Territorial intrusions and copulation patterns in red kites, Milvus milvus, in relation to breeding density

Two main paternity assurance strategies are generally found in birds: mate guarding and frequent copulations. The latter is expected particularly in species such as raptors that cannot guard their mates efficiently because of ecological constraints, such as frequent courtship feeding. I investigated the prelaying behaviour of red kites, in which the males courtship feed. I compared pair behaviour in situations of varying breeding density and simulated male territorial intrusions by presenting decoys. Males' certainty of paternity was likely to decrease with increasing breeding density, because of the proximity of other males and more frequent male territorial intrusions during the presumed fertile period. The percentage of time spent by males within their breeding territory during the prelaying period was positively related to the number of close breeding neighbours, suggesting territory surveillance and mate guarding. The kites copulated frequently and over a long period. Copulation frequency prior to and during laying increased with breeding density, and in isolated pairs in response to simulated male territorial intrusions. The results support the idea of paternity assurance through frequent copulations during the presumed fertile period of the female, and suggest that early copulation activity is related to functions other than fertilization, such as pair bonding or mate assessment. Copyright 2000 The Association for the Study of Animal Behaviour.     

Breeding density, cuckoldry risk and copulation behaviour during the fertile period in raptors: a comparative analysis

Extrapair copulations (EPCs) and fertilizations (EPFs) occur in many socially monogamous bird species, including raptors. In this group of species, males invest heavily in reproduction through the feeding of female and young, so the cost of cuckoldry is particularly high. The feeding of females by males, characteristic of most species, conflicts with mate guarding, so raptors are expected to use frequent within-pair copulations (WPCs) for paternity assurance. In this study, I reviewed information on copulation behaviour of diurnal raptors, and used regression analyses and phylogenetic comparative analyses to investigate relations between density, EPC frequency and WPC behaviour. EPCs occurred in most raptor species studied (68%, N=19), mostly during the presumed fertile period of females. EPC frequency, measured as the percentage of females engaging in EPC but not the percentage of extrapair copulations, was positively related to breeding density. The rate and relative duration of WPCs were also positively related to breeding density, but only WPC rate was positively related to EPC frequency (percentage of females). Frequently copulating species had relatively larger testes, underlying higher sperm production capacities. The results support the hypotheses that sperm competition intensity increases with breeding density, and that male raptors rely on frequent copulations to ensure paternity. Despite the risk, EPFs rarely occurred in raptors, suggesting efficient paternity guards.     

Territoriality as a paternity guard in the European robin, Erithacus rubecula

To investigate the relative importance of paternity defences in the European robin we used behavioural observations, simulated intrusions and temporary male removal experiments. Given that paired males did not increase their mate attendance, copulation rate or territory size during the female's fertile period, the most frequently quoted paternity assurance strategies in birds were absent. However, males with fertile females sang and patrolled their territories more regularly, suggesting that territorial motivation and vigilance were elevated when the risk of cuckoldry was greatest. In addition, there was a significant effect of breeding period on response to simulated intrusions: residents approached and attacked freeze-dried mounts more readily in the fertile period. During 90-min removals of the pair male in the fertile period, neighbours trespassed more frequently relative to prefertile and fertile period controls and appeared to seek copulations with unattended females. When replaced on their territories, males immediately increased both song rate and patrolling rate in comparison with controls. We propose that male robins sing to signal their presence, and increase their territorial vigilance and aggression in the fertile period to protect paternity. Copyright 2000 The Association for the Study of Animal Behaviour.     

Explicit experimental evidence for the role of mate guarding in minimizing loss of paternity in the Seychelles warbler

Extra-pair copulations (EPCs) (copulations outside the pair bond) resulting in extra-pair fertilizations (EPFs) are widespread in birds. To increase reproductive success, males should not only seek EPCs, but also prevent their females from having EPFs. Male Seychelles warblers (Acrocephalus sechellensis) follow their partner closely during the period when these females are most receptive (fertile period). The Seychelles warbler is the first species to offer explicit experimental evidence that mate guarding functions as paternity guarding: in territories where free-living males were induced to stop mate guarding during the pair female''s fertile period, the rates of intrusions by other males and successful EPCs (male mounting female) were significantly higher than those observed in the control group and in the absence of mate guarding the frequency of successful EPCs increased significantly with local male density. Male warblers do not assure their paternity through frequent copulations to devalue any sperm from other males: males do not copulate with their partners immediately following a successful EPC obtained by their partners, the frequency of successful within-pair copulations does not increase with the frequency of successful EPCs and females initiate all successful copulations and are capable of resisting copulation attempts.     

