Variation in dispersal ability among taxa affects community assembly and biodiversity maintenance within metacommunities. Although fungi and bacteria frequently coexist, their relative dispersal abilities are poorly understood. Nectar-inhabiting microbial communities affect plant reproduction and pollinator behavior, and are excellent models for studying dispersal of bacteria and fungi in a metacommunity framework. Here, we assay dispersal ability of common nectar bacteria and fungi in an insect-based dispersal experiment. We then compare these results with the incidence and abundance of culturable flower-inhabiting bacteria and fungi within naturally occurring flowers across two coflowering communities in California across two flowering seasons. Our microbial dispersal experiment demonstrates that bacteria disperse via thrips among artificial habitat patches more readily than fungi. In the field, incidence and abundance of culturable bacteria and fungi were positively correlated, but bacteria were much more widespread. These patterns suggest shared dispersal routes or habitat requirements among culturable bacteria and fungi, but differences in dispersal or colonization frequency by thrips, common flower visitors. The finding that culturable bacteria are more common among nectar sampled here, in part due to superior thrips-mediated dispersal, may have relevance for microbial life history, community assembly of microbes, and plant–pollinator interactions.

Highlights d Nectar bacteria Acinetobacter induce pollen germination and bursting d Pollen germination and bursting increase protein in solution d Germinability of pollen benefits Acinetobacter, but not nectar yeast

For many flower visitors, pollen is the primary source of non-carbon nutrition, but pollen has physical defenses that make it difficult for consumers to access nutrients. Nectar-dwelling microbes are nearly ubiquitous among flowers and can reach high densities, despite the fact that floral nectar is nitrogen limited, containing only very low concentrations of non-carbon nutrients. Pollen contains trace micronutrients and high protein content but is protected by a recalcitrant outer shell. Here, we report that a common genus of nectar-dwelling bacteria, Acinetobacter, exploits pollen nutrition by inducing pollen germination and bursting. We use time course germination assays to quantify the effect of Acinetobacter species on pollen germination and pollen bursting. Inoculation with Acinetobacter species resulted in increased germination rates within 15 minutes, and bursting by 45 minutes, as compared to uninoculated pollen. The pollen germination and bursting phenotype is density-dependent, with lower concentrations of A. pollinis SCC477 resulting in a longer lag time before the spike in germination, which is then closely followed by a spike in bursting. Lastly, A. pollinis grows to nearly twice the density with germinable pollen vs ungerminable pollen, indicating that their ability to induce and exploit germination plays an important role in rapid growth. To our knowledge, this is the first direct test of non-plant biological induction of pollen germination, as well as the first evidence of induced germination as a method of nutrient procurement, as the microbes appear to hijack the pollen’s normally tightly controlled germination mechanisms for their benefit. Our results suggest that further study of microbe-pollen interactions may inform many aspects of pollination ecology, including microbial ecology in flowers, the mechanisms of pollinator nutrient acquisition from pollen, and cues of pollen germination for plant reproduction.

22Dispersal, particularly variation in dispersal ability among taxa, affects community assembly in 23 individual communities and biodiversity maintenance within metacommunities. Although fungi 24 and bacteria frequently coexist, their relative dispersal abilities are poorly understood. Here, we 25 compare the incidence and abundance of culturable flower-inhabiting bacteria and fungi among 26 individual flowers. Using collections that span two coflowering communities across two years, 27we assess viable bacterial and fungal incidence and abundance within individual flower samples, 28and examine patterns across plant species that differ in flower traits. Our results demonstrate 29 that bacteria can be detected in more flowers and in greater numerical abundance than fungi, 30 particularly in flowers with more exposed corollas. For fungi, however, flowers with long 31 corollas were equally likely as exposed flowers to contain cells, and hosted higher numbers of 32 fungal cells, primarily yeasts. Across all flowers, bacteria and fungal incidence was positively 33 related, but within flowers containing microbes, bacterial and fungal incidence was negatively 34related, suggesting shared dispersal routes but competition among microbes within flowers. The 35 difference in dispersal abilities of bacteria and fungi identified here may have broad relevance 36 for community assembly of microbes and plant-pollinator interactions. 37