FORAGING CHOICE IN THREE SPECIES OF PIPILO (AVES: PASSERIFORMES): A TEST OF THE THRESHOLD CONCEPT (BAYESIAN, TOWHEE).

Abstract

Foraging by captive individuals of three species of towhees (Pipilo:Emberizinae) was investigated to determine if patch choice was associated with critical threshold levels of foraging success. Birds were offered a choice of feeding in three litter types. Once individual preferences for litter type were determined, I reduced the amount of food (millet seed) under a preferred litter each day, while maintaining high levels under a non-preferred litter. Thus, birds experienced a declining resource gradient under their preferred litter. During the bird's initial foraging period each day, I noted number of seeds taken and number of double-scratches made in each litter type, until the bird abandoned its preferred litter and shifted to the other patch. Foraging theory assumes that patch choice is made when foraging success declines as resources decline to some threshold level. I estimated success using information potentially available for decision-making by the birds. One of four success variables, mean seeds/scratch, best conformed to the expected pattern, providing supportive evidence for the threshold assumption. The three species changed patches at species-specific thresholds. Pipilo fuscus, a foraging generalist, switched to its non-preferred litter when food levels under the preferred litter were high. Pipilo aberti and P. erythrophthalmus, two foraging specialists, foraged in their preferred litters until food levels were much lower. Optimal Foraging Theory predicts that shifts in patch use are determined by between-patch comparisons of foraging success. Timing of towhee patch changes was not consistent with between-patch comparisons. An alternative expectation theory, Status-quo Foraging, which emphasizes comparisons of within-patch success, was more consistent with observed results. The rapid patch switching shown by P. fuscus may reflect greater flexibility of this species' foraging repertoire in its natural habitat. This result suggests that foragers with imperfect knowledge may use recent past experience to make foraging decisions. Optimal Foraging Theory provided a useful null model for comparing foraging theories. Deviations from predictions indicated how more realistic hypotheses can be constructed. The nature of available information plays an increasingly important role in modeling decision-making by imperfect foragers.