Breaking the rules: sex roles and genetic mating system of the pheasant coucal
Generally in birds, the classic sex roles of male competition and female choice result in females providing most offspring care while males face uncertain parentage. In less than 5% of species, however, reversed courtship sex roles lead to predominantly male care and low extra-pair paternity. These role-reversed species usually have reversed sexual size dimorphism and polyandry, confirming that sexual selection acts most strongly on the sex with the smaller parental investment and accordingly higher potential reproductive rate. We used parentage analyses and observations from three field seasons to establish the social and genetic mating system of pheasant coucals, Centropus phasianinus, a tropical nesting cuckoo, where males are much smaller than females and provide most parental care. Pheasant coucals are socially monogamous and in this study males produced about 80% of calls in the dawn chorus, implying greater male sexual competition. Despite the substantial male investments, extra-pair paternity was unusually high for a socially monogamous, duetting species. Using two or more mismatches to determine extra-pair parentage, we found that 11 of 59 young (18.6%) in 10 of 21 broods (47.6%) were not sired by their putative father. Male incubation, starting early in the laying sequence, may give the female opportunity and reason to seek these extra-pair copulations. Monogamy, rather than the polyandry and sex-role reversal typical of its congener, C. grillii, may be the result of the large territory size, which could prevent females from monopolising multiple males. The pheasant coucal’s exceptional combination of classic sex-roles and male-biased care for extra-pair young is hard to reconcile with current sexual selection theory, but may represent an intermediate stage in the evolution of polyandry or an evolutionary remnant of polyandry.
KeywordsExtra-pair paternity Parental care Polyandry Sex-role reversal Sexual size dimorphism
The classic sex roles, defined across animals as male competition and female choice (Darwin 1871; Vincent et al. 1992), are often intimately related to the form of parental care. In species with internal fertilization, such as mammals and birds, females rarely compete for mates but provide most of the parental care. This pattern has long been explained by the ‘cruel bind’ (Trivers 1972): as a result of internal fertilisation females are the first to care for the young giving males the opportunity to leave their mate and offspring to compete for additional matings (Dawkins and Carlisle 1976). These behaviours are then selected for because males seem able to maximise their potential reproductive rate through desertion, while maternal care increases the reproductive success of lone females (Andersson 1994; Clutton-Brock and Vincent 1991; Maynard Smith 1977). These explanations for sex role evolution have now come under increasing scrutiny through theoretical and experimental research (Kokko and Jennions 2008), which highlights how the limitation in additional reproductive opportunities and the influence of ecological factors have not been considered fully in the analysis of sex role evolution. Ecological studies on species with sex roles diverging from the classic pattern are therefore highly informative and sorely needed to enable a more complete understanding of the prevalence of classic sex roles in animals (Kokko and Jennions 2008). In birds, the critical role of female parental care for the evolution of the classic courtship sex-roles in birds is demonstrated by the proverbial exception that proves the rule. Female mating competition is found in less than 5% of avian species, and in all those species, the males provide most parental care (Lack 1968). In fact, the link between reversed courtship roles and reversed parental care is so strong that the term sex-role reversal has sometimes been applied to both behaviours (e.g. Jenni 1974; Owens 2002; Taplin and Beurteaux 1992).
Sex role reversal in birds has long fascinated biologists because of its potential to reveal how a species’ ecology and life history can shape sexual selection (Andersson 1994). A number of attempts have been made to determine common ecological and life history traits that set apart sex-role reversed species from the majority of birds. In addition to predominantly male care, four other factors have been emphasised: (1) ample food supply; (2) precociality; (3) polyandry; and (4) certainty of paternity.
A temporary abundance of food during the breeding season may facilitate the evolution of sex-role reversal (Andersson 1995). Abundant food allows females to produce multiple clutches at little cost, but only if they can delegate parental care to the male. Females then exceed males in their reproductive rate and the operational sex ratio is female-biased, leading to female competition and male choice (Clutton-Brock and Parker 1992; Colwell and Oring 1989).
Until recently, reversal of parental duties was known only in species with precocial young, where the cost of parental care is low (Clutton-Brock 1991). Consequently, all the avian clades in which sex-role reversal dominates show this developmental mode (Cockburn 2006). Following anecdotal reports of male parental care in African black coucals, Centropus grillii (Vernon 1971), a detailed study of this species provided the first unequivocal evidence for male parental care and female competition in a species with altricial young (Goymann et al. 2004, 2005). Female black coucals are 70% heavier than males and compete with each other, mainly by calling, and do not incubate or feed young. This extension of sex-role reversal to altricial species further strengthens the argument for a tight link between mating system and mode of parental care, rather than the mode of development.
