PREDICTED POLLEN DISPERSAL BY HONEYBEES AND 3 SPECIES OF BUMBLE-BEES FORAGING ON OILSEED RAPE - A COMPARISON OF 3 MODELS

1. Comparisons among animal pollinators of the spatial distributions o f pollen that they produce have typically been made among morphologica lly disparate pairs of species. In contrast, we investigated the poten tial extent of pollen dispersal by honey-bees (Apis mellifera) and bum ble-bees (Bombus lapidarius, B. pascuorum and B. terrestris) foraging in rows of oil-seed rape (Brassica napus cv Westar). 2. We estimated t he pollen carryover attributable to individual bees by using particula te fluorescent dye as a pollen analogue. Most of the dye was deposited at the first few flowers probed and smaller proportions were deposite d up to the 20th successively probed flower. We found no significant i nterspecific differences in dye carryover mediated by individuals of A . mellifera, B. lapidarius and B. terrestris with respect to either th e amount deposited or the rate of decline in deposition across success ively probed flowers. We present evidence that the dye produced a reas onably good analogue of pollen transfer. 3. Bees typically flew from o ne plant to another nearby in the same row and were strongly direction al in their movements. Bee species differed significantly in their mov ement patterns, with B. terrestris having the greatest mean move lengt h and directionality. 4. We used three kinds of model (a numerical sim ulation and two different sets of diffusion-advection equations) to at tempt to emulate bee movements. The predictions from all models were r easonably consistent with the observed bee movements, although the num erical simulation invariably made the most accurate predictions, parti cularly over the first few moves. 5. Predicted bee movements were comb ined with least-squares models of dye deposition to estimate the spati al dispersal of pollen by each bee species. All models ranked the bees in the same order of decreasing effectiveness in dye dispersal: B. te rrestris, A. mellifera, B. lapidarius, B. pascuorum, although, except for long-distance dispersal, there was only minor variation among the bee species in the predicted extents of dye dispersal (e.g. the models predicted that the median dispersal distance would be approximately t wo intervening plants irrespective of the species of bee). Overall, th e consensus of the models' predictions is that most of the pollen from a source plant is deposited on immediate neighbours, but that long-di stance pollen dispersal in this system extends over approximately 20-4 0 intervening plants from the originating plant, depending on the iden tity of the pollinator.