Brood Parasitism and the Evolution of Cooperative Breeding in Birds

See allHide authors and affiliations

Science  20 Dec 2013:
Vol. 342, Issue 6165, pp. 1506-1508
DOI: 10.1126/science.1240039

Power in Numbers

Avian brood parasites target particular bird species to raise their offspring, sometimes at great cost to the foster family. Feeney et al. (p. 1506; see the Perspective by Spottiswoode) analyzed the global distribution of brood parasitism and found a correlation with the occurrence of cooperative breeders across multiple taxa. For example, Australian fairy wrens breed both singly and in cooperative groups, but the group breeders are better able to resist parasites than lone pairs, indicating that the prevalence of cooperative breeding may be a response to brood parasites.


The global distribution of cooperatively breeding birds is highly uneven, with hotspots inAustralasia and sub-Saharan Africa. The ecological drivers of this distribution remain enigmatic yet could yield insights into the evolution and persistence of cooperative breeding. We report that the global distributions of avian obligate brood parasites and cooperatively breeding passerines are tightly correlated and that the uneven phylogenetic distribution of cooperative breeding is associated with the uneven targeting of hosts by brood parasites. With a long-term field study, we show that brood parasites can acquire superior care for their young by targeting cooperative breeders. Conversely, host defenses against brood parasites are strengthened by helpers at the nest. Reciprocally selected interactions between brood parasites and cooperative breeders may therefore explain the close association between these two breeding systems.

Cooperative breeding, in which three or more individuals contribute to the care of young in the nest, occurs in around 9% of birds (1). The distribution of this social system is strongly skewed toward two major hotspots: Australasia and sub-Saharan Africa (2) (Fig. 1A). Ecological correlates of this distribution include both variable, unpredictable environmental conditions (2) and stable, predictable conditions (3). Unsurprisingly, the broad-scale ecological conditions that favor the evolution and persistence of cooperative breeding in birds therefore remain controversial (2, 4, 5).

Fig. 1 Global patterns of richness in (A) avian cooperatively breeding passerine species and (B) obligate avian brood parasite species, during their breeding seasons.

Previous studies have proposed that cooperatively breeding species are more likely to be hosts of avian interspecific brood parasites than are noncooperative species (6, 7). We investigated the correlation between avian brood parasitism and cooperative breeding. Interspecific brood parasites lay their eggs in the nests of other birds, primarily passerines, and abandon their young to the care of the host (8). The cost of hosting a brood parasite can be immense, so hosts are typically under selection to evolve defenses against parasitism (8). One of the most ubiquitous host defenses is the mobbing of brood parasites (9). Collective mobbing by multiple individuals can provide a more effective defense than solitary or pair mobbing, thus providing a selective force for cooperative or colonial breeding (10).

To test this hypothesis, we first compared the global geographic breeding distribution of avian brood parasites and cooperatively breeding passerine species (11). We found a strong correlation between species richness in cooperative breeders and species richness in brood parasites [simultaneous autoregressive model (z) = 61.3, P < 0.0001, correlation coefficient (r2) = 0.68, Fig. 1], with both exhibiting the same geographic skew toward sub-Saharan Africa and Australasia [63% of avian brood parasite species breed exclusively within this region (8)]. This correlation remains strong after controlling for avian species richness (z = 21.0, P < 0.0001, r2 = 0.41, fig. S1).

This correlation could reflect a direct association between brood parasitism and cooperative breeding, or both breeding systems could be the outcome of a third variable, such as the high cost of parental care in variable environments (2). If there is a direct association, either because exploitation by brood parasites promotes cooperative breeding or because brood parasites favor cooperatively breeding hosts, we would predict that within a given geographic region, species that are hosts of brood parasites should be more likely to breed cooperatively than nonhosts. We tested this prediction using phylogenetic comparative methods for two regions with sufficiently well-studied avifaunas: Australia and southern Africa (11). These two regions encompass the phylogenetically diverse passerine and nonpasserine hosts of 21 cuckoo species, 6 honeyguide species, and 9 parasitic finch species. We used published classifications of the host status of Australian passerines [brood parasites exploit passerines exclusively in Australia (12)] and all southern African birds (13) and the modes of parental care in all bird species worldwide (1). Our analyses revealed a significant association between hosts of brood parasites and cooperative breeders in both southern Africa (Bayes factor = 18.36, strongly correlated; likelihood ratio test: χ2 = 60.28, P < 0.001; Fig. 2A) and Australia (Bayes factor = 17.34, strongly correlated; likelihood ratio test: χ2 = 11.66, P = 0.02; Fig. 2B). In southern Africa, 27.5% of hosts were cooperative breeders, compared to only 7.6% of nonhosts. Similarly, in Australia 52.8% of hosts were cooperative breeders, compared to 11.9% of nonhosts.