Black coucals also demonstrate another behavioural trait that is typical for sex-role reversed birds: simultaneous social polyandry (e.g. Butchart et al. 1999; Oring and Lank 1986; Whitfield 1995). Female black coucals defend territories that encompass the home-ranges of multiple males for which they produce multiple clutches (Goymann et al. 2004). Males on the other hand call infrequently and chase away intruding males only if they approach the female.
The heavy paternal investment of sex-role reversed species should only evolve if male certainty of paternity matches that usually reserved for females (Sheldon 2002; Wright 1998). Accordingly, most sex-role reversed species show extremely low levels of extra-pair paternity (e.g. Delehanty et al. 1998; Oring et al. 1992), and species with extensive male care but classic sex roles are often genetically monogamous, e.g. woodpeckers and sandpipers (Michalek and Winkler 2001; Pierce and Lifjeld 1998). Nonetheless, it is difficult to gauge the importance of low extra-pair paternity rates for the evolution of sex-role reversal, partly because all sex-role-reversed birds discovered so far belong to the non-passerines, and in this group, extra-pair paternity is generally very low with an average rate of only 5.4% of young (Griffith et al. 2002). However, polyandrous black coucals show an exceptional 14.2% extra-pair paternity (Muck et al. 2009), close to the mean found in passerines (17.6%; Griffith et al. 2002).
An abundant food supply, precociality, polyandry or the certainty of paternity are each likely to play a part in the evolution of sex-roles, but we know little about their relative importance or how they interact. This problem is exacerbated by the small number of sex-role-reversed species studied to date and an almost exclusive focus on waders (see review in Muck et al. 2009). Studies on other candidates for sex-role reversal are sorely needed and have the potential to dramatically shift our understanding of the evolution of courtship and parental care in birds (Muck et al. 2009; Voigt and Goymann 2007).
Here, we describe the social and genetic mating system of the pheasant coucal, Centropus phasianinus, an enigmatic species with a combination of features suggesting either classic or reversed sex roles. Duetting and possible monogamy imply classic sex roles in this species, while predominantly male care, female calling, reversed sexual size dimorphism, and its close taxonomic relationship to black coucals suggest sex-role reversal (Higgins 1999; Payne 2005; Taplin and Beurteaux 1992). Specifically, this study aimed to distinguish between three potential mating systems for pheasant coucals: (1) simultaneous polyandry with female competition as found in black coucals; (2) sequential polyandry with female competition; or (3) monogamy and male competition as found in other nesting cuckoos such as the roadrunner, Geococcyx californicus (Calder 1967). We used a combination of call, observational and molecular data to clarify the social mating system, and paternity analysis to establish the genetic mating system.
Materials and methods
Study species and population
The pheasant coucal is a nesting cuckoo with remarkable reversed sexual size dimorphism (Andersson 1995; Payne 2005). Females are about 47% heavier than males (445 vs. 300 g) and slightly larger (wing: 26.9 vs. 24.1 cm; tail: 37.2 vs. 34.6 cm) (Higgins 1999). All incubation and 80% of the feeding of 2–4 young in up to four annual breeding attempts is done by the male (Maurer 2008). Coucals produce two types of presumed territorial calls, descending whoops calls and scale calls, singly or in duets which can be heard over more than 1 km distance. Females produce both calls at a lower acoustic frequency than males (Maurer et al. 2008). Both sexes show a distinctive breeding plumage of black contour and tail feathers and brown, striped wing feathers, but differ slightly in the striping of the primaries. Adults can therefore be sexed in the field by calls, size comparison or pattern of the primaries (Higgins 1999; Maurer et al. 2008). Pheasant coucals inhabit open woodland with dense understorey along the east and north coast of Australia and in New Guinea (Higgins 1999; Payne 2005).
In northern Australia, males build the nests in grass tussocks near the ground or in Pandanus palms, Pandanus spp. (Maurer 2007a). Females usually lay 3–4 (mean 3.56) eggs per clutch with both clutch and egg size increasing over the season (Maurer 2007a). Egg size is only about half the c. 30 g expected for a species of the coucal’s body mass (Rahn et al. 1975). Laying occurs every second day, although the fourth egg may be laid 3 days after the third. The two eggs laid first usually hatch on the same day and hatching is complete after 3 days, suggesting that incubation starts with the second egg (Maurer 2007a). The incubation and nestling periods are extremely short for a bird its size, at only 15 and 12 days, respectively, but fledglings are unable to fly and depend on their parents for at least two more weeks (Maurer 2007a; Taplin and Beurteaux 1992). Despite the short incubation and nestling period, less than half the nests survive to fledging (Maurer 2007a). Nestling and adult population sex ratios are equal to parity (Maurer 2007a).
