Personality, anti-predation behaviour and behavioural plasticity in the chaffinch Fringilla coelebs

Animal 'personalities' or behavioural syndromes can be defined as individual differences in suites of correlated behaviours. Using chaffinches Fringilla coelebs temporarily held in captivity, we determined the extent to which activity level (AL), the behavioural response to a stressful situation (BRSS) and two anti-predation risk behaviours when exposed to a model predator were inter-correlated within individuals. We assayed AL and BRSS when the chaffinches were first put in an experimental room. Subsequently we recorded their initial response (freeze or escape) to the sudden appearance of a model hawk, in addition to their latency to resume activity afterwards as a measure of cautiousness ('latency'). Each bird was assayed in two situations that differed in the risk level posed and in the assumed optimal response required to minimize predation risk: (I) high risk, when the hawk flew directly over the chaffinch and the assumed optimal response was to escape; and (2) low risk, when the model hawk flew 2 m to the side of the chaffinch and, since the bird was not being targeted, the assumed optimal response was to freeze in order to avoid attracting attention. All of the variables had relatively high repeatabilities and were inter-correlated, providing strong evidence for a behavioural syndrome. Though propensity to freeze was much greater in the low risk treatment, AL was negatively correlated with freezing behaviour in both risk treatments, and therefore there was a 'behavioural carryover' across situations. While hypoactive individuals were more likely to freeze in the low risk treatment, and were therefore assumed to be better able to assess risk compared to hyperactive individuals, some hypoactive individuals also froze in the high risk situation when it was assumed to be inappropriate to do so. Despite this apparently maladaptive behaviour in the high risk situation, overall hypoactive individuals showed a greater degree of behavioural plasticity across treatments compared to hyperactives. Taken together our results support the hypothesis of differential predation risk among personality types. We discuss the implications of our results for current interest in personalities, and for predation risk and group-living theories.