Interactions between plants and soil microbes are increasingly recognized as an important component in the functioning of ecosystems. Because these interactions affect and are affected by soil abiotic conditions, restoration efforts must consider the interactions between the plant community, the soil community, and the soil abiotic conditions. We sampled soil from 20 independently restored tallgrass prairies and 8 natural prairie remnants in southern Manitoba. Soils from the restored sites had 4.5 times higher phosphate levels than soils from the remnants. In whole soil assays, big bluestem (Andropogon gerardii ) and Culver's root (Veronicastrum virginicum) had significantly greater growth in soil from the remnant sites. A second growth assay using sterile and inoculated soil from a subset of these sites showed that while big bluestem benefited from soil biota on both remnants and restored sites, the effect was twice as strong on the remnant sites. Our results suggest that plants on restored prairies are less reliant on soil microbes due to the higher fertility found within their soils. Our data suggests that like other ecosystems, residual high fertility in tallgrass prairies may facilitate invasion by non-native plants.

It is well established that plants and soil properties are interdependent. What is less known is the degree to which the distribution of plant and soil properties, in both space and time, match one another. We assessed plant community and soil conditions in permanent sample plots in a tallgrass prairie community from 2002 to 2008. Redundancy analysis showed that the dominant grass species were associated with changes in elevation, soil pH, and gravimetric water content, but not nitrogen or phosphate. Indicator species analysis suggested the community is best described by four cluster types, with half of the plots changing community type over the 6-year sampling period. These cluster types were generally spatially aggregated, with semivariograms showing a range from 80 to 346 m. This degree of spatial structuring was observed in the elevation and soil water content data, but not in measures of soil inorganic N or P. This suggests that natural plant communities do not necessarily show a strong correspondence to all soil variables, especially those that vary rapidly in time and space. The dynamic nature of vegetation patch distribution, combined with their size, may contribute to the loss of species from small tallgrass prairie reserves.