Factors influencing mate guarding and territory defence in the stitchbird (hihi)

Socially monogamous male birds are predicted to maximise their reproductive success by pursuing extra-pair copulations (EPCs) while engaging in anti-cuckoldry behaviour such as mate guarding. In the stitchbird, Notiomystis cincta, high levels of forced EPCs and a high proportion of nestlings resulting from extra- pair fertilisations lead to the prediction that males of this species should exhibit intense paternity guarding behaviours. While studying an isolated stitchbird population on Tiritiri Matangi Island New Zealand (3636'S, 17453'E), I collected daily behavioural data throughout the breeding season from 15 males in 2000/01 and 27 males in 2001/02. In this study, male stitchbirds demonstrated clear paternity guarding by exhibiting: (1) an increased likelihood of being close to their mate during her fertile period, (2) an increased initiation of mate contact during her fertile period, (3) switching from site-specific territorial defence during the pre-fertile period to defending an area centring on the their female partners location during her fertile period, and (4) an increased following of the female to communal feeding sites outside the territory during her fertile period. For polygynous males, mate guarding and territorial defence were conditional on which of their females was fertile. Additional evidence supporting the hypothesis that mate guarding in this species is a form of paternity assurance, rather than protection from harassment, is that males protected their partner from harassment by other stitchbird males but did not intervene when females were harassed by male bellbirds, Anthornis melanura. While mate-guarding intensity in many species is conditional on the stage of female fertility, male stitchbirds also modified their behaviour depending on the location of the female and the rate of intrusions by extra-pair males. Resident males adopted a best-of-a-bad-job tactic when they were unable to locate their female by defending an area around her last known location. Furthermore, when the rate of intrusions by extra-pair males increased they traded-off the area they could defend within their territory against their ability to guard the female. Territory takeovers were uncommon, but when they did occur older males displaced younger males and healthy birds displaced sick ones. Contrary to the prevailing view that mate guarding is a male response to female infidelity, male stitchbirds appear to use mate guarding primarily to prevent paternity losses from forced EPCs. Future assessments of mate guarding function should consider the possibility that mate guarding involves a combination of conflict and co-operation between the sexes.     

The effect of experimental male removals on extrapair paternity in the wheatear, Oenanthe oenanthe

To examine the role of the pair male in ensuring paternity in the wheatear, we removed pair males for 24 h during the fertile period (days 0 to +1 in 1992 and days -5 to -1 in 1993, where day 0=first egg date). Control males were removed during incubation. The frequency and duration of intrusions, and the frequency of extrapair copulations (EPCs), increased during removals in the fertile period, but not during incubation. The frequency of extrapair paternity (EPP) was marginally higher (25%) in experimental than control broods and the presence of the pair male seemed particularly important in ensuring paternity in those days immediately preceding laying. Females were selective over which males they copulated with in the absence of their mate and rejected the majority of attempted EPCs. Only extrapair males in better body condition than the removed pair male eventually gained successful matings. Experimental males in poorer body condition were also more likely to have EPP in their brood than experimental males in better body condition. Female copulatory behaviour was the most important factor in determining patterns of paternity, and mate-guarding behaviours by the pair male appeared to limit the females' opportunities to engage in EPCs by reducing the frequency and duration of intrusions, rather than to control female behaviour directly. Copyright 1999 The Association for the Study of Animal Behaviour.     

Responses of territorial and floater male Red‐winged Blackbirds to models of receptive females

Red‐winged Blackbirds (RWBL; Agelaius phoeniceus) have a polygynous mating system and, because territorial males commonly have harems of two to five females, some second‐year (SY) and after‐second‐year (ASY) males do not establish nesting territories, but become floaters. Previous studies have revealed high rates of extra‐pair copulations in this species and that sexually mature male floaters and territory owners do not differ in size, testosterone levels, or reproductive capability, suggesting that floaters may occasionally gain paternity. During May and June 2008, we observed the behavioral responses of floater males to taxidermic mounts (models) of female RWBL placed in a precopulatory position. Floaters intruded into territories during 46% of model presentations, with 20% of intrusions by ASY floaters and 80% by SY floaters. During intrusions, ASY floaters attempted to copulate with models 93% of the time compared to 80% for SY floaters. Copulations were successful during 30% of attempts by ASY males and 25% of attempts by SY floaters. The frequency of intrusions by ASY and SY floaters, attempted copulations by SY floaters, and successful copulations by ASY floaters increased as territorial males spent more time off their territories. Responses of floater males toward models in our study suggest that floater male RWBL attempt to exploit available breeding opportunities. The lack of evidence for extrapair young (EPY) fathered by floater male RWBL in previous studies, combined with our results indicating that the presence of territorial males limits floater intrusions, copulation attempts, and successful copulations, suggests that the reproductive success of floater males is limited in part by the aggressive behavior of territorial males.     

