Effects of variation in flower number on pollinator visits in Cirsium purpuratum (Asteraceae).

We examined the functional relationships between floral display and two types of bumble bee response, the visitation rate per plant and the number of flowers visited on a plant, in an artificially arranged field population of Cirsium purpuratum. To reduce the variance in data, we collected data for each day separately and adopted a Latin square design in selecting the focal plants within a day. We then tested several types of regressions to each set of data to find the best-fitting line accounting for the observed relationship between pollinator response and display size. We found that the visitation rate of bumble bees per plant was a decelerating function of floral display, and that the number of flowering heads visited on a plant increased linearly with display size. Predicted from the above two functions, the visitation rate per head was independent of floral display and nearly constant within each day. Our results suggest that conventional methods in collecting and analyzing data on pollinator visitation may yield large variance in data derived from temporal and spatial heterogeneity and that improved methods employed here are effective in reducing the variance and estimating patterns of pollinator response to floral display more accurately.