In the past two and a half decades, the evolution of foraging behavior has been studied mainly via the optimization approach. Recent work suggests an increasingly important role for the study of proximal mechanisms. The goal of the thesis was to examine the evolution of foraging behavior in bumble bees (Bombus spp.) from a mechanistic view. I studied three questions: how bumble bees move among plants while foraging in patch-free situations, how bumble bees decide to depart one multi-flowered plant and move to another, and how bumble bees choose the species from which to gather nectar.; Queen bumble bees foraging in meadows of larkspur (Delphinium nelsonii) were observed to intersperse short, near-neighbor interplant flights with longer flights during which they bypassed many plants. These "skip" flights followed a very consistent pattern during 4 seasons of observation. The flights were more likely when bees visited only a few flowers on a plant, suggesting that the flights were involved with foraging. A model constructed for the underlying mechanism was similar to several increment-decay models from foraging theory. The same model was successfully applied to the case of departure from multi-flowered plants; in tests with the monkshood (Aconitum columbianum) bees' departure probabilities after empty flowers depended on the amount of nectar placed in previous flowers. The results contrasted with previously published "threshold departure rules". In laboratory studies with captive bees, the effect of the increment-decay model was quantified and shown to be consistent with Bayesian theory in its dynamics, assuming a clumped distribution of nectar.; Laboratory experiments with captive bees were done to test whether bumble bees should be considered short-term or long-term rate maximizers. Short-term maximization had been suggested as the reason for apparent risk aversion in bees. Bees were tested for preference in artificial flower arrays that contained contrasting flower types containing equal mean amounts of nectar but differing in variance. In tests, bees were risk-averse in the short term, but in the long term (visits to 150 or more flowers) they became risk-indifferent. The results were consistent with a model derived from animal learning psychology.
展开▼