The research was conducted during the breeding seasons in the austral summer (November–March) of three consecutive years starting in November 2003. The study population consisted of c. 30 adults in an area of approximately 10 km2 near Howard Springs Hunting Reserve, Northern Territory, Australia (12.442°S, 131.102°E). This area of open savannah woodland is dominated by Eucalyptus miniata and dense spear-grass Heteropogon spp. undergrowth that is burnt down in the dry season but reaches 2–3 m in height during the summer monsoon.
Territory location and size
We mapped territories based on individual sightings, and calculated territory sizes in ArcView™ GIS 3.3 from minimum convex polygons around these coordinates (Figs. 1 and 2). Birds were located by their calls or at their nests, and then followed from a distance of 50–200 m for up to 3 h, until they vanished in the high grass or flew out of sight. Visits to each territory were made at least twice a week during the breeding season and between 6 and 62 coucal locations were mapped for each territory using either a Garmin e-trex Vista™ or a Magellan Meridian Gold™ Global Positioning System with accuracies of 5–10 m. New coordinates were recorded whenever the bird had moved more than 50 m into a new part of the territory. Birds that were not marked or individually recognisable were followed from their nests or regular song perches to define territories (Maurer et al. 2008), assuming that only a single male and female visit each nest. This assumption was supported by nest observations of recognisable birds and later by genetic analyses of feathers collected at the nest (Maurer 2008; Maurer et al. 2009).
Social interactions among territorial pheasant coucals appear to be mediated through sex-specific territorial calls (Maurer et al. 2008), as we never observed any physical aggression. We therefore noted all calls heard during 106 standardised dawn chorus sample periods to determine whether a territory was occupied at the time by a male and a female, and to quantify the number of calls given by each sex. Observations lasted for 1 h, starting 15 min before sunrise, from two fixed listening points each within hearing distance of 17 territories (Fig. 1). For each call, we noted the date and time, call type and sex of the caller. We established the territory position of the caller, even when they were out of sight, using a compass and a territory map similar to Fig. 1, which allowed us to narrow down the number of possible territories a bird may be calling from. This information, together with an estimate of the distance of the bird and knowledge about the relative position of its neighbours, usually sufficed to assign a caller unequivocally to a territory, due to the coucals’ preference for widely scattered regular call perches. Unassigned calls were omitted from all analyses. Reliability of the territory assignments was tested in 27 of the 106 observation hours: an assistant tracked down the caller and confirmed the position with the primary observer via two-way radio. All 38 birds located this way had been assigned correctly (Maurer, unpublished data).
DNA sampling, genotyping and parentage analysis
Individuals were genotyped at 2–8 variable microsatellite loci, specific to pheasant coucals (Maurer et al. 2005). Over the three seasons, a total of 24 adults and 108 nestling pheasant coucals from 42 breeding attempts of 30 pairs were genotyped. DNA was extracted from blood samples (all nestlings and 4 adults), feathers (18 adults from 14 nests), or tissue (1 adult) with the CTAB or Ammonium-Acetate method (Nicholls et al. 2000; Weising et al. 1995; Maurer et al. 2009). The PCR products were analysed on a 3100 DNA sequencer (Applied Biosystems, Foster City, CA, USA) and allele sizes were determined automatically and checked manually using GeneMapper 3.7 software (Applied Biosystems) and an internal size standard (GS 500 LIZ). Sex was determined genetically (Griffiths et al. 1998). In total, a genotype with two or more loci was available for at least one of the adults in 21 out of 42 broods.
Due to variation in the amount and quality of DNA extracted from the feather and blood/tissue samples, the number of loci genotyped and thus the certainty of genetic identity differed between samples. Feather samples, with 2–5 loci genotyped, resulted in probabilities of correctly excluding an adult from parentage (exclusion probability) of 0.61–0.8 (one parent known) and 0.8–0.97 (both parents known), whereas for blood samples, with all 8 loci genotyped, the exclusion probabilities were 0.92 and 0.99 for one or two parents known, respectively. The 108 nestlings were genotyped at a mean of 6.85 ± 0.16 loci (range 3–8). The probability that two birds were genetically identical if sampled only at the two least variable loci was 0.12, while for all eight loci the probability of identity was <0.001.
