Traplining in bumblebees (Bombus impatiens): a foraging strategy's ontogeny and the importance of spatial reference memory in short-range foraging.

To test the relative importance of long-term and working spatial memories in short-range foraging in bumblebees, we compared the performance of two groups of bees. One group foraged in a stable array of six flowers for 40 foraging bouts, thereby enabling it to establish a long-term memory of the array, and adjust its spatial movements accordingly. The other group was faced with an array that changed between (but not within) foraging bouts, and thus had only access to a working memory of the flowers that had been visited. Bees in the stable array started out sampling a variety of routes, but their tendency to visit flowers in a repeatable, stable order ("traplining") increased drastically with experience. These bees used shorter routes and converged on four popular paths. However, these routes were mainly formed through linking pairs of flowers by near-neighbour movements, rather than attempting to minimize overall travel distance. Individuals had variations to a primary sequence, where some bees used a major sequence most often, followed by a minor less used route, and others used two different routes with equal frequency. Even though bees foraging in the spatially randomized array had access to both spatial working memory and scent marks, this manipulation greatly disrupted foraging efficiency, mainly via an increase in revisitation to previously emptied flowers and substantially longer search times. Hence, a stable reference frame greatly improves foraging even for bees in relatively small arrays of flowers.