<p>How do herbaceous plant species coexist in the diverse communities of the arid ecosystems of the world? We review understanding for southwestern North America where these communities are threatened by invasive species and climate change. Relative abundances of these species fluctuate from year to year due to their different regeneration niches (germination and growth requirements), which are sensitive to environmental fluctuations. Such regeneration niche differences were an early hypothesis, with much support, for how these species coexist. Seed predation provides other coexistence mechanisms supported by empirical studies. Indeed, dramatic effects of seed predators on community dynamics have been observed through the controls they impose on invasive plant species. Various studies have included measurements of frequency-dependent predation whereby seed predators shift their preferences to the most abundant species as densities change, and predator partitioning, where the different prey species have different patterns of susceptibility to the various predator species. Due to insufficient studies of predator partitioning, we conducted proof-of-concept experiments on seed predation by ants in an herbaceous plant community in the Sonoran Desert. We show that various ant species discriminate differently between the various plant species, meaning that the plant species have different seed-predation niches, potentially contributing to their coexistence. More extensive studies are needed to assess the roles of seed predators in maintaining the diversity of these plant communities. Deficient also, are studies of density feedback that measure how much a specific mechanism promotes recovery of a species from low density, and allows that species to coexist with other members of the plant community. Failure to assess such density feedback empirically in coexistence studies is pervasive and weakens empirical claims for most mechanisms of species coexistence. Key advances can be made by developing and implementing methods to link density feedback to proposed mechanisms in the field.</p>