Abstract
The importance of species richness to ecosystem functioning and services is a central tenet of biological conservation. However, most of our theory and mechanistic understanding is based on diversity found aboveground. Our study sought to better understand the relationship between diversity and belowground function by studying root biomass across a plant diversity gradient. We collected soil cores from 91 plots with between 1 and 12 aboveground tree species in three natural secondary forests to measure fine root (≤ 2 mm in diameter) biomass. Molecular methods were used to identify the tree species of fine roots and to estimate fine root biomass for each species. This study tested whether the spatial root partitioning (species differ by belowground territory) and symmetric growth (the capacity to colonize nutrient-rich hotspots) underpin the relationship between aboveground species richness and fine root biomass. All species preferred to grow in nutrient-rich areas and symmetric growth could explain the positive relationship between aboveground species richness and fine root biomass. However, symmetric growth only appeared in the nutrient-rich upper soil layer (0–10 cm). Structural equation modelling indicated that aboveground species richness and stand density significantly affected fine root biomass. Specifically, fine root biomass depended on the interaction between aboveground species richness and stand density, with fine root biomass increasing with species richness at lower stand density, but not at higher stand density. Overall, evidence for spatial (i.e. vertical) root partitioning was inconsistent; assumingly any roots growing into deeper unexplored soil layers were not sufficient contributors to the positive diversity–function relationship. Alternatively, density-dependent biotic interactions affecting tree recruitment are an important driver affecting productivity in diverse subtropical forests but the usual root distribution patterns in line with the spatial root partitioning hypothesis are unrealistic in contexts where soil nutrients are heterogeneously distributed.