The foraging movements of bees are said to be influenced by the spatial distribution of the plants whose flowers they visit, However, Manasse found that bee-mediated gene flow was unrelated to the degree of aggregation among the plants in the linear arrays that he studied. To investigate this apparent contradiction further. I manipulated the levels of aggregation in linear plant arrays and examined both the effect on the movements of the bumblebee pollinators and the likely consequences for pollen dispersal and gene flow. I studied pollinator movements among plants of oil-seed rape (Brassica napus cv. Westar) that were arranged in rows that each contained 50 plants. In each experiment one row of plants was spaced at uniform intervals and the remaining rows contained aggregated plant distributions. Experiments were conducted at two scales, short rows (10 m) and long rows (20 m). The plants were visited almost exclusively by worker caste individuals of Bombus lapidarius. Bumblebees were followed over successive visits within a row and the location of each visited plant and the number of flowers probed there was recorded. Within each row, I examined the following components of bumblebee behaviour: the mean number of flowers probed per plant visit; the mean interplant flight distance; and the mean directionality (proportion of moves in an individual's favoured direction). The collective movements of bumblebees within each row were characterised by the rate of change of the bees' mean squared distance from an original location. The spatial dispersion of pollen from a paternal plant was estimated in each row by combining a quantitative description of bumblebee movements in terms of displacement from a paternal plant with a model of pollen deposition. Bumblebees were observed to visit many plants successively within a row and movements between rows were infrequent. In both long and short rows, three components of bumblebee movement behaviour were significantly modified by the level of plant aggregation. With increasing plant aggregation, bumblebees exhibited a decrease in directionality and in the number of flowers probed per plant visit and they exhibited an increase in the number of plants by-passed during interplant flights. Despite the effects of plant aggregation on the components of bumblebee behaviour, the bumblebees' collective rate of movement away from an origin within a row was unrelated to the level of plant aggregation. This was consistent with the finding that plant aggregation did not significantly affect either the estimates of mean pollen dispersal distances measured in metres or a plant's expected number of female mates. The modifications to bumblebee behaviour that were caused by differential aggregation among rows apparently combined, fortuitously, to buffer the spatial extent of pollen dispersal. Consequently, my results are consistent with Manasse's finding that pollinator-mediated gene flow was unrelated to plant spatial heterogeneity at these scales. However, I also conclude that the effects of plant aggregation on gene flow will depend on the spatial scale of the inter-patch distances.