Parentage was assigned manually by aligning the available genotypes of adult pheasant coucals with each other and the young of their nest and the population. This laborious approach was chosen over an automated analysis using paternity analysis software, e.g. CERVUS (Marshall et al. 1998), because by design such software requires a more complete representation of the population and more genetic markers than were available in this study to produce reliable results. To account for the fact that many adult feather samples could only be genotyped at three loci (Maurer et al. 2009), thus reducing the probability of detecting two mismatched alleles, extra-pair parentage data are presented first as a conservative estimate with young mismatching the social parent at two or three loci and then as a maximum value based on young mismatching the adults at a single locus only. Given the very low rates of spontaneous mutation at avian microsatellite loci of between 0.5 and 1.5% (Brohede et al. 2002), this approach can provide a reliable estimate of population-wide levels of extra-pair paternity.
Social and genetic mating system
Given the combined difficulties of catching birds and observing them in dense vegetation, we deduced the social and genetic mating system from a combination of direct observation of territory use, call analysis, observation of nest attendance and molecular evidence on the identity of individuals attending nests. In order to assess the social mating system of pheasant coucals, we compared the adult genotypes within and between territories. This comparison of the parents of successive or neighbouring breeding attempts throughout the breeding season relied on broods for which either both parents or only a single parent were known. If only one of the parents had been sampled, the genotype of the missing parent was inferred from the alleles present in most or all of the young once the alleles found in the known parent had been accounted for.
Territory sizes and percent differences in overlap of territories were assessed in two-tailed paired t tests. Sex differences in calling during different stages were compared between territories using one sample t test with equal distribution as the null hypothesis. t tests were compared for each. 95% confidence intervals are given for all means. Generalised linear mixed models (GLMM) were used to analyse the differences in calling behaviour of male and female pheasant coucals over time of the breeding season. These models controlled for the pair and its particular nesting attempt as random factors. Statistical analysis was conducted in SPSS 14.
Social mating system
Schematic assessment of potential pheasant coucal, Centropus phasianinus, mating systems based on adult genotypes on each territory and for each brood
Mating system and definition
Simultaneous polyandry: ♀ territories contain multiple ♂ territories
Broods on neighbouring ♂ territories share ♀ genotype
Each of 14 neighbouring broods had a different ♀ genotype
Sequential polyandry: ♀ changes ♂ territory between breeding attempts
Consecutive broods on same ♂ territory differ in ♀ genotype
The same ♀ laid the clutch for a ♂ 2nd brood in 7 of 8 cases
Social monogamy: ♀♀ and ♂♂ maintain a common territory
Consecutive broods on same ♂ territory share ♀ genotype
The same ♀ laid the clutch for a ♂ 2nd brood in 7 of 8 cases. No ♀ laid for her ♂ and neighbour
Territory observations and genetic analyses show that the birds are not simultaneously polyandrous. Female ranges did not encompass the territories of more than one male, and there was no genetic evidence for females laying clutches in territories of neighbouring males.
The territories of female pheasant coucals each overlapped substantially with the territory of only a single male (Figs. 1 and 2). Similarly, each male territory overlapped those of only one female but exceeded it in size in season two and tended to be larger in season three (paired sample t test season two: territory size male = 18.45 ha ± 5.3, range: 7.22–41.25 ha, female = 14.71 ha ± 6.0, range: 3.28–46.45 ha; t12 = 2.4, p = 0.04; season three: territory size male = 28.89 ha ± 8.0, range: 13.45–56.86 ha, female = 23.35 ha ± 6.8, range: 9.65–43.99 ha; t9 = 2.0, p = 0.07). Genetic data show that some of the pairs contained the same individuals in both seasons.
The genetic analyses are consistent with social monogamy. A comparison between the parental genotypes of neighbouring broods was possible for seven territories in each of two breeding seasons. In the first season; both adult samples were available for two broods and just the paternal genotypes for five. In the second season; both parental genotypes were known for four broods and the paternal genotypes for three broods. In none of these seven cases did the known or inferred maternal genotype match the known or inferred maternal genotype of a neighbouring brood.
