Evolution of avian roosting behaviour: a test of the information centre hypothesis and of a critical assumption

Abstract

After a number of carrion crows, Corvus corone corone, had fed on a rich food source, many more birds appeared at this source after an intervening visit to the communal roost suggesting that information was transferred between birds. The majority of individuals that appeared on the following days, however, had learnt where food was by themselves the day before, and not by following successful foragers from the roost to their rich food source. The information centre hypothesis contains the important, but previously untested, assumption that the number of birds present at the site at a given time represents the total number of birds so far informed. This assumption can be evaluated only by estimating the daily turnover of birds at the food patch, which necessitates working with marked individuals. The present study, which used marked birds, shows the assumption is unjustified. The apparent information transfer effect was produced by a turnover of birds at the food patch, which resulted in more birds knowing about the food source than were actually present. The following morning the majority of these knowledgeable birds arrived at the food source early, producing the same effect as would be predicted by the information centre hypothesis.