Most vascular plants are dependent on mutualistic associations with arbuscular mycorrhizal fungi (AMF) for improved growth and survival. These soil-born fungi colonize plant roots and provide their hosts with increased abilities to uptake scarce resources in exchange for carbohydrates. Although it is becoming clear that AMF-mediated processes are important in herbaceous plant communities in temperate regions, it is not known if similar processes are equally important in other systems such has tropical forests. Accordingly, my dissertation research not only demonstrates that community composition of AMF in tropical forests of Panama differ across spatial scales, but also that differences in AMF composition have the potential to influence the makeup of tropical tree communities. Specifically, I show that AMF spore communities of mainland differ from those of small forest fragments, thereby suggesting that forest fragmentation may promote AMF community divergence. I further demonstrate that both growth and allocation of seedlings are dependent on the species identity and life history of the fungus colonizing its roots. Species comprising the AMF community of Panama differ in dispersal strategies. The majority of species are restricted to slow dispersal through the soil; however, a subset of the species has evolved the ability to be dispersed by rodents. Seedlings of Eugenia nesiotica and Apeiba membranacea hosting rodent-dispersed AMF species grow faster than those hosting AMF not dispersed by rodents. However, in seedlings of two shade-tolerant tree species (Randia armata and Virola surinamensis ) associated with rodent-dispersed AMF under low light, net assimilation rates are reduced, thereby suggesting that these fungi may be more costly when seedlings are growing in the forest understory. Finally, I show that when seedlings of different species of tropical trees are provided with identical communities of AMF, host-dependent divergence of AMF communities occurs relatively quickly, with such divergence being of sufficient strength to differentially influence performance of the next generation of tree seedlings. Taken as a whole, this research provides strong evidence that plant-AMF interactions in the tropics are species specific, and such interactions are most likely important in shaping both plant and AMF communities.