Population density and the intensity of paternity assurance behaviour in a monogamous wader: the Curlew Numenius arquata

We compared paternity assurance behaviour and related displays in high (1.6 pairs perkm²) and low (6.7 pairs per km²) density populations of the Curlew Numenius arquata breeding on arable farmland in western Finland. There was little evidence of individuals pursuing extra-pair copulations or males exhibiting paternity assurance behaviour. Furthermore, there was no variation in the frequency of intrusions or in the intensity of paternity assurance behaviour relative to lay date. However, intrusions and approaches to the pair female by extra-pair males were more frequent at high breeding density, and pair males remained closer to their female, followed her more and exhibited a higher frequency of copulatory behaviours in the high-density population. Therefore, high breeding density appeared to increase opportunities for individuals to copulate outside the pair-bond and resulted in more intense male paternity guards. Male song flight displays were also more frequent in the high-density population. Territorial absences by males were infrequent in both areas. The increased frequency of intrusions and interactions between non-pair individuals (male-male and male-female) at high density were probably the major factors in explaining area differences in the intensity of paternity assurance and territorial behaviour.     

Decoy presentations as a means to manipulate the risk of extrapair copulation: an experimental study in a semicolonial raptor, the Montagu's harrier (Circus pygargus)

Mate guarding and frequent copulations are two alternative paternity assurance strategies found in birds. In species with intensecourtship feeding, like raptors, the "frequent copulation"strategy is expected because male food provisioning conflictswith mate guarding. We evaluated experimentally the paternityassurance behavior of a semicolonial raptor, the Montagu'sharrier Circus pygargus, using decoy presentations to simulateterritorial intrusions. Breeding pairs were exposed to maleand female decoys at different periods during the female's reproductive cycle. Agonistic responses to decoys were intra-sexual,and the timing and intensity of male attacks toward male decoyssupported responses related to the risk of extrapair copulation(EPC): Male aggression peaked during the presumed fertile periodand almost disappeared after clutch completion. During thefertile period, copulation rate was significantly higher, andcopulations lasted longer, during male decoy presentations than during controls. Males also spent more time close to the femaleduring male decoy presentations compared to controls, bothduring the early prelaying and fertile periods, but not duringincubation. In the fertile period, males also increased presencetime close to the female in the hour following the removal of the male decoy. Conversely, female decoy presentations hadno significant effect on copulatory behavior or male presencetime. These results showed that the risk of EPC can be experimentallymanipulated by the means of decoy presentations, simulatingmale territorial intrusions, and that male Montagu's harriersincrease their short-term copulation frequency and female surveillancewhen they perceive themselves at an increased EPC risk.     

The adaptive significance of non-contact mate guarding by males of the dragonfly,Nannophya pygmaea Rambur (Odonata: Libellulidae)

InNannophya pygmaea, ovipositing females were frequently disturbed by conspecific males. Disturbed females often copulated with one of these males or flew away from the pool. Females which flew away from the pool due to male disturbance often returned later the same day and mated with different males. A territorial male would guard his ovipositing mate by hovering above her, presumably trying to prevent her from moving out of his territory. A non-territorial male would also guard his mate in a similar way, both at a vacant water area which was not occupied by any territorial males, or within the territory of a resident male. In addition, both territorial and non-territorial males chased intruding males in an attempt to prevent their mates from being stolen. Territorial males defended their mates better than non-territorial males. Both males and females often mated more than once in the course of a single day. Some territorial males copulated with a new female while another mate oviposited in their territories. This observation supported the “multiple mating hypothesis” proposed by Alcock (1979) and Uéda (1979) but other evidence suggested that this is an inadequate explanation for the non-contact guarding ofN. pygmaea.     

