We investigated differences in host infection by a desert mistletoe, Phoradendron californicum, and examined one of the processes that contributes to these differences: variation in seed deposition among host individuals and species. In the Sonoran Desert, P. californicum parasitizes the sympatric leguminous trees Olneya tesota, Cercidium microphyllum, Prosopis velutina, Acacia constricta, and Acacia greggii. We hypothesized that seed deposition depends on host height and crown architecture. At a site in Arizona, frequency of infection did not reflect host relative abundance. Olneya tesota was parasitized at a higher frequency than expected from its abundance and maintained the highest mistletoe loads per individual host. In contrast, P. velutina was infected less frequently than expected. Infection frequency increased with host tree height for all hosts. Mistletoe seed deposition by avian dispersers differed among host species and was disproportionately high in O. tesota and P. velutina. Seed deposition was higher in infected than in non-infected host trees, and increased with tree height in O. tesota but not in C. microphyllum. We suspect that increased seed deposition with height in O. tesota may be due to the preference of seed-dispersing birds for higher perches. Some host tree species, such as C. microphyllum and A. constricta, probably received fewer mistletoe seeds because birds avoid hosts with dense and spiny crowns. Mistletoe populations are plant metapopulations in which host trees are patches and the frequency of infection in each host species/patch type is the result of interspecific differences in the balance between mistletoe colonization and extinction. From this perspective, our study of host use and seed dispersal is a metapopulation study of patch occupancy and propagule distribution among available patch types. Our seed-dispersal study demonstrates that the mechanisms that create pattern in patchy plant populations can be investigated in mistletoe systems.