Kernel Distributions in Main Spikes of Salt‐Stressed Wheat: A Probabilistic Modeling Approach

Grain development in wheat ( Triticum aestivum L.) is a complex process that responds to interactions among primary genotypic factors and the environment. This study was conducted to determine the effects of salinity stress on kernel occurrence and kernel mass distributions within the main spike. Mexican semidwarf wheat cuitivars Yecora Rojo and Anza were grown in sand tanks in the greenhouse with saline and nonsaline nutrient solutions. At harvest, each spikelet position and grain position was identified and the weight of every kernel was determined. Hierarchical multiple linear regression models were derived and fit to the kernel mass patterns. In a similar manner, logistic regression models were derived and fit to the kernel occurrence patterns. Kernel mass was shown to be highly dependent on spikelet location long the spike, kernel position within the spikelet, and salinity stress. Results from the logistic regression models confirm that these same factors affect kernel occurrence. Both types of statistical analysis are advantageous, since the changes in kernel occurrence and kernel mass distributions due to each of the above factors can be easily detected and studied.