Endophytic fungi live within tissues of plant hosts without causing symptoms of disease. These fungi are broadly split into the taxonomically and ecologically cohesive Clavicipitaceous endophytes, which infect grasses, and the taxonomically diverse non-Clavicipitaceous endophytes, which are found in nearly all plants and have diverse ecological strategies. My dissertation has two sections: Section A investigates the intersection of Clavicipitaceous endophyte ecology with other ecological theory, including invasion ecology (Chapter II) and community ecology and climate change (Chapter III); Section B investigates the ecology of one group of non-Clavicipitaceous endophytes, the Xylariaceae, using a culture-based study in Ecuador (Chapter IV) and a next-generation sequencing based endophyte survey in Taiwan (Chapter V). Section B is centered on testing the Foraging Ascomycete (FA) hypothesis—the idea that some decomposer fungi may adapt an endophytic lifestyle to escape limitations in primary substrate in both time and space. In Chapter II, I utilized a host-specific Epichloë endophyte present ubiquitously in the European native range of the Pacific Northwest (PNW) invasive grass Brachypodium sylvaticum to test theories of invasion. In Chapter III, I examined the grass Agrostis capillaris in the context of a climate manipulation experiment in prairies in the PNW to elucidate patterns of interaction between multiple symbionts (Epichloë endophytes, dark septate root endophytes, and arbuscular mycorrhizal fungi) within single hosts across climatic variation. In Chapter IV, I began to test the FA hypothesis by examining spatial relationships of Xylaria endophytic fungi in the forest canopy with Xylaria decomposer fungi on the forest floor in a remote Ecuadorian cloud forest. In Chapter V, I build on the results from the previous study, using a novel technique to examine spatial ecology of the Xylariaceae, pairing traditional mycological collection with the preparation of a next-generation sequencing metabarcode library of endophytes over a much greater area. This dissertation includes previously published and unpublished coauthored material.

The fungal microbiome, or “mycobiome” of plants is diverse and important to host health, but the fluxes of fungi among plant hosts and with the surrounding environment are poorly understood. In chapter two, we employed sterile culture techniques and spatial sampling to examine leaves as possible vectors for transfer of their endophytic fungi from the canopy to substrate on the forest floor, as predicted by the Foraging Ascomycete hypothesis. Some foliar endophytic fungal species are also present as wood-decomposing fungi on the forest floor, that transfer of mycelium across these two life history stages can occur, that endophytic life history stages are buffered from environmental conditions in comparison to wood-decomposing fungi, and that spatial linkages between the two life history stages can be observed. In another study, described in chapter 3, wood and leaf wood endophytes were sampled across a 25 ha plot, to explore landscape patterns of mycobiomes, and to explore the concept of a core microbiome in aerial plant tissues. We found that core microbiomes may be observed in a real ecological setting, but that the concept of core must be carefully defined and that some level of buffering from disturbance may be necessary to allow core microbiomes to assemble. In chapter four, we return to examine some of the assumptions and implications of the Foraging Ascomycete hypothesis, with an agent-based model. We model the conditions under which dispersal through falling leaves may represent a fitness-enhancing dispersal strategy for fungi, and that deforestation as is currently underway throughout the world may have impacts on fungi that rely upon a canopy- inhabiting life stage for dispersal. In chapter five, some challenges associated with environmental sampling of microbes using illumina© MiSeq sequences are critically examined. We find that biases introduced by random sampling at various stages of IVenvironmental DNA extraction and illumina© MiSeq sequencing are not well corrected by currently accepted bioinformatic algorithms. In addition, information loss from differential extraction, PCR amplification, and sequencing success, requires that users of MiSeq read libraries to interpret read abundances carefully. This dissertation includes previously published, co-authored material.