What is maintaining variation in personalities in wild populations of great tits and blue tits?

Abstract

It is now commonly recognized that animals have different personalities and that these personalities are moderately heritable, repeatable, and may influence fitness components such as survival and reproduction. Much of this knowledge has been gained by studying birds, and in particular great tits Parus major. In nature, great variation in personalities seems to exist, from shy to bold individuals, i.e. a shy-bold continuum. It has been suggested that the various personalities have equal fitness over time because of environmental variation, and thus represent Evolutionary Stable Strategies (ESS). An alternative is that some individuals with low fighting ability, e.g. because of poor rearing conditions, play a shy personality ( Best of a bad job strategy). I wanted to address this problem by studying anti-predator behaviour in great tits and blue tits Cyanistes caeruleus in a woodland area in Norway. A total of 107 great tit and 75 blue tit local recruits (grown up in the area) were caught in the autumn and several measures of anti-predator behaviour were recorded. Subsequent winter survival and reproductive success (i.e. laying date, clutch size, number of fledged and body mass on day 15) were observed. As the methods adopted in this study to measure personality traits were previously little used, analyses of repeatability was performed in order to determine whether the methods proved to be a reliable way of measuring personality. The results showed quite high repeatability within observers, but low repeatability between observers and within and between anti-predator variables in general. Possible explanations for this are discussed. Two significant relationships, both for great tits, were obtained between fitness measures and anti-predator behaviour (survivors screamed less than non-survivors, and birds with low breathing rate laid more eggs than birds with high breathing rate). However, due to the large number of tests, and data from one year only, no firm conclusions could be drawn. Improved methods and data from more years are needed.