The importance of mycorrhizal symbioses to plant fitness and community structure is clearly established but the actual mechanisms causing these effects are not well understood. Arbuscular mycorrhizal (AM ) fungi vary widely in their symbiotic effects on their plant hosts ranging from mutualists to parasites. Genotypes of both fungus and host, as well as environment, seem to control AM function. However, there are currently no established principles to accurately predict when particular plant-fungus-soil assemblages will be mutualistic and when they will be parasitic. To begin to formulate a more unified hypothesis of mycorrhizal function, we tested the predictions of three complementary models. The trade balance model predicts mutualistic AM associations when soil P is limited and photosynthate (light and N) is not limited, and parasitic associations when the limitations are reversed. The functional equilibrium model predicts that mutualistic AM associations will produce more arbuscules and extraradical hyphae than less beneficial ones. The co-adaptation model states that, over a sufficient number of generations, ecotypes of plants and AM fungi should become adapted to each other and to their edaphic environment—mutualistic function is predicted if suites of co-adapted plant and fungal genotypes are most successful in exchanging carbohydrate for minerals and have the highest fitness. We tested these models by studying AM symbioses in Andropogon gerardii in grasslands at Fermi National Laboratory, and Cedar Creek and Konza Prairie LTER sites. Soils at these sites form a gradient of N and P availability. Four experiments were conducted using plants, soils, and AM fungi from each site.

Results/Conclusions

Results of these experiments support the trade balance, functional equilibrium, and co-adaptation models of mycorrhizal function. AM symbioses were strongly mutualistic at Fermi and Konza, the two grasslands with low soil P, and they were less mutualistic and sometimes parasitic at Cedar Creek where soil P availability is very high. Also, AM benefits were reduced when light was reduced. As predicted, formation of arbuscules and extraradical hyphae was positively correlated with mutualistic function; and both plants and AM fungi had the highest fitness when grown with their local soil and home symbionts. These findings help elucidate the factors controlling mycorrhizal function in natural grasslands and may be used to guide management decisions involving mycorrhizae.