A test of local adaptation to soil microbial communities in tallgrass prairie plant assemblages

Background/Question/Methods

Establishment of diverse plant communities is the cornerstone of some restoration efforts.  Local plant sources are frequently promoted for re-introduction because ecotypic variation results from local adaptation.  If ecotypes of dominant species support different soil microflora, plant-soil feedback could result and contribute to ecotypic variation.  We tested whether native plant assemblages from Kansas and Illinois were locally adapted to their ‘home’ soil microbial community by reciprocally crossing soil and plants in a greenhouse experiment.  We obtained seeds and soil from two remnant prairies in eastern Kansas and central Illinois with similar species composition, but differing climates.  Seeds of four species (Andropogon gerardii, Elymus canadensis, Lespedeza capitata, Oligoneuron rigidum) common to both locations were collected, germinated, and transferred to pots to create 4-species assemblages from each region.  Soil from an Illinois agricultural field (AG) was also used as an inoculum treatment to increase relevance to restoring prairie in formerly cultivated soil.  Kansas and Illinois plant assemblages were subjected to a fully factorial combination of soil inocula (KS, IL, AG) and soil sterilization treatment (sterilized or live).  Plants were harvested after 20 weeks and soil was analyzed for microbial composition using phospholipid fatty acid (PLFA) markers.

Results/Conclusions

Microbial PLFA composition in sterilized soil differed from live KS and IL prairie soil, but not AG soil. Live IL and KS inoculated soils had similar microbial communities, but KS soil was dissimilar to AG soil from Illinois (P<0.001).  Sterilization decreased PLFA biomass (P=0.005), although microorganisms persisted.  Biomass of grasses differed among soil inocula.  Andropogon gerardii from Kansas had highest ANPP when inoculated with live AG soil (P=0.005). Elymus canadensis from Illinois exhibited highest ANPP when inoculated with live IL prairie soil (P=0.020).  Addition of sterilized KS soil increased ANPP of A. gerardii (P=0.004) and E. canadensis (P<0.001) plants from Kansas relative to plants inoculated with live KS soil.  In AG soil, sterilization increased ANPP of E. canadensis plants from Kansas (P=0.009) and Illinois (P=0.008).  ANPP of L. capitata decreased with soil sterilization (P<0.001) in all treatments, as did O. rigidum from Kansas inoculated with KS (P=0.015) and AG soil (P=0.009).  No clear benefit of ‘home’ or ‘live’ soil was detected for Kansas or Illinois plants assemblages.  Thus, ecotypic variation among plants from different regions cannot be easily attributed to an extended phenotype effect on soil communities, particularly from multispecies plant-soil feedback studies.