collected from Kaikoura in 2010, we thank Jim Briskie for facilitating fieldwork and helping with permit applications, Jack van Berkel for providing facilities at the Edward Percival Field Station, and Justin Rasmussen and Tom Walker for help with nest searching and monitoring. We thank two anonymous reviewers for useful comments.
CONFLICT OF INTEREST
The authors declare that they have no competing interests.
Open Research
DATA AVAILABILITY STATEMENT
Data used to generate Figures 2 and 3 and additional figures and tables in the Supporting Information is available from Dryad Digital Repository https://doi.org/10.5061/dryad.95x69p8pf (Attisano et al., 2022).
