Overyielding of maize cultivar mixtures is associated with interaction-driven depth partitioning of roots

Context of problem: The use of cultivar mixtures to increase agricultural productivity is well-studied and overlaps conceptually with research on ecological diversity and ecosystem function. However, how roots of different cultivars, or any intraspecific taxa, might interact in ways that promote yield advantages in mixtures is unknown. Objective of research question: Our goal was to understand the effects of maize cultivar mixtures on productivity, and to quantify the responses of root traits and spatial distributions to mixtures as potential mechanisms for increases in productivity. Methods: Productivity and root traits were measured in a field experiment including eight different maize cultivar monoculture treatments and 2−, 4−, 6−, 8−cultivar mixtures in 2018 and 2019. Root traits and distributions were analyzed with Fourier Transform Infrared spectroscopy (FTIR) to quantify root weight density, and for the first time for any conspecifics, spatial root segregation, measured as the proportion of roots of each cultivar in a specific soil volume, in response to neighbors. Results: Cultivar mixtures increased maize grain yield by 3.1 % and aboveground biomass by 6.3 %, across all mixtures combined, compared to the average of the eight monocultures, but not the best cultivar each year, and specific two-cultivar combinations (LY66/XY335) showed even greater yield increases, up to 9 % for grain yield and 12 % for aboveground biomass. Overyielding was derived almost completely from positive complementarity effects and not selection effects. However, 2−, 4−, 6−, and 8−cultivar mixtures increased root surface area, root length density and root weight density per volume in the topsoil (0–0.2 m) compared to those cultivars in monocultures, suggesting substantial plastic responses in mixtures. Most importantly, two-cultivar mixtures showed significantly more spatial root segregation than predicted from monocultures, which may have alleviated competition. Across different mixtures, the degree of root spatial segregation, in mixtures relative to in monocultures, correlated with grain mass and aboveground biomass productivity. Conclusions: Our findings emphasized that appropriate cultivar mixtures could increase productivity via the plastic responses of cultivars to “non-self” conspecifics, including greater root segregation, which could contribute to sustainable agroecosystems. Implication or significance: Although computationally attributed to complementarity, our results are fundamentally different than classic niche partitioning based on inherent differences in root traits. Instead, they demonstrate interaction-driven plasticity among cultivars.