Fig. 2 Random phylogenetic trees for (A) 892 bird species in southern Africa (gray divisions represent orders) and (B) 129 passerine species in Australia (gray divisions represent families).

Orange circles indicate cuckoo hosts, and blue circles indicate cooperative breeders.

Three non–mutually exclusive processes could explain why brood parasite hosts are more likely to be cooperative breeders: (i) Brood parasites might selectively target cooperative breeders to maximize the care of their offspring (7); (ii) cooperative breeders may be more obvious targets as a result of the increased activity of helpers near the nest (6, 7); and (iii) cooperative breeders may be better able to defend their nests against brood parasitism (7), selecting for cooperative breeding in hosts. To investigate whether one or more of these processes underpin the patterns uncovered by our comparative analysis, we conducted field observations and experiments on the facultatively cooperative superb fairy-wren Malurus cyaneus. In this species, some pairs breed unassisted, whereas others are assisted by up to six nonbreeding helpers. This allowed us to investigate how cooperative breeding might change the outcome of interactions with brood parasites. In southeastern Australia, superb fairy-wrens are the primary host of Horsfield’s bronze cuckoo, Chalcites basalis (12), and can suffer high annual rates of brood parasitism (14).

We began by investigating whether cuckoos might gain a selective advantage by preferentially targeting cooperative breeders for parasitism, using superb fairy-wren breeding and parasitism data (11). Cuckoo chicks grew slightly faster when reared by groups of three or more (n = 30 cuckoo chicks, day of the nestling period × group size; F1 = 7.46, P = 0.009), with a predicted mean (±SE) mass on day 12 of 22.6 g (±0.5 g) if reared by a pair and 23.4 g (±0.5 g) if reared by a group. The chance of surviving to fledge was also greater for nestlings reared by larger groups, because predation rates decreased with increasing group size [generalized linear mixed model (GLMM): χ21 = 4.31, P = 0.04]. Although superb fairy-wrens commonly reject cuckoo chicks (14), the incidence of chick rejection was not correlated with group size (n = 72 cuckoo chicks, logistic regression, χ21 = 0.6, P = 0.44). Overall, then, our analyses provide support for hypothesis (i). We find that brood parasites can gain a fitness advantage for their offspring by associating with cooperative breeders, because they offer superior provisioning and a more effective defense against predators.

However, our analyses also show that this potential fitness advantage was seldom realized by Horsfield’s bronze cuckoos parasitizing superb fairy-wrens, even when considering data from two sites over 500 km apart. Large groups were significantly less likely to be parasitized than small groups at both Campbell Park (GLMM: χ21 = 7.68, P = 0.006) and Serendip Sanctuary (χ21 = 5.01, P = 0.027; Fig. 3). Therefore, our results do not support hypothesis (ii): Cuckoos were not drawn to parasitize cooperative breeders because they are more salient targets for exploitation. Instead, we find support for hypothesis (iii), because cooperative breeding facilitates defense against brood parasites. We quantified the fitness advantage associated with better defenses against parasitism in large groups using data from Campbell Park (11). Relative to small groups, the reduced probability of parasitism in large groups increases the production of young by 0.2 fledglings per group per season. Therefore, both parents and related helpers gain a fitness advantage from cooperative breeding when interacting with brood parasites.

Fig. 3 The percentage of large and small superb fairy-wren groups that were parasitized by Horsfield’s bronze cuckoos at Campbell Park and Serendip Sanctuary, Australia.

