Background/Question/Methods Understanding the mechanisms that generate differences in species composition among communities is a central goal of ecology. Dispersal limitation and niche restrictions can work independently or interact to generate the communities we observe, and identifying their relative importance remains a challenge. Arbuscular mycorrhizal (AM) fungi belong to a ubiquitous taxon whose members form a root symbiosis with the majority of terrestrial plants. These fungi have been detected in both geothermal and non-thermal grasslands in Yellowstone National Park, where these grasslands provide islands of good habitats in an otherwise wooded landscape of low habitat quality. Utilizing natural differences in plant community composition, soil temperature and chemistry among grasslands, and their interspersed locations throughout the park, our objective was to evaluate the role of niche restrictions and dispersal for AM fungal community assemblage. We identified AM fungal phylotypes in roots collected from 5 geothermal and 3 interspersed non-thermal grasslands using molecular techniques. Results/Conclusions Fungal phylotype richness did not differ between geothermal and non-thermal grasslands. Certain phylotypes, such as Glomus intraradices, appeared to be generalists and occurred in all grasslands, whereas others, such as a previously undescribed Archaeospora phylotype and a Paraglomus phylotype were only found in acidic geothermal areas. Communities that were closer together were not more similar, suggesting that dispersal limitation of AM fungal phylotypes is of minor importance. Ordination of fungal communities showed two separate clusters; one cluster including the four acidic geothermal sites and one cluster including the three non-thermal sites and the alkaline geothermal site. This ordination mirrored the ordination of the soil chemical properties, indicating that soil chemistry may constituted a major filter, whereas the plant community composition and soil temperature play a minor role in structuring these communities. The abundance of Gl. intraradices was positively correlated with soil pH (r = 0.84, p <0.01), but also showed a negative relationship with the abundance of the Archaeospora and Paraglomus indicator phylotypes of acidic geothermal areas. To separate the role of the abiotic environment from competition for the observed distribution of these phylotypes, a controlled greenhouse experiment with selected geothermal isolates is underway and will be discussed.