Does diversity deliver? How variation in individual knowledge and behavioural traits impact on the performance of animal groups

Have you ever been stood with a group of friends (or worse, colleagues) in an unfamiliar city trying to choose which restaurant to dine at? I expect so, but do you remember how you made your choice, all the while trying to stick together and not lose one another? There are a number of ways you could have arrived at your decision. You may have followed the decision of the group member who set-off purposefully down the High Street, or you all may have agreed to go to the restaurant declared 'the best in town' by the person with local knowledge of the city. Such situations are just as common in the animal world. Swap 'restaurant' to 'foraging patch' and we have a description for the type of foraging decisions faced by almost all social animals, every day of their lives. But these are not trivial decisions; a number of consecutive bad choices-where a group are led to poor foraging areas, or risky habitats-can be fatal. Such decisions are made even more difficult where animals face unexpected dangers or an environment which is constantly changing, something all animals are increasingly encountering in our rapidly changing world. One way animal groups may cope with uncertainty-and perform well in such environments-is by drawing upon one another's particular expertise or information, or having individuals bold enough to try out new options. For instance, we know that summing information possessed by individual group-members can increase the collective cognition of groups, a concept which has attracted much attention in the media and been termed 'the wisdom of crowds'. Larger and/or more diverse groups may also contain individuals with different skills and experiences, which increase the chances of a group solving a given task. But to date, studies that explicitly investigate the effect of such diversity on group performance have so far been restricted to human teams acting for financial gain, and have produced mixed results. In the proposed fellowship, I will use a combination of experimental and theoretical data to test the hypothesis that 'diversity delivers' in a group setting, and ask: What makes a winning team? To answer this question, I will present highly gregarious freshwater fish, the nine-spined stickleback, with a series of problems not unlike the 'which restaurant?' example I used in my opening sentences. Combining highly sophisticated video tracking of individual fishes movements, with foraging devices made using kitchen egg-timers and laboratory petri dishes, I will explore the role of variability in individual knowledge and personality. Questions I will tackle will include: If the most informed individual in a group is also the boldest in the group, does the group make more accurate decisions as a consequence of following this bold individual's choices? If a group is composed of a majority of poorly informed bold individuals and a minority of well-informed shy individuals, is the group able to utilise the information possessed by the latter? How different does a groups information or personality composition have to be before they cannot agree on a place to forage, and split up? The unique experimental set-up proposed will also allow me to measure the actual benefit to individual fish (in terms of amount of food eaten) in these different contexts. All of these experiments will take place with groups of just five fish, given the time needed to individually assess individual personality and train fish to different levels of 'expertise'. However, I will then use a series of computer simulations based on what I find in these experiments to make predictions about the outcome where shoals of fifty or a hundred individuals are concerned, which I will then test in the laboratory. Overall, this project will be the first attempt to manipulate the distribution of both information and personality within groups, and will provide a new insight into the function of diversity for social animals when coping with uncertainty.