To better understand the factors shaping the composition of arbuscular mycorrhizal fungal (AMF) communities we surveyed the AMF communities colonizing the roots of and existing in the soil beneath 20 individuals of Andropogon gerardii and 25 individuals of Sorghastrum nutans in a serpentine grassland. Because hyphae of different AMF species are mostly morphologically indistinguishable a method other than visual inspection must be used. We cultured the fungi in the greenhouse and scored species presence or absence in a sample using spore morphology. For each individual plant we created pairs of greenhouse trap cultures, one culture with only roots as the fungal inoculum and another with both roots and field soil. To create the pairs of cultures we first removed the roots from the soil sample, cut them into 2 cm pieces, and randomly divided the pieces in half. We then sterilized half the soil, added back roots and sterile sand, and planted the two cultures. We also measured six physical soil properties: gravimetric soil water content, nitrate N, ammonia N, P, Ni, and Ca:Mg. Using Canonical Correspondence Analysis (CCA), we investigated the relative importance of these environmental variables in structuring for AMF community composition. Describing both the root-only and whole soil communities allowed us to determine if soil factors influence which AMF are colonizing roots independent of their presence in the soil.

Results/Conclusions

Thirteen AMF species were identified in the cultures. The most common species, Glomus aggregatum, was present in 41 of 90 cultures or 32 out of 45 paired cultures. The next most common were G. claroidium, present in 21 pairs, and Scutellospora calospora, present in 20 pairs. The composition of the AMF community did not tend to cluster by host plant species or type of inoculum. Additionally, there were no strong patterns linking AMF community composition and soil properties. There appeared to be an inoculum-type cultivation bias. Cultures started with roots were missing, except in a single case, the four large-spored species (diameter > 120µm) present in cultures started with whole soil inocula. There was also a relatively large number of cases, 24, where a particular species was present in the root-only trap culture but not in the paired whole soil culture which included roots. These cases indicate a lack of consistency in culturing AMF from roots. Taken together these findings indicate that root-only trap cultures should not be relied upon to give a true representation of an AMF community.