1. Avian cognition and the evolution of multimodal warning displays

Toxic insects use bright and colourful patterns to signal their unpalatability to visually hunting predators. Birds are able to learn to avoid warning colours, especially when they are novel or are conspicuous colours that contrast with their backgrounds. However, aposematic insects can also produce odours and sounds if they are inspected or attacked by potential predators. For example, this peacock butterfly also hisses when it is attacked: but why hiss as well as have warning colours?

The answer is to be found in understanding predator cognition. Foraging birds learn to avoid the colour patterns which seem to be designed to be particularly easy to learn. This project studies how acoustic components of insect warning signals enhance predators' abilities to make associations between colours and unprofitability. I use touchscreen Skinner boxes specifically designed to present visual images and sounds to birds. I look at how sounds improve the way that birds learn about and remember visual patterns. Whilst I study fundamental cognitive processes in birds, I use the data I collect to test my ideas that they can drive the evolution of multisensory signals (see Publications).

2. Avian cognition and the evolution of insect defence chemicals

This is a classic picture from research carried out in the sixties, where blue jays were given monarch butterflies to see how well they learned to avoid them... Monarchs contain emetic chemicals that made the birds ill after they had eaten them - not surprisingly, the birds were very good at remembering their warning patterns and subsequently avoided them.

John Skelhorn is researching the role of defence chemicals in aversion learning for his PhD. His work shows that the way that birds perceive and learn about unpalatable chemicals could explain not only the evolution of colour patterns, but also the diversity of insect defence chemistry. He has done this using non-toxic chemicals so as not to make the birds ill, but has found general learning principles that have not previously been considered in warning signal evolution.

If you would like to know more about John's research, visit his webpage.

3. Interactions between unlearned colour biases and aversion learning

This ladybird is reflex bleeding in response to being 'attacked' by a human predator. This is another example of a multimodal warning signal (see above), since the liquid that the ladybirds is secreting smells very strongly of a chemical, called pyrazine. This odour is produced by many warningly coloured insects. My PhD work found that this odour could produce colour biases in naive young birds, in that they avoided food that was coloured red or yellow (both these colours are commonly used by insects in their warning patterns). I have since found that other novel odours, sounds and even tastes can elicit the same colour biases (see Publications). This suggests some general mechanism in predators that adaptively biases their foraging decisions away from prey that could be toxic.

I am now interested in how these unlearned visual biases affect the learning process. In particular, I am currently looking at how they might constrain the learning process, perhaps making some colour-toxin associations easier to learn than others.

4. Survival, risk-taking and hunger: effects of predator state on prey choices and evolution of colour patterns

Times in a bird's life can be hard, for example when it is cold, or when prey are hard to find. How hungry a bird is is likely to affect the foraging decisions it makes: a potentially toxic insect may be very acceptable if a bird is hungry and it is its only chance of survival. Recent theoretical models of the evolution of warning signals and mimicry have incoportaed internal state to look at how hunger and risk might affect the evolution of visual signals.

In collaboration with Dr Melissa Bateson and her PhD student, Craig Barnett, we are looking at how hunger affects birds' decision-making, which could be important in the evolution of visual mimicry.

To learn more about this project and the role of state in foraging choices, please see Craig's website.