We found no evidence for sequential polyandry, as females did not move between male territories, and there was no case of a succession of males on a female’s territory. Individuals of each sex were sighted repeatedly on only one territory throughout the breeding season (Figs. 1 and 2). These observations are supported by a genetic analysis of sequential broods cared for by the same male on his season-long territory. We were able to test for continuity of the female genotype throughout the breeding season on six territories, four of which had two nesting attempts and two had three nesting attempts. Both parental genotypes were known for three of those territories, while for the other three territories the female genotype could be inferred from the genotypes of the young and the male. Together, these broods allowed eight comparisons between sequential broods, and in only one of these comparisons (12.5%) did the female genotype differ between broods. Death or dispersal of the first female may explain this rare occurrence better than sequential polyandry, as the female genotype was not recovered from any other nest.
Both sexes were frequently detected during regular visits made to 11 coucal territories in the third season. Each territory was visited between 26 and 82 times (mean = 51 visits ±8.3), and on all but two territories, one male and one female were heard or seen during more than 50% of the visits in which any bird could be detected (mean percentage when both sexes detected = 62.7% of visits ±6.4, range 38.5–82.4%, n = 615 visits; Fig. 3). Given the difficulty of detecting birds in dense vegetation, it is likely that birds were often present but undetected. Supporting this idea, individually identifiable birds of each sex were seen repeatedly on the same territory but never on other territories (Figs. 1 and 2) or at different nests (Maurer 2008).
The number of observations on which the male of a territory called (or initiated a duet), minus the number of observations on which its female called (or initiated a duet, respectively) given as means (±SE) per stage of the breeding cycle. A positive mean value implies a prevalence of male calls or male initiated duets. The t tests assess whether the call number differed significantly from equity
Observations, mean (SE)
Laying, n = 8
t7 = 4.4
Duets, joining sex
t7 = −0.7
Incubation, n = 9
t8 = 2.8
Duets, joining sex
t8 = −1.2
Nestling, n = 6
t5 = 1.2
Duets, joining sex
t5 = 0.8
Fledgling, n = 6
t5 = 2.7
Duets, joining sex
t5 = −0.9
Pheasant coucals had a high rate of extra-pair young. Overall, we had parental genotypes for 21 broods with 59 young, and 11 (18.6%) of young and 10 (47.6%) of broods had at least one extra-pair young. Using a single mismatch as exclusion criterion, 19 (32%) of young in 16 (76%) of broods had at least one extra-pair young. No extra-pair sires were identified, as we could not uniquely match any of the genotyped neighbouring males to the respective extra-pair young at all loci.
Frequency of extra-pair paternity versus types of adult samples and for different combinations of available parental genotypes using a single mismatch criterion
Total both parents known
Total male known
In two broods where both parents were genotyped, a single nestling mismatched the genotype of the putative mother at locus CP11, suggesting intra-specific brood-parasitism or mutation. Unfortunately, these females were genotyped from feathers, so no DNA remained to test for genotyping error in a repeat analysis. Both mismatches were excluded from all analyses of extra-pair paternity.
Exceptionally for birds, pheasant coucals combine social monogamy and predominantly male care for demanding altricial young with male calling competition (Clutton-Brock 1991; Cockburn 2006). Despite these substantial investments, males suffer extra-pair paternity in approximately half their broods, one of the most extreme rates of extra-pair paternity found in any non-passerine (Griffith et al. 2002). The combination of these traits is difficult to explain within the framework of sexual selection theory (Andersson 1994). However, the release from the assumption that a more male-biased operational sex ratio leaves increased male–male competition as the only alternative for males to increase their overall reproductive success (Kokko and Jennions 2008) may help explain the high rates of parental care found in many socially monogamous bird species with high rates of extra-pair paternity.
Social mating system
Given the difficulty of studying a cryptic and reclusive species like the pheasant coucal, our strategy was to assess which of the candidate social mating systems—simultaneous polyandry, sequential polyandry or monogamy (Andersson 1995)—was best able to explain the observational and genetic data. Only social monogamy is consistent with all the available evidence.
Evidence for social monogamy
Pheasant coucals appear to be socially monogamous rather than polyandrous: (1) the territories of 25 males and females overlapped markedly with only one bird of the opposite sex; (2) one bird of each sex called from each territory during the majority of morning choruses (57%), even though females called less frequently than males; (3) focal searches of individual territories consistently detected not more than one adult of each sex, and individually identifiable birds were never seen outside their own territory; (4) each female genotype occurred only in broods on one male’s territory for a total of 14 comparisons of neighbouring males; and (5) sequential clutches on a male’s territory were laid by the same genetic mother, except in one of eight cases, arguing against sequential polyandry.
