Bumblebee Foraging Behavior: Changes in Departure Decisions as a Function of Experimental Nectar Manipulations

Abstract

The nectar-gathering foraging behavior of bumblebees (Bombus flavifrons and B. bifarius) was observed on control and nectar-enriched inflorescences of three Rocky Mountain herbs, Aconitum columbianum, Delphinium nelsonii and D. barbeyi, to determine whether bumblebees actually forage in a manner consistent with the learning mechanism proposed by Ollason. When gathering nectar from two of the three species, bumblebees encountering a series of enriched plants initially increased the number of flowers per inflorescence that they visited relative to control inflorescences. After the blossoms from a small number of enriched racemes were probed, the average number of flowers visited per plant returned to control levels. These results are consistent with the predictions made by Ollason and suggest that foragers' assessment of habitat quality is modified during a foraging bout. These results are used to estimate the extent of bumblebees' memories of previously encountered nectar rewards. The small estimated size of bumblebee memories may be adaptive under spatially heterogeneous reward schedules. INTRODUCTION Food resources for animals often occur in patches. Two problems shared by foraging organisms are, therefore, how to assess the quality of various food patches and how to decide when to leave one patch in search of another of higher quality. In an attempt to predict how an optimally foraging organism would deal with food patches of variable quality, Charnov (1976) formulated the marginal value theorem. In brief, the theorem states that a forager should leave its present patch when its rate of net energy intake drops below the average rate of net energy gain for the entire habitat. However, an animal's correct use of Charnov's rule requires a complete knowledge of the resource structure of the environment. It is unrealistic to expect that any organism would ever possess such knowledge, so how is it that observed foraging behavior is often consistent with the model's predictions (e.g., Krebs et al., 1974; Cowie, 1977; Pyke, 1978b, 1982; Heinrich, 1979; Hodges, 1981; Best and Bierzychudek, 1982; Cibula and Zimmerman, 1984)? Ollason (1980), recognizing that animals have incomplete knowledge of their environment, proposed a mechanism by which animals could make approximately optimal departure decisions. His rule states that foragers should continuously compare their current feeding rate with the average rate at which they remember feeding. When their current rate drops below their remembered average they should leave the patch in which they are feeding in search of another. A distinct assumption of Ollason's model is that animals, while gathering food, use the acquired information about reward schedules to continually update their estimate of overall habitat quality. Thus, Ollason's mechanism involves learning on the part of the foraging organism. The decisions actu'Address all correspondence to M. Zimmerman.