Environmentally friendly approaches to increasing food production include using the positive effects of plant biodiversity, such as in intercropping. Functional traits are key drivers of these positive effects, in part because variation in functional traits can increase niche partitioning. However, we know little about how variation in functional traits affects the long-term stability of yield in agroecosystems. We conducted a five-year field experiment with five different cropping systems; maize/peanut, maize/soybean, and maize, peanut and soybean monocultures. We compared the productivity of monocultures to intercrops and then compared plasticity in functional traits at different rates of nitrogen supply between the cropping systems. Intercropping generated plasticity, measured here as the inverse of the coefficient of variation, in three functional traits of maize (height, stem diameter, ear height), which probably increases aboveground spatial niche differentiation, and decreases the intraspecific competition of maize. Intercropping also increased the stability of grain yield and aboveground biomass. Plasticity in functional traits of maize correlated positively with year-to-year temporal stability (CV^-1) of grain yield and biomass of maize and with the total productivity of the agroecosystem.

Synthesis and applications. Our study provides evidence of the greater productivity and temporal stability of species-diverse intercropping systems. Interspecific interaction-induced plasticity suggests a unique mechanism for biodiversity effects on ecosystem function, which adds to our understanding of fixed, or inherent, variation in traits among genotypes or species. Theoretically, our findings provide novel insights into how interspecific interactions contribute to ecosystem service, especially in yield temporal stability, by increased trait plasticity of the dominant crop, maize. The results also had implications for applying intercropping in the sustainable management of food-production systems with the use of more crop species. Greater stability in production has the potential to provide a stable income for farmers.

This field experiment was arranged as a split-plot design with two factors and ran from 2017 to 2021. The factors were four nitrogen (N) applications (N0, N1, N2, and N3) and five cropping systems – maize in monoculture (Zea mays L. cv. Xian-yu No. 335), soybean in monoculture (Glycine max (L.) Merr. cv. Ji-yu No. 47), peanut in monoculture (Arachis hypogaea L. cv. Bai-sha No. 1016), intercropped maize/soybean, and intercropped maize/peanut. N application was the main treatment, and cropping system was the sub-plot treatment.

Peanut and soybean monocultures received N-fertilizer treatments at 0, 40, 80, and 120 kg N ha^-1 (N0 through N3) each year. Maize monocultures received either 0, 180, 240, and 300 kg N ha^-1 (N0 through N3) each year. N applications for the maize/peanut intercrop were 0, 110, 160, and 210 kg N ha^-1 (N0 through N3) each year, the same as in the maize/soybean intercrop. One-third of the N was applied as basal fertilizer, and two-thirds was divided into two portions applied at the maize stem elongation stage and the pre-tasseling stage. N supplied to maize vs. legumes differed due to standard crop requirements, but maize in monoculture and intercrops received same N additions and legumes in monoculture and intercrops received the same N additions. The N level in the N2 treatment was based on the recommendation of local farmers, and N1, and N3 were chosen to test reasonable increases and decreases in N supply relative to N2. The N application rate for maize/peanut and maize/soybean was the average of the monocultures. The topdressing fertilizer was only for maize, and the amount was 1/2 that of the monoculture because the proportions of maize was in intercropping was 0.5 of that in the monocultures.. All plots received 52 kg ha^-1 P fertilizer as calcium superphosphate and 83 kg ha^-1 K fertilizer as potassium sulfate before sowing. Throughout the five-year duration of the experiment, no large-scale disease outbreaks were observed, and no pesticides were applied. Maize borers were effectively managed by releasing Trichogramma wasps, and all plots were manually weeded.