The optimization of a weed control program depends on early and reliable prediction of the impact of weeds on crop yield. This research aimed to evaluate the effects of rice cultivars and plant arrangements on flooded rice grain yield loss and to identify the explicative variable that provides the best adjustment to the model tested. Two field experiments were carried out during the 2000/2001 growing season. Treatments consisted of three rice cultivars (BRS-38 Ligeirinho, IRGA 417, and BR-IRGA 409), two row widths (15 and 25 cm), and plant populations (ten and six levels for the first and second experiments, respectively) of the rice cultivar EEA 406, mimicking red rice infestation.. Data were analyzed using the non - linear regression model of rectangular hyperbole, adjusted independently for the three studied factors. Results of the first experiment were used to evaluate data adjustment to the model and results of the second to validate the model generated in the first. IRGA 417 cultivar presented greater competitive ability than the other two. Reduction of row width increased the ability of rice cultivars to compete with the concurrent genotype. The rectangular hyperbolic model was suitable to estimate flooded rice grain yield loss due to interference of concurrent plants. The explicative variable leaf area of EEA 406 cultivar, in general, was equivalent to plant population as an independent variable in the mathematical model tested.
Oryza sativa; red rice; competition; modeling