Results published previously further support the conclusion that pheasant coucals are socially monogamous. First, nestling care was provided by a single male and female (Maurer 2008). Second, genetic analyses of feather samples collected near the nest also support the presence of a single male and female (Maurer et al. 2009). Third, duetting is typical for species with a strong pair-bond (Hall 2004, 2009) and occurs regularly in pheasant coucals (Maurer et al. 2008), but not in the polyandrous black coucals (Goymann et al. 2004).
Mating system data for other coucals and nesting cuckoos are scant. Pheasant coucals differ from the only other coucal species studied in detail, the African black coucal, in which simultaneously polyandrous females call to defend territories encompassing those of up to four males. Anecdotal evidence from other coucal species suggests a variety of mating systems, but with monogamy widespread (Andersson 1995; Payne 2005). The nesting cuckoos of the new world show monogamy and male competition, as in the roadrunner (Calder 1967), but polyandry has been observed at least once (Ralph 1975), while cooperative breeding cuckoos show female competition in the form of egg tossing (Payne 2005).
Surprisingly, male pheasant coucals gave more long-range calls than females in all stages of the breeding cycle except the nestling period when both sexes called equally rarely, suggesting that males are the more competitive sex. The design and context of these calls is consistent with a double function in territorial defence and mate attraction (Klump 1996; Maurer et al. 2008; Staicer et al. 1996), both components of competition within the sexes. This suggests classic sex roles despite extensive male care and reversed sexual size dimorphism. Female competition may still occur but could be restricted primarily to quiet and hidden, physical interactions.
Further support for classic sex roles in the pheasant coucal comes from a comparison with their polyandrous congener. The calling behaviour of African black coucals is almost a complete reversal of that found in the pheasant coucal. Females call more than males and compete for territories, while males mainly compete for matings through physical interaction (Geberzahn et al. 2009; Goymann et al. 2004).
Genetic mating system
Pheasant coucals have a high rate of extra-pair paternity (18.6% of nestlings) that exceeds the average rate of extra-pair paternity in socially monogamous non-passerines and even passerines (5.4 and 17.9% of nestlings, respectively; Griffith et al. 2002). Overall, 47.6% of coucal broods examined contained at least one extra-pair offspring. These values are surprising when compared to the low extra-pair paternity rates of other duetting species (Gill et al. 2005; Hall and Magrath 2000), but match the 14.2% extra-pair young found in the pheasant coucal’s polyandrous congener with male-only care, the African black coucal (Muck et al. 2009). Due to the sampling difficulties experienced for adult pheasant coucals, these values could represent an over- or under-estimate. On the one hand, our exclusion criteria of two or more mismatches or a single mismatch, while not without precedent (e.g. Gill et al. 2005), could include mismatches caused by rare spontaneous microsatellite mutations (Brohede et al. 2002) or, more likely, scoring error, which increased the estimate of extra-pair paternity. If scoring errors alone were the source of our high rate of extra-pair paternity, we should, however, have found young to mismatch the maternal genotypes as much as they did the paternal genotypes. On the other hand, the lack of full parental genotypes for feather-sampled adults may have caused some extra-pair young to go undetected.
Paradoxically, it is the male’s extreme parental investment—the very reason why we expect genetic monogamy in pheasant coucals—which may offer the best proximate explanation for their high rate of extra-pair paternity. Males start incubation after the second egg is laid (Maurer 2008), and so like black coucals are unable to guard their female when the remainder of the clutch is fertilised (Muck et al. 2009). A female could then obtain extra-pair matings unhindered and unnoticed by her social mate. Neither sex has been observed to make extra-territorial forays for matings (Maurer 2007b), so that it remains unclear whether the male (as in black coucals) or the female solicits the extra-pair copulations and when (Goymann et al. 2004).
Adaptive function and evolution of the coucal breeding system
The sex that provides most of the parental care in birds usually engages little in territorial defence or competition (Andersson 1994), but in a few polygamous bird species both sexes compete for territories independently. (e.g. Heinsohn et al. 2005; Langmore 1998; Butchart et al. 1999; Emlen and Wrege 2004). The pheasant coucal’s combination of male care and competition is therefore not without precedent, but it appears exceptional in its degree. In the following, we discuss aspects of coucal biology that may help explain the function and evolution of this breeding system.
