The principle of competitive exclusion holds that the strongest competitor for a single resource can exclude other species. Yet in many systems, more similar species appear to stably coexist than the small number of limiting resources. Understanding how and when similar species can stably coexist has taken on new urgency in managing biological invasions and their ecological impacts. Recent theoretical advances emphasize the importance of predators in determining coexistence. The effects of predators, however, can be mediated by behavioral changes induced in their prey as well as by their lethality. In this dissertation, I ask how considering multiple trophic levels changes our understanding of how a grass invasion (Pennisetum ciliare) affects species diversity and dynamics in southeastern Arizona. In considering interactions with plant consumers, and with the predators of those consumers, this research reveals more general ecological processes that determine species diversity across biological communities. I first present evidence from a grass removal experiment in the field that shows increased emergence and short-term survival of native perennial plants without grass. This is consistent with Pennisetum ciliare causing the observed concurrent decline in native plant abundance following invasion. I then present results from greenhouse and field studies consistent with that suppression of native plants being driven primarily through resource competition rather than increased rodent granivory. Granivorous rodents do not solely function as consumers, however, because they cache their harvested seeds in shallow scatter-hoards, from which seeds can germinate. Rodents thus act also as seed dispersers in a context-dependent mutualism. The primary granivores in areas invaded by Pennisetum ciliare are pocket mice (genus Chaetodipus), which have a well-studied tendency to concentrate their activity under plant cover to avoid predation by owls. Because the dense canopy of the grass may provide safer refuge, I hypothesized the pocket mice may be directly dispersing native seeds closer to the base of the invasive grass. Such a behavior could increase the competitive effect of the grass on native plant species, further driving the impacts of the invasion. By offering experimental seeds dusted in fluorescent powder and tracking where the seeds were cached, I show that rodents do preferentially cache experimental seeds under the grass. This dispersal interaction may be more general to plant interactions with seed-caching rodents across semi-arid regions that are experiencing plant invasions. Finally, I ask how the predator avoidance behavior exhibited by these rodents affects their ability to coexist with one another. Not only could their diversity affect that of the plant community, but the effects of plant invasions can cascade through other trophic levels. Theoretical understanding of how similar predator avoidance strategy alters coexistence had not yet been developed, however. Instead of a field study, therefore, I modified a general consumer-resource model with three trophic levels to ask whether avoidance behavior by the middle trophic level alters the ability of those species to coexist. I found that more effective avoidance behavior, or greater safety for less cost, increased the importance of resource partitioning in determining overall niche overlap. Lowering niche overlap between two species promotes their coexistence in the sense that their average fitness can be more different and still permit coexistence. These results provide novel understanding of behavioral modifications to population dynamics in multi-trophic coexistence theory applicable to this invasion and more broadly.