Wild birds learn to avoid distasteful prey by watching others
How do predators know to avoid brightly-coloured toxic prey? A collaboration of researchers has put social information theory to the test in a reliable real-world system to find the answer – by copying what others do, or do not, eat.

An international team of researchers from Finland, New Zealand, Colombia and the U.K. have provided the first evidence that wild birds can learn to avoid distasteful prey by observing what others eat.

“We’ve known for a long time that predators, like birds, associate brightly coloured warning signals with the danger of eating certain prey types. However, we’ve never been able to demonstrate in the wild how predators learn about these aposematic prey advertisements. If predators do not recognize the signal, then the prey are highly vulnerable to naïve predators. This is a big problem that prey face each year when juvenile predators arrive. Since aposematism is widespread in nature, we wanted to solve this problem in a real-world setting” explains one of the lead-authors Rose Thorogood, now at the University of Helsinki.

At the established Madingley Wood field site in Cambridgeshire, UK, Liisa Hämäläinen, a doctoral student at the University of Cambridge, used an innovative combination of field experiments and social network analyses to investigate the capacity for social information transmission among bird predators and identified potential implications for predator-prey coevolution. The team’s results offer solutions to evolutionary problems in predator-prey dynamics and support theoretical predictions from social learning theory. But more broadly, the research speaks to the role of social information amidst the powerful dynamics of ecology and evolution, and hence opens the door to new studies in other coevolutionary systems.

According to this experiment, one avenue is through social learning. The researchers set up pairs of bird-feeders within the Madingley Wood field site. One feeder would dispense brightly-dyed almond flakes that were left naturally tasty (undefended prey). The other would dispense differently-coloured almond flakes with additives to make them disgustingly bitter (aposematic prey). Local blue tits (Cyanistes caeruleus) and great tits (Parus major) could gather around the feeders and take their pick of food, but also observe the feeding attempts of others.

In between the experimental sessions, the feeders dispensed plain, uncoloured almond flakes. During this time, the researchers used RFID to record predator visits to the feeders. Will Hoppitt, a statistical consultant from Royal Holloway University of London, then generated a social network based on the likelihood for individuals to forage together, and compared this to the patterns observed in the birds’ choices of coloured almonds.

The analysis showed that birds could learn to avoid the bitter almonds within eight days, with adults being quicker to learn than the juveniles. Importantly, information about the bitter almonds appeared to flow through the predicted social links, especially from adults to juveniles. This offers a solution to the problem of naïve predators, as they can learn by watching the behaviour of others instead of through trial-and-error. Thereby, the selection pressures exerted on their aposematic prey are reduced.

”These results greatly extend the current state of our knowledge gained from studying predator learning under controlled lab conditions. It demonstrates that social interactions both within and across species allow predators to learn very quickly, and in turn allow aposematic prey types to persist across naïve predator generations. This highlights that social information transmission is likely to play a critical role in eco-evo dynamics and in many other coevolutionary systems, including host-parasite and plant-pollinator relationships. It’s exciting to see what we can find from further field-based as we dive deeper into understanding the social layers of coevolutionary processes” Thorogood sums up.

The international collaboration was spearheaded by Liisa Hämäläinen and Rose Thorogood at the University of Cambridge (now at Macquarie University and University of Helsinki, respectively), has recently published a pioneering insight into the networked social layer of eco-evo dynamics using a long-standing field system. The collaboration brings together expertise in social networks, coevolutionary dynamics, and fieldwork from across Europe.

Original article:

Hämäläinen, L., Hoppitt, W., Rowland, H.M., Mappes, J., Fulford, A.J., Sosa, S., Thorogood, R. Social transmission in the wild can reduce predation pressure on novel prey signals. Nat Commun 12, 3978 (2021). https://doi.org/10.1038/s41467-021-24154-0 (open access)

Contact details for lead authors:

Focus on the social layer

Original text written by Stephen Heap (drstevilphd.com)

The coevolutionary trajectories of interacting species twist and turn according to the interplay of ecology and genetics (ecology and evolution: eco-evo). The context of ecology shapes the selection pressures on different genes, whilst changing gene frequencies alter the prevailing ecological conditions. But much of our understanding of eco-evo dynamics focuses on the relationships between interacting species, neglecting the social interactions that occur within species. In other words, the ecological component of eco-evo dynamics often includes a critical social layer, yet this layer is seldom investigated outside of simulations or strict lab environments.

Meanwhile, state-of-the-art research in social ecology is uncovering the processes of information transmission across social networks. The capacity for social networks to store and transmit information is important because it allows individuals to inherit behaviours from others (social learning), which builds upon the inheritance that comes through genes. The literature highlights the possibilities for social learning to impact selection pressures, and thereby be involved with eco-evo dynamics. But again, this theory lacks evidence from natural systems.