experiments. Bees increased their selectivity as the number of variable traits increased from 1 to 3 in different experiments (Fig. 1.2). All three response variables showed the same trends. As among-trait variability increased, more bees showed preferences for some of the flower types available (Fig. 1.2a, F = 8.14, df = 9, p = 0.021, rz = 0.50); also, our measure of constancy increased (Fig. 1.2b, F = 15.88, df = 9, p = 0.004, r2 = 0.66) and Bateman's Index approached +1, indicating that moves among flowers of the same type were increasingly common (Fig. 1.2c, F = 31.51, df =9, p = 0.0005, rz = 0.80).
In contrast, increased variation within states of a single trait (Fig. 1.2, 1-4 states of one trait) did not increase preference (Fig. 1.2a, F = 0.0006, df = 6, p = 0.98, rz = 0.0001), constancy (Fig. 1.2b, F = 0.01, df = 6, p = 0.92, rz = 0.002), or Bateman's Index (F = 0.04, df = 6, p = 0.57, rz = 0.07).
The observed increase in selective foraging behavior is not explained simply by increases in the number of different flower types learned - this is clear from the behavior of bees in experiments testing the same number of flower types. In arrays with 4 flower types, variation within a single trait (4 colors, Experiment 6; floral complexity, Experiment 7) provoked less constancy than the same amount of variation among traits (color and complexity, Experiment 8; color and size, Experiment 9; Fisher's exact test, p < 0.05).
Are these patterns of selective foraging behavior consistent with Darwin's hypothesis or the search image hypothesis.? Darwin's idea predicts that bees should be constant only when there are costs associated with switching among flower types differing in handling methods. Because bumble bees experience negligible costs associated with switching between two different handling methods (Laverty 1994&; Gegear & Laverty 1998), bees were not expected to display constancy in any experiment. Although bees did forage randomly when flower types differed only in complexity, bees were constant when flowers varied in more than one trait. In addition, bees were constant in several experiments that did not involve variation in complexity. When bees are presented with multiple floral signals, the search image idea predicts that individuals should focus on one floral feature (e.g., yellow color) and restrict their visits to flowers with similar features. Some bees tended to show constancy to color in some experiments, but most did not do so in experiments in which color alone was variable. The observed patterns are not accounted for by either hypothesis.
Was this article helpful?