Information and exploitation: Patch assessment strategies in birds and mammals.

Abstract

Animals are decision-makers. While we cannot examine directly their decision-making process, we can observe the results of such decisions: their behavior. Behavioral ecologists, attempt to understand and predict the behavior of animals. To do so, clearly requires an understanding of the kinds of information individuals possess about their environment. Here, I explore how foragers exploit resource patches. When resource items are discrete and hidden, individuals must estimate the quality of encountered patches. Solitary foragers can use two sources of information. First, they can obtain information by sampling a patch. This is called patch-sample information. Second, foragers may use their prior experience and learn the distribution of resources in the environment. Individuals that combine these two sources of information are called Bayesian foragers. Foragers that cannot distinguish patch differences must rely on patch-sample information. Those that can, should employ Bayesian estimation because this generates better patch estimates. I examined the patch assessment ability of a wide variety of species. When patch variation is small, Arizona pocket mice (Perognathus amplus), and round-tail ground squirrels (Spermophilus tereticaudus) use patch-sample information; when patch variation is great, they both are Bayesian foragers. Gambel's quail (Callipepla gambelii) and mourning doves use patch sample information regardless of patch variation, while Merriam's kangaroo rat appears to always employ Bayesian foraging. If patches are spatially predictable and quality is temporally predictable, foragers can use this information to generate patch estimates prior to exploitation. Inca doves (Columbina inca) do just this. Black-chinned hummingbirds (Archilochus alexandri) are Bayesian foragers in predictable environments but rely much more on patch-sample information in unpredictable environments. Group foragers can use one additional source of patch information: they can observe the foraging success of other group members. I call this public information. In a series of Monte Carlo simulations, I show that Public information allows groups to estimate patch quality faster than solitary foragers and may prevent the underutilization of patches. Individuals in groups that cannot obtain public information can be faced with a dilemma because each group member will generate a different patch estimate. These individuals should abandon the patch when the first individual does so.