The influence of variability in nectar reward per flower on the foraging behavior of individual bumblebees (Bombus sandersoni Fkln.) and paper wasps (Vespula vulgaris L.) was explored during late summer 1979. The area around each colony was enclosed in mosquito netting, and an artificial floral patch, yellow and blue cardboard squares randomly mixed and fixed beneath wells in a clear sheet of plexiglass, was established within the enclosure. The nectar reward per flower (i.e., the volume) was manipulated to generate different degrees of variability within a floral type. The first set of experiments had equal quantities of nectar in all yellow and blue flowers. Individual foraging bees showed a significant preference for yellow. However, when blue flowers were kept constant (2 μL in each) and yellow made variable (6 μL in every third flower and none in the remainder), individual foragers avoided the variable yellow even though the expected reward from yellow was equal to blue. As a control, the variable floral type was reversed with yellow now constant. Individual foragers then showed an equally strong avoidance of blue flowers. In order to insure that bees were avoiding uncertainty rather than merely avoiding empty flowers, a new set of experiments was run with some nectar in every flower. The constant floral type contained 2 μL of nectar in every flower while the variable type contained 5 μL in every third and .5 μL in the remainder. Again the foragers preferred the constant floral type; however, the preference was not as great as when there were empty flowers. Analysis of the sequence of visits made by individual forages seems to indicate that bees learned to distinguish the variable floral type during the foraging bout. Similar experiments were run on the wasps. When flowers were of equal quality, the wasps preferred yellow to blue. When one floral type was variable and the other constant, the wasps preferred the constant type. For the same magnitude of variability, however, the bumblebee's avoidance of the variable floral type was more pronounced than was the wasps' response. The results of these experiments are then analyzed within the context of a more general model of foraging behavior in variable environments.