The study of personality is typically the domain of psychologists. Recently, however, animal researchers have been keen to delve into the personality traits of non-human animals. “During the past couple of years, animal personality has become a very active area of research,” says Max Wolf, a PhD student at the University of Groningen in the Netherlands. “But many questions about the evolutionary value of variation in animal personality have still not been answered.” Such questions, as it turns out, became the focus of his graduate thesis.

Researchers had observed variations in certain behaviours among animals — for example, some are more aggressive than others. Furthermore, such differences tend to be consistent — an individual who is more aggressive today is likely to be more aggressive in the future. And an individual that tends to be more aggressive is also inclined to be bolder in exploring its territory. Such observations suggested that something akin to human personalities exists in other animals. But why would evolution give rise to different personality types?

Wolf's mentor, theoretical biologist Franz Weissing, and his colleagues noticed that many personality traits described in the literature could be categorized as 'risk-taking' behaviours. They reasoned that the predilection for or against taking risks — aggressive and bold individuals take more risks than non-aggressive and shy individuals — might be explained by how much an individual stands to lose in terms of future reproduction. For example, an individual that has a good chance of mating in the near future would have a lot to lose, and would therefore evolve to be consistently risk-averse in different situations.

Starting with this basic intuition, Wolf began working towards a mathematical model to study the idea. He first scoured the literature and discussed ideas with his colleagues, then used the resulting information to develop the model. “It took months to arrive at a satisfying model structure,” he says. The trick was to simplify the problem enough so that it became tractable — for example, by assuming that an individual lives for only a couple of years and reproduces only once a year — but not so much that it lost its most crucial ingredients. The model had to provide repeated choices for each behavioural situation, such as whether or not to run away from a predator, throughout individuals' lives.

In the model (on page 581), the degree to which an individual has invested in future reproduction determines how that individual behaves in risky situations. This model can now be tested experimentally by looking for a correlation between the reproductive profile of an animal and its propensity to take risks. “Ultimately, what we want to achieve with our theoretical work is to influence empirical work,” says Wolf.

With almost two years left of his PhD, Wolf is currently applying his ideas to more specific ecological situations. For example, in a Dutch population of oystercatchers, some birds start breeding early in life on low-quality territory, whereas others delay reproduction by queuing for high-quality territories. Wolf's model predicts that the birds that queue for the high-quality territories have more to lose than those that have already reproduced. Thus, the queuing birds should be relatively shy and non-aggressive. “Applying our ideas to more specific ecological scenarios will make it easier for researchers to test these ideas,” says Wolf.