Camouflage & Predator Learning

Being spotted by a predator can spell the end for many animals, so camouflage is a vital first line of defence.


No camouflaged object can match its background perfectly, allowing predators to home in on weak-spots in the protection. Learning to “break” camouflage in this way is most likely to occur when predators encounter the same type of prey repeatedly, possibly allowing them to build up a mental template or “search image” of the prey’s appearance, or perhaps by focussing on salient textural or colour cues. However, the importance of predator learning has been largely ignored in previous camouflage research. We have a major on-going project, funded by a BBSRC grant to Martin Stevens and John Skelhorn (Newcastle), with Jolyon Troscianko, to further our understanding of how camouflage has evolved to evade a world full of sophisticated visually-guided predators and to defeat both their visual and cognitive systems.

A number of different camouflage strategies have been proposed and investigated, including background matching (resembling the general appearance of the background), disruptive coloration (breaking up an object’s outline), and distractive markings (diverting the predator’s attention away from object edges).

Here we are testing whether some types of camouflage are easier to learn to find over repeated encounters than others. We will also test how easily predators can switch from one camouflage type to another after forming a search image, whether certain camouflage types are more difficult to form search images for, and whether specific camouflage features make them more or less salient for learning.

Much of the work will focus on humans as model predators, using touch-screen computer experiments and artificially generated prey. We use online games (play here!) in addition to controlled experiments on calibrated touch screens to generate these data. However, humans are just one group of visually guided animal – other species may differ in how they respond and learn about camouflage. We are therefore also running parallel experiments with domestic chicks, training them to peck at screens and rewarding them with meal worms. The individual chicks have no prior experience of finding camouflaged prey, but have a sophisticated visual system ready for search tasks within hours of hatching, and are highly motivated by meal-worms.

Watch The Video

You can watch a video below of how we train the birds to attack computer-generated stimuli.