Two tests of a stochastic dynamic programming model of daily singing routines in birds.

Many hypotheses have been put forward to account for the dawn chorus in birds. Few of these, however, are able to account for variation in song output over the whole day, or for differences in daily singing routines between species, individuals, seasons and environmental conditions. One hypothesis that does offer a more general explanation is based on a stochastic dynamic programming (SDP) model of daily singing routines. This model relates the relative costs and benefits of feeding and singing at different times of day to the size of a bird's fat reserves and calculates the optimal daily routines of singing and foraging that will maximize the amount that the bird can sing while avoiding starvation. The use of SDP models in behavioural ecology has become well established, but they remain largely untested empirically. I tested two predictions of the SDP model of daily routines of singing, using free-living European robins Erithacus rubecula. The results supported both predictions: (1) food supplementation causing unpredictable short-term increases in foraging success increased subsequent song output; and (2) changes in ambient temperature were positively associated with changes in subsequent song output. Copyright 1999 The Association for the Study of Animal Behaviour.