Simulating Potential Kernel Production in Maize Hybrid Seed Fields

In maize (Zea mays L.) hybrid seed production, achieving the optimum seed yield per unit land area often is based on limited information about the quantity of pollen shed by the male and practical experience synchronizing pollen shed by the male inbred with silk emergence by the female inbred. We recently reported that kernel production per hectare could be simulated fairly accurately under pollen-limited conditions from simple measures of pollen shed and silking dynamics. The objective of this study was to determine whether a simple mechanistic description of the flowering dynamics of male and female inbreds could be used to simulate and optimize kernel production in seed production fields. We estimated kernel production on the basis of flowering dynamics in six commercial seed fields located near Washington, IA, in 2002, which differed in the quantity of pollen production and silk emergence. In all cases, the fields were managed and harvested by standard seed industry methods. Harvested kernel number varied from 8.4 to 23.1 million kernels per female hectare. Simulated kernel number was closely correlated with these measured values (r ² = 0.98). This result indicates that relative differences in kernel production can be assessed directly from inbred flowering dynamics. Examples are provided to show how inbred management can be modeled to optimize harvested kernel number for a given inbred pair. Model simulations, however, overestimated harvested kernel number by 11%, on average, which implies that other plant factors, such as pollen viability, prolificacy, pollen capture by the canopy, or kernel abortion in response to leaf removal during detasseling might have limited kernel production across the six seed fields. Information about these variables can be incorporated readily into the kernel set model to improve its accuracy. This study indicates that kernel production in a hybrid seed field can be simulated from simple measures of inbred flowering dynamics. The model is a useful tool for optimizing harvested kernels for an established inbred pair or for defining initial management protocols for new combinations of inbreds.