Territory size and mating system
Two contrasting hypotheses illustrate how territory size could explain the evolution of polyandry and male parental care. Owens (2002) found that low nesting density and the associated lack of re-mating opportunities promotes male-only care and polyandry in bird lineages. By contrast, in the black coucal, small male territories and high breeding density may promote polyandry and male care, because it allows females to monopolize more than one male (Goymann et al. 2004). Our data support Goymann et al.’s (2004) hypothesis as male pheasant coucals have large territories, which presumably make it hard for females to control more than one male territory and become polyandrous. This interpretation could suggest that pheasant coucals have lost polyandry but not male-biased care either because of conditions at the study site or during the evolution of the species.
The discrepancy in territory size between black and pheasant coucals may be in compensation for a lower density of resources in the pheasant coucal. While food abundance for either species has not been measured directly, our pheasant coucal population showed possible evidence of food limitation in form of staggered egg laying and frequent loss of the youngest nestlings, probably to starvation (Maurer 2007a). These findings lend support to the idea that food abundance influences the mating system in coucals (Andersson 1995, 2005).
Reversed size dimorphism
The reversed sexual size dimorphism of pheasant coucals is puzzling in a bird species with classic sex-roles, especially since no traditional explanation for this reversal applies to pheasant coucals (Andersson 1994, 2005; Blanckenhorn 2005). They lack display flights, which may select for increased agility and reversed size dimorphism in waders and raptors (Andersson and Norberg 1981), and do not have female-only incubation and nest defence (Andersson and Norberg 1981; Maurer 2008; Szekely et al. 2000). Andersson (1995) thus proposed three alternative explanations for reversed size diorphism in coucals: (1) female rather than male competition, as now shown for black coucals but unlikely for pheasant coucals; (2) female resource storage for continued egg production, which is unlikely to be necessary in the monogamous pheasant coucal with 2–4 clutches of four, relatively small eggs laid in 2 day intervals (Owens 2002; Rahn et al. 1975; Slotow 1996), especially given that males can raise the brood on their own; and (3) selection for small male size to improve eneregetics which may fit the male pheasant coucal’s provisioning effort and wing-load index (Andersson 1995; Maurer 2008).
Alternatively, Blanckenhorn (2005) proposed ‘the ghost of a sexual size dimorphism evolution past’ as an intriguing explanation for size dimorphism with no obvious current selective explanation. Accordingly, reversed size dimorphism in pheasant coucals may have resulted from a mating system of sex-role reversal in the species’ evolutionary history. Such previous polyandry may also help explain other aspects of the pheasant coucal’s unusual mating system, for instance the limited female care and a high rate of extra-pair paternity.
Pheasant coucal females may benefit from extra-pair copulations through fertility assurance, genetic benefits or direct benefits (Birkhead and Møller 1992; Griffith et al. 2002; Petrie and Kempenaers 1998).
Fertilisation in pheasant coucals appears problematic, and this may select for extra-pair copulations. At least one egg failed to hatch in 30% of nests (Maurer 2007a). If this was caused by infertility rather than embryo death, it could be a result of the hormonal changes that males undergo for incubation. In male black coucals, testosterone levels decreased between the mating and the feeding stage (Goymann and Wingfield 2004), and decreasing testosterone levels could diminish sperm production (Jones and Lin 1993), although the testosterone levels of male black coucals during feeding appear sufficient to ensure sperm production. Females may then actively seek extra-pair copulations with males that do not currently care for a brood, to ensure that their eggs are fertilised.
Genetic and direct benefits might also help explain extra-pair paternity in coucals. Females cannot raise young without male help, and so some will be forced to accept partners of low genetic quality or else forgo reproduction. These females may then use extra-pair copulations with superior but already bonded males to obtain better quality offspring. Furthermore, pheasant coucals may engage in extra-pair copulation to benefit from additional courtship feeding (Maurer 2007b), which can increase a female’s egg-production and reproductive success (Carlson 1989; Nisbet 1973).
Why do male pheasant coucals provide the majority of care, when they are regularly cuckolded? Males should be selected to continue care, if present and prospective future broods contain the same proportion of extra-pair young (Maynard Smith 1978; Westneat and Sherman 1993). This could be the case in coucals, where females can freely seek extra-pair fertilisations at least for the last eggs of the clutch, when males are already incubating (Muck et al. 2009). On the other hand, young hatched from the last laid eggs were least likely to fledge (Maurer 2007a) and so the proportion of extra-pair young fledging from a nest may be small and reduce the cost of cuckoldry to the male.
