Direct and indirect effects of arbuscular mycorrhizal fungi on plant community composition.

Mycorrhizal fungi both increase and decrease plant diversity. This inconsistency of existing research suggests that to predict how mycorrhizas affect community composition, we need to understand the mechanisms by which they effect these changes. At the community level, I divide these mechanisms into direct and indirect effects of fungi on plant performance. In a community-level experiment, I found that indirect effects of fungi had stronger impacts on plant community composition than did direct effects, and exacerbated the impacts of plant interspecific interactions at low productivity, but ameliorated them at high productivity. Fungi reduced the negative impacts of competition on one species, Hieracium, apparently by allowing nutrient transfer via the hyphal network. However, in an individual-level experiment performed to determine the mechanisms by which fungi influenced interactions, fungi exacerbated the impact of competition. I suggest that because the matrix vegetation was largely non-mycotrophic, transfer of nutrients from the overstory was minimal. Thus, mycorrhizas do not necessarily confer a competitive advantage to the more responsive species. In this system, mycorrhizas generally have a negative impact on species biomass. This suggests that plants maintain the symbiosis because it confers an advantage on some trait other than biomass. Using Achillea millefolium , I tested the hypothesis that seed number was determined largely by plant biomass, while tissue nutrient concentrations affected seed germination success. Although fungi did negatively affect biomass and positively affect tissue phosphorus concentrations, using a path analysis, I found that total plant biomass positively affected both seed number and germination success. Maternal tissue nutrient concentrations did not affect seed success in this species. Ecologists commonly assume that it is possible to extrapolate from response by a fixed group of species to how community level response changes. Using a meta-analysis on species level data, I demonstrate that, as described by Simpson's paradox, it is impossible to make this extrapolation. Although both the community- and individual-level experiments suggest that the importance of indirect effects increases with increasing nutrient availability, it is not possible to assume that communities growing at high rather than low nutrient availability are more strongly affected by the indirect effects of fungi.PhDBiological SciencesEcologyUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/127033/2/3029284.pd