Microbial legacy effects vs. plant-soil feedbacks: Effects on plant community composition and productivity

Background/Question/Methods

Plant-soil feedbacks can stabilize species co-existence and determine the relative abundance of plant species within communities.  However, little is known about how feedbacks may change across sites with different land-use histories or the relative importance of plant-soil feedbacks as drivers of plant community structure compared to between site variation in the composition and function of soil microbial communities.  We collected soils from two remnant and two restored tallgrass prairie sites and conditioned these soils with nine species of tallgrass prairie forbs and grasses.  These conditioned soils were then used to inoculate community mesocosms which were planted with seedlings of each of our nine plant species.  In a parallel experiment we followed similar methods using soils from an additional study site, and we also measured the abundance of mycorrhizal fungi and rhizobia in conditioned soil to determine their contributions to plant-soil feedbacks.

Results/Conclusions

We found effects of both soil conditioning treatment and site (the original source of the soil inoculum) on plant community composition and total community productivity.  Plant communities inoculated with soils originally collected in remnant prairie were more productive than plant communities in soils originally from restored prairie.  Seven of our nine study species were significantly affected by site effects, but there were not consistent differences in plant community composition between remnant and restored prairie soil.  Soil conditioning treatments also affected total productivity and composition of plant communities.  The magnitude of productivity effects was similar to the differences in productivity between remnant and restored prairies.  However, plant community composition did not differ between most conditioning treatments, and strong effects were only observed for soils conditioned by Tradescantia ohiensis.  We did not observe any interactions between site of inoculum origin and soil conditioning treatments, indicating that plant-soil feedbacks are likely to be consistent across initial variation in soil communities.  We also observed significant differences in the total abundance of mycorrhizal fungi and rhizobia in soils conditioned by different species, but this explained little variation in total productivity or species abundances within our mesocosms.  These differences did explain the strong effects of Tradescantia; mycorhizal fungi abundance was very low in Tradescantia soils as compared to other conditioning treatments.  Overall, our results indicate that variation in the composition of soil communities resulting from previous associations with different plant species and due to between site variation can affect the composition and function of plant communities.