The vast majority of terrestrial plant species live in symbiosis with arbuscular mycorrhizal fungi (AMF). AMF and plants live in complex networks, with roots of individual plants hosting multiple AMF, and single AMF colonizing multiple plants concurrently. Through the exchange of resources, the two partners of this symbiosis can have great effects on each other, effects which can ripple through both communities. What determines the patterns of associations between the partners is still largely unknown. In this dissertation, I examine a variety of factors, and in particular host identity, that could drive the community composition of AMF in roots. I began by surveying the diversity of AMF in roots of 12 plant species at a remnant bunchgrass prairie in Oregon, U.S.A. (Chapter 2). To do that, I first designed new primers for use in polymerase chain reaction (PCR) to specifically amplify DNA from all Glomeromycota species. Using those primers, I found 36 distinct AMF phylogenetic groups, or operational taxonomic units (OTUs) in the roots from the prairie. The proportion of OTUs in the basal order Archaeosporales was greater than in many other environmental surveys. I also conducted an in silico analysis to predict how effectively previously published primers would detect the whole diversity of OTUs I detected. I then assayed AMF community composition in the roots of 50 plants from nine plant species (Chapter 3). To do that, I designed primers specific to 18 of the OTUs detected in the initial field survey and used them to test for the presence of each OTU in the roots individual plants. I used that data to test if AMF community composition in individual roots correlated with host identity, spatial distribution, or soil characteristics. I found host identity was associated with both the richness and the structure of root AMF communities, while spatial distribution and soil characteristics were not. Finally, I performed an experimental test of the effect of host identity and community context on AMF community assembly (Chapter 4). I grew plants from four native perennial plant species, including two common and two federally endangered plants, either individually or in a community of four plants (with one plant of each species). I analyzed the AMF community composition in the roots of all plants after 12 weeks of growth with exposure to a uniform mix of field soil as inoculum. I found that host species identity affected root AMF richness and community composition, and community context affected AMF richness. Only one of the endangered species was highly colonized by AMF, and I did not detect unique AMF communities associated with it. This dissertation provides information on the diversity of AMF at a remnant bunchgrass prairie, an ecosystem which has been the subject of very few studies of AMF. Although a complex mix of factors interact to determine AMF community composition in roots, this work provides strong evidence that host identity plays a major role in that process. / Graduation date: 2013