Our findings show that male pheasant coucals engage in more vocal competition behaviour than females, although they provide most of the parental care. Surprisingly, male pheasant coucals also suffer a high frequency of extra-pair paternity despite social monogamy. These results do not fit comfortably into the current theoretical framework of sex-role evolution. The contrasting mating systems of African black coucals and pheasant coucals demonstrate the diversity of social systems long suspected in coucals, and caution against assuming a universal link between sex-roles and parental care (Andersson 1994; Clutton-Brock 1991). In particular, the unusual mating system of pheasant coucals may illustrate a transitional stage to or from polyandry, and further studies could clarify the mechanisms behind the evolutionary transitions to polyandry and male care in birds and the role of the female (Andersson 1995, 2005; Cockburn 2006; Owens 2002).
We thank S. Musgrave, A. Quellmalz, S. Quinlan, C. Smith, M. Starling, R. Noske and W. Goymann with planning and conducting the fieldwork, and N. Beck, S. Cooney, N. Langmore and two anonymous reviewers for comments on earlier drafts. The study received support from the Stuart Leslie Bird Research Award, the Cayley Memorial Scholarship, the Ingram trust, and The North Australia Research Unit. The work was conducted under permits from Parks Northern Territory (16973) and the Ethics Committee of the Australian National University (F.BTZ.56.03) and complies with the laws and regulations of the Commonwealth of Australia.
- Andersson M (1994) Sexual selection. Princeton University Press, PrincetonGoogle Scholar
- Birkhead TR, Møller AP (1992) Sperm competition in birds: evolutionary causes and consequences. Academic, LondonGoogle Scholar
- Calder WAI (1967) Breeding behavior of the roadrunner, Geococcyx californianus. Auk 84:597–598Google Scholar
- Clutton-Brock TH (1991) The evolution of parental care. Princeton University Press, PrincetonGoogle Scholar
- Hicks RK, Restall R (1992) Pheasant coucal Centropus phasianinus attacking birds caught in a mist net. Muruk 5:143Google Scholar
- Higgins PJ (1999) Handbook of Australian, New Zealand and Antarctic birds. Volume 4: parrots to Dollarbirds. Oxford University Press, MelbourneGoogle Scholar
- Jenni DA (1974) Evolution of polyandry in birds. Am Zool 14:129–144Google Scholar
- Jones RC, Lin M (1993) Spermatogenesis in birds. Oxf Rev Reprod Biol 15:234–264Google Scholar
- Klump GM (1996) Bird communication in a noisy world. In: Kroodsma DE, Miller EH (eds) Ecology and evolution of acoustic communication in birds. Cornell University press, Ithaca, pp 321–338Google Scholar
- Lack D (1968) Ecological adaptations for breeding in birds. Methuen, LondonGoogle Scholar
- Maurer G (2007a) Ecology and evolution of sex-roles in the pheasant coucal Centropus phasianinus. Australian National University, CanberraGoogle Scholar
- Maurer G (2007b) Love is in the air: arboreal copulations in the pheasant coucal Centropus phasianinus. North Territory Nat 19:48–50Google Scholar
- Maynard Smith J (1978) The evolution of sex. Cambridge University Press, CambridgeGoogle Scholar
- Oring LW, Lank DB (1986) Polyandry in spotted sandpipers: the impact of environment and experience. In: Rubenstein DR, Wrangham RW (eds) Ecological aspects of social evolution. Princeton University Press, Princeton, pp 21–42Google Scholar
- Payne RB (2005) Bird families of the world: the Cuckoos. Oxford University Press, OxfordGoogle Scholar
- Pierce EP, Lifjeld JT (1998) High paternity without paternity assurance behaviour in the purple sandpiper, a species with high paternal investment. Auk 115:602–612Google Scholar
- Rahn H, Paganelli CV, Ar A (1975) Relation of avian egg weight to body weight. Auk 92:750–765Google Scholar
- Staicer CA, Spector DA, Horn AG (1996) The dawn chorus and other diel patterns in acoustic signalling. In: Kroodsma DE, Miller EH (eds) Ecology and evolution of acoustic comminucation in birds. Cornell University Press, IthacaGoogle Scholar
- Trivers RL (1972) Parental investment and sexual selection. In: Cambell B (ed) Sexual selection and the descent of man. Aldine, Chicago, pp 136–179Google Scholar
- Weising K, Wolff K, Nybom H, Meyer W (1995) DNA fingerprinting in plants and fungi. CRC, Boca RatonGoogle Scholar