Caching economics: jays cache more when handling times are short and habitats are poor

In this paper we develop a simplified and experimentally tractable version of Andersson & Krebs’s (1978, Animal Behaviour, 26, 707–711) classical model of caching behaviour. We conducted three experiments using captive blue jays, Cyanocitta cristata, to test the predictions of this model. These experiments explored the effects of three theoretically important variables: availability time (Ta), handling time (h) and background foraging rate (γ). In experiment 1, we tested the predicted effect of availability time, and we found that blue jays cached more food items when availability time was short. In experiment 2, we examined the interaction between availability time and handling time, and we found a significant effect of handling time (jays cached more when handling times were short), but no effect of availability time (apparently contradicting experiment 1). In experiment 3, we considered the predicted interaction between handling time and habitat richness (as represented by the background foraging rate). Experiment 3 revealed significant effects of both handling time and background, but no interaction between the two variables. We discuss the implications of these experiments for the further development of our model.