It pays to be distasteful and smelly: Agent based models suggest that microbial metabolism may interact with pollinator learning and memory to shape floral microbiome assembly

Nectar-inhabing microbes rely on animals for dispersal and vary in dispersal limitation. Many of these microbes produce pollinator-perceptible volatile compounds or modify pollinator-relevant nectar chemistry. While it is intuitively clear that these signaling and reward strategies should shape microbe dispersal, the consequences of microbial effects on nectar attractiveness and palatability for nectar microbiome assembly are unclear. Here we us an agent-based simulation model of pollinator learning and memory to test how microbial cues and chemistry changes impact microbial dispersal. Inspired by empirically measured microbial traits, we modeled focal microbes that 1) either did or did not produce volatile cues, and 2) modified nectar palatability for pollinators positively or negatively. Preliminary simulations suggest that microbes which make nectar less palatable and produce a volatile cue of their presence colonize new flowers more quickly and have higher landscape level incidence than competitors. Alternatively, microbes that increase pollinator preference and signal microbial presence spread among metacommunities slower, but reach higher densities in the flowers where they occur. Currently we are testing how the success of these strategies might depend on floral longevity and interact with pollinator memory. These models are a first step in predicting how microbial metabolism impacts dispersal and generate testable hypotheses about the relationship between microbial signaling/reward strategy and landscape level patterns of floral microbiomes. Further, the models highlight the importance of animal cognition in shaping dispersal limited epi-floral communities, and raise the question of whether similar dynamics might be important for animal dispersed microbes in other systems.