Subsequent experimental analyses of behavior at the nest revealed how larger groups are able to escape parasitism more frequently than smaller groups. We found that superb fairy-wrens were more aggressive toward mounts of a cuckoo than of a nest predator (eastern brown snake, Pseudonaja textilis), a predator of adult birds (collared and Eurasian sparrowhawk, Accipiter cirrocephalus and A. nisus, respectively), a predator of both adults and nestlings (pied currawong, Strepera graculina), or a nonthreatening control (white-plumed honeyeater, Lichenostomus penicillatus; linear mixed effects model on number of mobbing calls: χ24 = 53.95, P < 0.0001; Fig. 4A). Further, cuckoo-targeted mobbing was elicited by a referential vocalization. Superb fairy-wrens produce whining alarm calls (15) that are structurally unlike any other calls in their repertoire (Wilk’s λ = 0.11, exact F8,56 = 13.75, P < 0.01; fig. S3) and do so exclusively when confronting a cuckoo (Friedman test: χ24 = 54.72, P < 0.0001; Fig. 4C). With playback experiments, we found that this call elicits a more rapid approach by group members than mobbing alarm calls or a control sound (a parrot call, GLMM: χ22 = 68.05, P < 0.0001; Fig. 4D). Once mobilized, the strength of these defenses increases with group size. Large groups were more vigilant around their nest (GLMM: χ21 = 8.03, P < 0.004), spent more time mobbing the cuckoo than smaller groups (Kruskal Wallis test: χ21 = 5.42, P = 0.02; Fig. 4B), and ultimately were less likely to be parasitized. Thus, superb fairy-wrens possess cuckoo-specific nest defenses, which are enhanced by helper contributions and which can explain the lower parasitism rates experienced by large groups.

Fig. 4 (A) Mean number of mobbing alarm calls produced by 15 fairy-wren groups in response to different model types: cuckoo (CK), currawong (CR), honeyeater (HE), snake (SN), and sparrowhawk (SP).

(B) Mean time spent mobbing the cuckoo model (<0.5 m from the model) by small (n = 27 groups) versus large (n = 5 groups) groups. (C) Mean number of whining calls produced in response to the five model types. (D) Mean number of individuals that approached playbacks (n = 20 each) of fairy-wren mobbing alarm calls, a control sound, and fairy-wren whining alarm calls. Error bars denote the standard error.

Our findings show a pronounced association between avian brood parasitism and cooperative breeding in birds, on a global scale. Our field data suggest that a two-way process underpins this relationship. On the one hand, brood parasites can gain a fitness advantage by preferentially exploiting the superior care provided by cooperatively breeding groups. On the other hand, the genetic relatives of offspring raised by cooperatively breeding families potentially gain fitness from the superior defenses that the extended family collectively mounts against brood parasites. Defense against brood parasitism is therefore an important kin-selected fitness advantage associated with cooperative breeding [see also (16)]. In superb fairy-wrens, we have shown this two-way process at work, but in other cooperatively breeding hosts, especially those with a less protracted coevolutionary relationship with brood parasites, only the first part of the process may be evident. The challenge remaining for future work is to determine the extent to which brood parasites have influenced the biology of cooperatively breeding species.

Supplementary Materials

Materials and Methods

Figs. S1 to S3

Table S1

References (17–24)

References and Notes

  1. Materials and methods are available as supplementary materials on Science online.
  2. Acknowledgments: We thank our field assistants, L. Joseph for assistance with specimens, and A. Cockburn for helpful discussions. W.E.F. was supported by the Canberra Birds Conservation Fund and Australian Geographic. R.A.M. and M.L.H. were supported by Australian Research Council (ARC) Discovery Grant DP110103120. N.E.L. was supported by ARC Discovery Grant DP110101966. Data are archived on Figshare. N.E.L. and W.E.F. conceived the study; macroecological analyses were done by M.S. and J.A.S.; comparative analyses were done by I.M.; field data were collected by N.E.L., R.M.K., R.H., M.L.H., and R.A.M.; model and playback experiments were done by W.E.F.; and N.E.L. and R.M.K. wrote the paper.
View Abstract

Navigate This Article