Anther-smut disease caused by Ustilago violacea is transmitted by insects that visit diseased flowers and then deposit spores on healthy Silene alba plants. The transmission rate of this disease therefore depends on spore deposition by the vector species and preferences for infected vs. healthy hosts. We observed natural populations of S. alba to document the most abundant insect visitors, which include syrphid flies, andrenid bees, bumblebees and a variety of nocturnal moths. Using bumblebees and moths in flower preference studies, we showed that bumblebees preferentially visited healthy flowers, but the strength of this preference declined if the bees had prior exposure to diseased flowers. Nocturnal moths showed less discrimination with respect to disease status when plants were arranged in a field population of S. alba, and preferentially visited plants with more flowers. A laboratory analysis of spore deposition by bumblebees showed that whereas most spores were deposited on the first several flowers visited, flowers beyond the 15th visit may still receive enough spores to produce a new infection. Spore deposition was also influenced by changes in vector behavior associated with spacing in artificial arrays of S. alba plants. Models were fitted to deposition data to compare estimates for spore dispersal rates and the initial number of spores acquired by vectors. This study demonstrated several ways that vector behavior can influence rates of disease spread in natural populations. Preferences for healthy vs. infected hosts can cause disease dispersal to vary from that expected by randomly foraging insect vectors. Changes in visitation behavior associated with host density can affect the magnitude and distance of spore deposition in host populations.