Mate switching and copulation behaviour in King Penguins

Extra‐pair paternity (EPP) in monogamous birds may result from either extra‐pair copulations (EPCs) or mate switching. In this study of King Penguins in South Georgia, we observed no EPCs at all, an effect of very efficient mate guarding. Onshore males fast and need not divert attention to foraging or to defending nest or territory, as this species has neither. However, we found that mate switching was common. On average 38% (range: 29%–56%; three years pooled) of the birds established pair bonds with at least one initial partner before switching to the partner they bred with (i.e. the “pair mate”). Of the observed copulations of 44 studied females, 22% were with initial partners and 78% with the pair mate. This and the high proportion of mate switching suggest that roughly 10% of the females could have received sperm from males other than the pair mate. The average copulation frequency was 0.026 h^?1, resulting in an estimated 8.2 copulations per clutch (which consists of one egg). That more copulations than necessary for fertilisation occur suggests that males try to protect paternity by sperm competition, and that this is a result of the potential for EPP due to mate switching in King Penguins. All observed copulations except one took place between days 13 and 5, with the peak 7.5 days prior to egg‐laying. The birds found their pair mates (often not the same as in the previous year) on average about 10 days before egg‐laying, and always established themselves at the outskirts of the colony about 8 days before egg‐laying. Thus, most copulations occurred around the time the birds joined the colony. We suggest that it is adaptive to obtain a breeding spot early, because the colony will grow and pairs joining later will protect the offspring. Additionally, we suggest that early copulation outside the colony is adaptive because of the risk of failing to fertilise the egg when copulating among aggressive neighbours inside the dense colony. Based on these two arguments we suggest a “safe place hypothesis” to explain the early copulation peak in King Penguins.     

Mate guarding in the Seychelles warbler is energetically costly and adjusted to paternity risk

Males may increase their fitness through extra-pair copulations (copulations outside the pair bond) that result in extra-pair fertilizations, but also risk lost paternity when they leave their own mate unguarded. The fitness costs of cuckoldry for Seychelles warblers (Acrocephalus sechellensis) are considerable because warblers have a single-egg clutch and, given the short breeding season, no time for a successful replacement clutch. Neighbouring males are the primary threat to a male's genetic paternity. Males minimize their loss of paternity by guarding their mates to prevent them from having extra-pair copulations during their fertile period. Here, I provide experimental evidence that mate-guarding behaviour is energetically costly and that the expression of this trade-off is adjusted to paternity risk (local male density). Free-living males that were induced to reduce mate guarding spent significantly more time foraging and gained significantly better body condition than control males. The larger the reduction in mate guarding, the more pronounced was the increase in foraging and body condition (accounting for food availability). An experimental increase in paternity risk resulted in an increase in mate-guarding intensity and a decrease in foraging and body condition, and vice versa. This is examined using both cross-sectional and longitudinal data. This study on the Seychelles warbler offers experimental evidence that mate guarding is energetically costly and adjusted to paternity risk.     

The couple that sings together stays together: duetting,aggression and extra-pair paternity in a promiscuous bird species

When individuals mate outside the pair bond, males should employ behaviours such as aggression or vocal displays (e.g. duetting) that help assure paternity of the offspring they care for. We tested whether male paternity was associated with aggression or duetting in the red-backed fairy-wren, a species exhibiting high rates of extra-pair paternity. During simulated territorial intrusions, aggression and duetting were variable among and repeatable within males, suggesting behavioural consistency of individuals. Males with quicker and stronger duet responses were cuckolded less often than males with slower and weaker responses. In contrast, physical aggression was not correlated with male paternity. These results suggest that either acoustic mate guarding or male–female vocal negotiations via duetting lead to increased paternity assurance, whereas physical aggression does not.     

Extrapair paternity in the common sandpiper, Actitis hypoleucos, revealed by DNA fingerprinting

The prevalence of extrapair paternity in many socially monogamous passerines has not been mirrored in most monogamous nonpasserines studied to date. Here, we investigated the reproductive behaviour of a socially monogamous shorebird, the common sandpiper, using multilocus DNA fingerprinting. Given the high level of paternal care in the species, and the likely high costs in allocating care between kin and nonkin in species with precocial young, we predicted low levels of extrapair paternity similar to other monogamous shorebirds. We found the social mating system to be predominantly monogamous although one polyandrous pairing was identified. Of 83 offspring from 27 broods, 13 (15.7%) young from five (18.5%) broods were identified as being extrapair. There was no evidence of intraspecific nest parasitism or quasiparasitism. In this population, territorial intrusions were carried out largely by males but did not appear to be related to seeking extrapair copulations (EPCs). Seventy copulation attempts were observed and most were within-pair (84%). Six of eight EPC attempts occurred outside the territory of the female's social mate. Copulation rates were significantly higher just before and during egg laying than at other times during the study. At least two females that reared extrapair young had associated with males other than their eventual mates on arrival, suggesting that some females use rapid mate switching as a mating tactic, facilitated perhaps by the asynchronous arrival among both sexes in this population. Why some female sandpipers mate promiscuously remains unresolved.     

