" Counts of visits to first-nearest, second-nearest, third-nearest, farther-than-third-nearest neighboring Penstemons. b Frequencies of transitions between Penstemons and other flower species, where P = Penstemon and Other = any other species.
c Bateman's Index (see Waser 1986) and the ratio of Other to P are given as summaries of constancy. The mean bout length (last column) is the number of plant visits observed before losing the bee; longer bouts indicate slow-moving bees that are easy to follow. d Data for bee "Blue" are divided into early (23-28 July) and late (5 August) periods.
that collected pollen while young may turn to nectar collecting with age, or vice versa. "Blue," for example, accumulated small corbicular loads during all bouts from 23-28 July, but by 5 August was no longer carrying visible loads.
Even though the relative efforts made into pollen and nectar foraging are genetically controlled (Robinson & Page 1989), there is also strong plasticity in the way in which individuals react to colony needs (Cartar 1992a; Plowright et al. 1993; Fewell & Bertram 1999). There have been recent attempts to understand such task allocation in bee colonies by self-organization models in which each bee is an automaton that differs from other colony members only in the response threshold to particular stimuli in and outside the nest (Bonabeau et al. 1997; Pankiw & Page 2000). Even if these models explain some of the observed behavior, there are potential difficulties, because they neglect the individuality of pollinators beyond their inborn thresholds. All animals encounter a basic difficulty when they set out to perform a novel skill: they generally need to learn that skill, even if it has innate components. The investments in learning different types of foraging activities (and the costs of interference when switching) can be substantial (Dukas & Visscher 1994); therefore, we cannot understand task allocation and task switching without quantifying these costs (and how bees perceive them). Surprisingly, however, one review of new breakthroughs in task allocation (Gordon 1996) avoids such terms as "learning" and "memory" altogether.
Was this article helpful?