Parent birds commonly face the problem of distinguishing their own brood from foreign chicks. Learnt chick-recognition evolves only when parents do not mistakenly learn to reject their own young.
Many species of cuckoo steal parental care by laying their eggs in nests belonging to other birds, but at that stage their potential victims are often well equipped to defend themselves against exploitation1. Hosts of the common cuckoo (Cuculus canorus), for example, have such finely tuned powers of discrimination that they can often eject a cuckoo egg from their nest, even though it closely resembles one of their own (Fig. 1). Just two or three weeks later, however, host birds seem bizarrely incapable of identifying the monster chick that has taken over their nest (Fig. 2, overleaf), and they continue to care for the cuckoo until it becomes independent1. Research by Shizuka and Lyon (page 223 of this issue)2 helps to explain how such exquisitely adaptive host behaviour can be followed by actions that seem so absurdly maladaptive.
Previous work shows that host birds learn to recognize the appearance of their eggs the first time they breed, and thereafter reject any eggs that look odd by comparison1. So why don't hosts similarly learn to recognize their nestlings? The following hypothesis3 provides an answer. Imagine a naive parent unlucky enough to receive a cuckoo egg the first time she breeds. The cuckoo egg is one among several host eggs in the nest, so there is a high chance that the host will correctly learn the appearance of her own eggs. However, the host is unable to learn the appearance of her own nestlings correctly because the cuckoo chick evicts all the host young from the nest soon after it hatches. Instead, she mistakenly learns to recognize the foreign nestling as her own and then, in subsequent breeding attempts, rejects her own offspring because they look odd by comparison.
Now imagine a host that doesn't learn to recognize chicks. She might similarly fall victim to the cuckoo in her first breeding attempt, but at least has the prospect of subsequently rearing her own young. So hosts that don't learn to recognize their nestlings leave more descendants than those that do. This means that the only possible way for learnt chick-recognition to evolve is if there are safeguards against costly mistakes in the learning process3,4.
In the first detailed test of this hypothesis, Shizuka and Lyon2 show how hosts protect themselves against learning errors, and demonstrate that this has enabled the evolution of learnt nestling-recognition. The authors focus on the American coot, Fulica americana, a bird that cheats on its own kind by laying eggs in nests belonging to other American coots. In this species, the brood hatches over several days, and foreign chicks are typically among the offspring that hatch later. Coot chicks are dependent on parents in the days after hatching, and foreign chicks are raised alongside host offspring. Shizuka and Lyon suspected that hosts might be able to recognize and reject foreign coot chicks, because these offspring typically suffered higher rates of mortality than host young, even after controlling for their later position in the hatch sequence. They wondered whether hosts might learn the appearance of chicks that hatched first and then reject any later-hatching chicks that seemed foreign.
To test this idea, Shizuka and Lyon presented parents with one sort of chick at the start of the nestling period (the 'referents'), and allowed them to learn the characteristics of these chicks for a day. During the next four days, parents were given test chicks, half of which were from the same brood as the referents and half of which were not, and the chicks' survival was monitored to detect evidence of learned discrimination. The experiment showed that the test offspring that were unrelated to the referents were least likely to survive to independence. A subsequent experiment ruled out the possibility that parents were simply rejecting chicks of the minority type. So parents learn to recognize their offspring by imprinting on the chicks that hatch first and they are inclined to reject any later-hatching chicks that seem odd by comparison. By using their first-hatched chicks as referents, parents are unlikely to mistakenly learn the wrong sort of offspring as their own.
The experiment therefore supports the hypothesis3 that learnt chick-recognition can evolve only when the learning process is error-free. Although the traits involved in coot chick-recognition remain to be described, classic behavioural studies suggest that both acoustic and visual cues are probably involved. For example, colonially nesting gull and swallow species are not cuckoo hosts, but they risk feeding alien young when their offspring leave the nest and intermingle with other fledglings of the same species5. Experiments show that parents start to discriminate against foreign chicks just before their own nestlings become mobile, and they identify offspring by the structure of their calls and by unique plumage patterns on their head5. Although American coots are not colonial, their chicks can swim soon after hatching and they occasionally stray onto foreign territories, where they are attacked by adults6. Perhaps this means that learnt chick-recognition has evolved in American coots as a general defence against feeding any alien young, rather than as a specific defence against parasitic chicks.
Might cuckoo hosts learn to recognize their nestlings in the same way as American coots? Intriguingly, the eggs of virtually all parasitic cuckoo species hatch in advance of host young1, and it is tempting to speculate that their shorter incubation periods have been selected to prevent the evolution of learnt nestling-recognition in their hosts. Nevertheless, there is increasing evidence that some hosts can recognize and reject cuckoo chicks7,8,9. The common theme in these diverse studies2,7,8,9 is that host discrimination against parasitic chicks can evolve as long as there are mechanisms in place to minimize the costs of accidentally rejecting host young.
Discrimination need not involve learnt chick-recognition4,8: unlearned rules of thumb may enable hosts to reject foreign nestlings rather than their own. For example, hosts may avoid exploitation by abandoning offspring that are alone in the nest4,7, or that take an unusually long time to fledge8, because these characteristics are reliably associated with cuckoo nestlings. Where there is learnt nestling-recognition, learning that is confined to a sensitive period2, or guided by an innate template4, can reduce the chance of error. In this regard, learnt chick-recognition resembles song-learning in birds, which is often confined to periods when an appropriate tutor bird is present and which can be guided by an auditory template that prevents birds from learning the song of the wrong species10.
The sight of a tiny songbird feeding a monstrous cuckoo chick overflowing its nest (Fig. 2) is one of the most extraordinary in nature. The message emerging from the latest research is that, absurd as it may seem, such behaviour will persist as long as there are no measures to prevent hosts from routinely rejecting their own chicks.
Davies, N. B. Cuckoos, Cowbirds and Other Cheats (Poyser, 2000).
Shizuka, D. & Lyon, B. E. Nature 463, 223–226 (2010).
Lotem, A. Nature 362, 743–745 (1993).
Langmore, N. E. et al. Behav. Ecol. 20, 978–984 (2009).
Beecher, M. D. Behav. Genet. 18, 465–482 (1988).
Lyon, B. E. Anim. Behav. 46, 911–928 (1993).
Langmore, N. E., Hunt, S. & Kilner, R. M. Nature 422, 157–160 (2003).
Grim, T. Proc. R. Soc. Lond. B 274, 373–381 (2007).
Sato, N. J., Tokue, K., Noske, R. A, Mikami, O. K. & Ueda, K. Biol. Lett. doi:10.1098/rsbl.2009.0540 (2009).
Catchpole, C. K. & Slater, P. J. B. Bird Song: Biological Themes and Variations (Cambridge Univ. Press, 2008).
About this article
How to spot the tricks of the cheat: detecting eggs and chicks of brood parasites as a consequence of coevolutionary arms races
Japanese Journal of Ornithology (2012)
Biological Reviews (2011)