Multiple mating and repeated copulations: effects on male reproductive success in red flour beetles

In many insects, both sexes mate multiple times and females use stored sperm for fertilizations. While males frequently engage in two distinct behaviours, multiple mating (with different females) and repeated copulations (with the same female), the reproductive consequences of these behaviours for males have been quantified for only a few species. In this study, males of the red flour beetle, Tribolium castaneum, were found to be capable of mating with as many as seven different virgin females within 15 min. Across sequential copulations with virgin females, there was no decline in either male insemination success or average female progeny production over 48 h. However, when males copulated with previously mated females, there was a significant decline in male paternity success across sequential copulations, possibly due to male sperm depletion. In separate experiments, T. castaneum males were found to engage in two to six repeated copulations with the same, individually marked female. These repeated copulations did not increase male insemination success, short-term female fecundity, or male paternity success. Repeated copulations may possibly play a role in sperm defence. This study indicates that males may frequently engage in multiple matings, but these additional matings may lead to diminishing male reproductive returns.     

Socio-sexual behavior of female northern muriquis (Brachyteles hypoxanthus)

Female northern muriquis (Brachyteles hypoxanthus) are known to engage in frequent copulations with multiple partners, a pattern that in other primates has been attributed to various functions such as confusing paternity, reducing male aggression, or ensuring fertilization. However, in some female primates, promiscuity is restricted to times when conceptions are unlikely. We investigated whether female northern muriquis might exhibit a similarly mixed strategy by examining their mating, social, and activity patterns during their conception cycles versus other times. Systematic behavioral data were collected during an 18-month period between August 2001 and February 2003 on 13 adult females in a well-studied group at the RPPN-Feliciano Miguel Abdala, Minas Gerais, Brazil. Females mated on an average of 12.5+/-7.9 days during the study period, and spent significantly less time resting and engaging in non-sexual social behaviors, and significantly more time in sexual behaviors on days that they copulated than on days they did not. Three of the eight females for which sufficient data were available copulated significantly more often with their spatially closest non-kin associates, and four of five females that could be analyzed copulated significantly more often with their most frequent non-kin embrace partners. Comparisons between conception and non-conception periods revealed no differences in female activity budgets or in either the number of copulations or the number of different mating partners per female. Our results suggest that some females mate preferentially with close associates and social partners, but there is no indication that females alter their behavior during the cycles in which they conceive.     

Monogamy in a feeding generalist, <Emphasis Type="Italic">Chaetodon trichrous</Emphasis>, the endemic Tahitian Butterflyfish

Butterflyfishes have been well studied for their feeding ecology and mating systems. In particular, studies of corallivorous butterflyfishes have supported models of monogamy based on their predictable, low quality food; a patch of coral that is economically defensible by a pair. Moreover, pairs often exhibit trade-offs in territorial defense (greater by males) and feeding (greater by females) that improve their reproductive success. However, this model has not been well tested for more generalist feeders. In addition, recent hypotheses for monogamy in fish have emphasized parental care, but butterflyfishes do not provide parental care. This study tests five hypotheses for monogamy in the endemic Tahitian butterflyfish, Chaetodon trichrous: 1) uniform distribution of limiting resources, 2) joint defense of a territory, 3) low mate availability, 4) predator detection, and 5) benefits of cooperative behavior. Chaetodon trichrous was the most abundant butterflyfish in bays. Pairs jointly patrolled feeding territories. They preferentially fed over hard substrate other than live coral, however, this substrate was available outside of territories. They also ate plankton. Pairs were sorted positively for size, and all pairs were heterosexual. Males were larger than their partners, but females fed at higher rates. These results suggest that C. trichrous is monogamous, but reject the hypotheses that pairs form for joint defense of a territory (pairs swam together), that pairs remain together because of low mate availability (frequent interactions with neighbors), or that pairs form for predator detection (no homosexual pairs). Monogamy in C. trichrous is associated with the uniform distribution of hard substrate, although this resource is not limiting. Further, the higher feeding rate of females may represent a benefit provided by their monogamous mates.     

