LAUSR

Abstract In semi-arid production regions having a high inter-annual rainfall variability, farmers use low plant densities, that promote tillering in maize (Zea mays L.). Tillering in maize is a scarcely researched trait, even though it offers vegetative and reproductive plasticity. However, the production of tillered maize crops may be risky due to a possible early soil water depletion and/or low reproductive efficiency. Here, we explored the potential effects of tillers (0, 1 and 2 tillers pl−1) on the water economy, plant growth and kernel set of maize crops at a low plant density (2 pl m-2) through the combination of simulations (using an adapted version of APSIM Sorghum model) and field experiments. The simulated scenarios included six locations across Argentina and Australia and three levels of available soil water content at sowing (50 %, 70 % and 90 %). In humid environments, simulated crops with increased number of tillers pl-1 expressed high vegetative plasticity (up to three-fold increases in leaf area index at flowering), driving high values of crop evapotranspiration during the pre-flowering period (EtaPF). In drier environments, simulated EtaPF was similar between crops with different tiller number pl-1. Manipulative field experiments were established to parameterize the function between kernel number per plant (KNP) as a function of plant growth rate around flowering (PGRF) for tillered and non-tillered plants. At low PGRF, tillered plants had lower KNP than non-tillered plants, though the opposite was evident at high PGRF. Interestingly, when simulated PGRF values were used to predict kernel number m-2 (KN m-2), increases in KN m-2 of tillered maize crops in humid environments were larger than the decreases in drier environments. Consequently, farmers could benefit by choosing tillering maize genotypes in the long term when sowing low plant densities.