Using real predators and robot prey to investigate the importance of predators in prey responses

Predator and prey are locked in a dynamic interplay of evolved behaviour; predators attempt to maximise their net intake while unsuccessful prey do not live to see another day. This often culminates in dramatic behavioural displays, as cheetahs chase gazelles across the savannah, gannets dive bomb sardine shoals, and bats hunt moths through echolocation. However, while recording and analysing this is possible, studying this type of behaviour is problematic as ideally we would manipulate the response and movement of prey, which is technically very difficult. Without manipulation, it is difficult to say with any certainty that such behaviour has evolved to help animals avoid predation. This has limited our understanding of dynamic responses to predation as the behaviour of predators is often based on unsupported assumptions rather than empirical data. By using recent technological developments, this project will explore how short-term behavioural avoidance by prey has evolved to minimise risk, how predators learn to counteract these strategies, and whether this locks predator and prey into a cycle of adaptation and counter-adaptation.

Predatory fish (the European perch, Perca fluviatilis) will be allowed to encounter, detect, attack and try to consume (i.e. 'kill') artificial robotic prey, small items of food magnetically connected to robots under the testing tank. While the predator's behaviour is unconstrained, the behaviour of the prey are under full control by a computer program. To allow responsive behaviour by the prey, the predatory fish will be tracked by a computer in real time and this information will be fed back and made available to the prey. The complexity of the prey's reaction will be manipulated, from relatively simple behaviours such as fleeing from the predator or aggregating toward other prey, to the more complex such as balancing aggregation with fleeing. In the latter stages of the project, the longer term dynamics of predator-prey interactions will be examined, where the frequency of different prey strategies will change over time based on the previous success of each strategy. This will simulate an evolutionary process and how the predators adapt their subsequent behaviour to these changes in prey strategies will allow an examination of the role of learning in these interactions.

The main focus of interest will be on the responses of prey groups. These collectives (flocks, schools, herds, swarms) continue to fascinate scientists, filmmakers and the public alike. One shared reason is the almost incomprehensible degree of coordination between individuals in some of these species; the evening displays of European starlings (Sturnus vulgaris), for example, attract visitors to sites across the U.K. and is also the subject of a major European Commission research project. There is currently much research interest in these 'self-organised' systems across disciplines, including physics, computer science and engineering as well as biology, and there are applications in controlling pedestrian crowds and pest swarms such as locusts. The proposed method will allow us to examine with unprecedented detail how this group behaviour protects prey from predators. For example, it is often thought that polarisation (where individuals in groups line up to face the same direction) facilitates information transmission between individuals, as a response to detecting a predator or food source. By manipulating this behaviour, it can be tested whether this actually is effective at avoiding real predatory attacks.

In addition to methodological advantages, importantly the system also circumvents the ethical problem posed by exposing live prey to predators which has previously hampered progress in this area. The project will develop a flexible system for studying behavioural interactions between individuals of the same or different species and will be of interest to a range of researchers working on other questions.

The system I seek to develop meets an important national priority towards reduction, refinement and replacement of animals in research. Not only will I be able to examine predator behaviour without killing or injuring live prey animals, but using artificial prey allows a much greater degree of standardisation of prey appearance and behaviour so that the variability in the data is reduced, and hence fewer predator individuals are needed. The new system will also record multiple attacks per predator, again meaning fewer individuals are needed. Although studies of predator-prey behaviour may not represent a large proportion of research using animals, any development in allowing a research question to be answered in the U.K., rather than the work being conducted overseas under more lax legislation, is progress.

The project's primary aim is to achieve high-quality basic research, thus applications to economic or social advances, beyond the effect on replacing and reducing the use of vertebrate animals in research, are unlikely to be a direct result of the project. However, predator-prey interactions have been used in military applications previously, and as the project seeks to understand attack-and-avoid behaviour within natural systems this could provide bio-inspired solutions to problems such as missile targeting or avoidance. With modelling of the empirical findings by Dr. Iain Couzin, Dr. Colin Tosh and others to generalise the underlying mechanisms, these mechanisms could be applied to engineering applications and potentially contribute to the UK's defence industry. More broadly, the project adds a unique perspective to the study of self-organisation and collective behaviour, a field which has clear applications in understanding, and where necessary controlling, human behaviour at high densities (such as during crowd control, evacuation procedures, and pedestrian traffic).

The choice of predator species may have implications for non-academic impact which could be explored in detail in future grant applications. The European perch has been introduced far beyond its natural range as a common game fish and in Australia and New Zealand has had a major negative impact on native freshwater communities. Part of the perch's success is likely to be due to its ability in learning to successfully predate native prey that was previously unfamiliar. The latter two years of the project will examine predator learning in detail, and these findings may have implications for the importance of learning by invasive predator species.

Both predator-prey interactions and collective animal behaviour have great popular appeal and our findings will attract attention from both the scientific and more general media. As an example, the previous study that this proposal seeks to develop was covered by over 15 online articles and a YouTube video explaining the video in lay terms has been viewed over 14,000 times (see Pathways to Impact). I will also undertake activities at public engagement events. More focused impact will be realised by visiting local schools to explain to GCSE and A level pupils the motivation, method and approach for the project and, more generally, how scientific research works. I will particularly emphasise the importance of the different skills required by the project from across disciplines, and encourage interested pupils to apply to university to study STEM subjects.

The research will provide valuable opportunities for undergraduate and postgraduate students. Taught students will be encouraged to undertake self-contained projects in parallel to the research and I will provide support, training and guidance for these students. Importantly the inter-disciplinary nature of the project will help broaden the usually narrow view that most undergraduates have of biology before entering the workplace. Experience with the proposed project will give students a more diverse understanding of approaches to problem solving.