Prospecting in a solitary breeder: chick production elicits territorial intrusions in common loons

In many species, young animals learn about various breedingpatches in one year and use what they have learned to settlein a promising patch the next. Common loons (Gavia immer) seemgood candidates for such prospecting as prebreeders and displacedbreeders intrude frequently into breeding territories defendedby monogamous pairs yet engage in no extrapair copulations.We tested 3 hypotheses for prospecting in loons. The permanentattributes hypothesis gained little support as we found no consistentdifferences in quality between territories and no physical orbiotic trait that predicted reproductive success in a territory.We found some support for the conspecific attraction hypothesisas intruders were attracted to conspecifics in a lake in theshort term; however, intrusions were not more frequent in territoriesthat had experienced regular use by a pair the previous yearthan in territories that had previously been vacant. Instead,the increase in intrusion rate after a year of chick productionsupported the habitat-copying hypothesis, which states thatfloaters use the presence of chicks as a cue to target territoriesfor future attempts at territorial takeover. Despite this systemof prospecting, founding of new territories was common. Onestriking finding was the tendency of territorial breeders toconceal chicks from flying intruders, perhaps to avoid futureterritorial takeover.     

Extrapair paternity in the Hoopoe Upupa epops: an exploration of the influence of interactions between breeding pairs, non-pair males and strophe length

Previous studies of the Hoopoe Upupa epops have shown that the strophe length of male songs influences female mate choice, and is correlated with female reproductive rates and male production of fledglings in the male’s own brood. However, frequent interactions between breeding pairs and non‐pair males suggests that extrapair copulations could occur and could affect the real number of fledglings sired by males, and therefore the relationship between strophe length and breeding success. Here we analyse the incidence of interactions between breeding pairs and non‐pair males, and of extrapair paternity, the interrelation of these parameters, the influence of male strophe length on them, and whether extrapair fertilizations affect the correlation between strophe length and breeding success of males, in a colour‐ringed population of Hoopoes in south‐eastern Spain. Multilocus DNA‐fingerprinting revealed that 10% of the broods contained offspring sired by extrapair males, representing 7.7% of the chicks. However, the interactions between pairs and non‐pair males were more frequent, with more than 25% of broods being visited by non‐pair males, and about 10% being helped (fed or defended) by males other than the nest owner. Most of these relationships were apparently attempts by visitor males to obtain copulations with paired females, or to obtain access to such females or nests in future breeding attempts. However, there was no significant link between the detection of interactions with alien males in a nest and the occurrence of extrapair paternity in it, indeed extrapair paternity was found in only 30% of the nests with interactions, and therefore the detection of visits or helping by non‐pair males cannot be considered evidence of extrapair paternity in visited or helped broods. Males that sang with long strophes never suffered losses of paternity within their broods, while 25% of males that sang with short strophes did. However, these differences were not significant. Nevertheless, strophe length of males was significantly positively correlated with per brood and seasonal production of fledglings after accounting for losses of paternity within their own broods.     

Unusual sex roles in a highly promiscuous parrot: the Greater Vasa Parrot Caracopsis vasa

We describe the unusual mating system of the Greater Vasa Parrot Caracopsis vasa . The dull black plumage of males and females is similar but females are significantly larger than males. Females are promiscuous and copulated with at least five different males. Copulations were either short (1–3 s) or very long (mean 35.9 min), and long copulations involved a copulatory tie facilitated by the male's enlarged cloacal protrusion. Multilocus DNA fingerprinting of 17 broods showed that all were of mixed paternity, and that some broods had three fathers. Males never visited nests directly, but during the incubation and chick-rearing periods females came off the nest and were fed regurgitated fruit by multiple males. Four had band-sharing coefficients that suggested they were unrelated. Males copulated with and provided food for several widely separated females simultaneously. During the chick-rearing period females defended a territory around the nest from conspecific females, developed conspicuous orange skin on the head (through feather loss), and uttered loud, complex vocalizations that we refer to as 'song' from prominent perches near the nest. Males showed none of these traits. Females with high song rates attracted more males and as a result received more food than other females. Play-back experiments in which female song rates were either increased or decreased, attracted more or fewer males respectively. We propose that female song, conspicuous head colour and territoriality have all evolved as a result of competition between females for the food provided by males. The selective pressures favouring this highly unusual breeding system in the Greater Vasa Parrot are unclear but some sort of ecological constraint, such as food availability, may be important.