Factorial and 'self vs. other' plant soil feedback experiments produce similar predictions of plant growth in communities

In species-factorial plant-soil feedback (PSF) experiments plants are grown on replicate soils from each other potential plant in a community. Species-factorial experiments are expected to provide the best estimate of PSF effects, but are rarely used because they require large sample sizes. As a result, the extent to which species-factorial data improve understanding of PSF effects on plant community dynamics remains unknown. Published data were used to parameterize a PSF model with either species-factorial or non-species factorial (i.e., 'self vs. other') data. Model predictions were compared to plant growth in communities. Species-factorial PSF data did not improve predictions of plant growth in communities relative to 'self vs. other' PSF data. Both PSF models performed slightly better than a null model without PSFs. Conceptually, species-factorial experiments should detect more PSFs and better predict plant community dynamics. However, because PSF effects were generally small, most likely to change competitive outcomes for species with similar intrinsic growth rates, and because 'self vs. other' approaches capture much of the information in species-factorial approaches, species-factorial- and 'self vs. other-' informed models produced similar predictions. Results suggest the 'self vs. other' approach produces accurate PSF values with many fewer replicate samples.